Loading...
The URL can be used to link to this page
Your browser does not support the video tag.
Home
My WebLink
About
Railbelt Intertie Reconnaissance Study Vol. 2 Forcast of Electric Demand in the Alaska Railbelt 1989
Alaska Energy Authority LIBRARY COPY RAILBELT INTERTIE RECONNAISSANCE STUDY Volume 2 Forecast of Electricity Demand in the Alaska Railbelt Region: 1988 - 2010 April 1989 Institute of Social and Economic Research University of Alaska Anchorage |i. Alaska Power Authority RAILBELT INTERTIE RECONNAISSANCE STUDY VOLUME 2 FORECAST OF ELECTRICITY DEMAND IN THE ALASKA RAILBELT REGION: 1988-2010 Prepared for Alaska Power Authority P.O. Box 190869 Anchorage, Alaska 99519-0869 Prepared by Institute of Social and Economic Research University of Alaska Anchorage 3211 Providence Drive Anchorage, Alaska 99508 Steve Colt, Principal Investigator O. Scott Goldsmith, Project Manager In collaboration with Adams, Morgenthaler & Company, Inc. Anchorage, Alaska Regional Economic Research, Inc. San Diego, California James E. McMahon, Ph.D. San Francisco, California 30 April 1989 RAILBELT INTERTIE RECONNAISSANCE STUDY VOLUME NUMBER 1 10 11 LIST OF VOLUMES VOLUME TITLE Economic and Demographic Projections for the Alaska Railbelt: 1988-2010 Forecast of Electricity Demand in the Alaska Railbelt Region: 1988-2010 Analysis of Electrical End Use Efficiency Programs for the Alaskan Railbelt Fuel Price Outlooks: Crude Oil, Natural Gas, and Fuel Oil Anchorage-Kenai Transmission Intertie Project Anchorage-Fairbanks Transmission Intertie Expansion and Upgrade Project Railbelt Stability Study Northeast Transmission Intertie Project Estimated Costs and Environmental Impacts of Coal-Fired Power Plants in the Alaska Railbelt Region Estimated Costs and Environmental Impacts of a Natural Gas Pipeline System Linking Fairbanks with the Cook Inlet Area Benefit/Cost Analysis Forecast of Electricity Demand In the Alaska Railbelt Region: 1988-2010 FINAL REPORT April 1989 Contributors The Institute of Social and Economic Research served as the prime contractor for this project. Scott Goldsmith managed the project and Steve Colt served as Principal Investigator. The residential end use telephone survey was designed by Jack Kruse, administered by Karen Foster, coded by Darla Siver, and analyzed by Eric Larson and Alan Mitchell. Rosalynd Frazier administered the commercial mail survey. John Wenger served as scheduling supervisor for the commercial on-site survey. Marybeth Holleman located commercial survey candidate buildings, and gathered primary data on Railbelt energy market structure. Debbie Guderjahn cleaned and coded the commercial sector on-site data set. Theresa Hull prepared several sets of interpretive graphics. Jim Kerr processed the Municipality of Anchorage’s property appraisal files. Toby Lesniak developed our analysis of historical load factors and regional coincidence. Alan Mitchell matched utility billing data to the end use survey records and reduced the commercial on-site and mail survey data. Adams, Morgenthaler, & Company conducted the field portion of the commercial sector on-site survey and provided engineering assistance throughout the project. The commercial survey instrument was designed by David Adams and administered by Max Moorhead. Rob Herrett performed lighting cost and performance calculations. David Crews conducted simulation analyses of commercial buildings. Regional Economic Research supported the commercial sector modeling effort. Stuart McMenamin is the author of COMMEND-PC 3.0, the commercial model. McMenamin assisted with sample frame selection, data development, parameter choice, and model mechanics. James McMahon, Ph_D. is a staff scientist at Lawrence Berkeley Laboratories and the author of the residential end use. model. McMahon analyzed residential appliance efficiency and capital costs for input to the residential forecast. Acknowledgments This project was largely devoted to primary data collection and depended for its completion on the cooperation of several organizations and scores of individuals. We are particularly grateful to: The Alaska Railbelt Utilities for providing data and staff time. Mike Kiech of ML&P deserves special thanks for his prompt responses to a long stream of data requests. Municipal, Borough, and State Government Agencies throughout the Railbelt for critical floorstock data. The Alaska Power Authority for funding a project focusing on primary data collection. APA Project Manager Dick Emerman offered helpful criticism and heroic patience throughout the project. EXECUTIVE SUMMARY ... 2.0... ccc eee teen ene 1. INTRODUCTION 1a Purposeiand Struchites! a5 oe ces 6 tines ses hems mole 12, 1987 Railbelt Eléctnicity: Demand 2.6 ove s news came oes 1:3 End Use Rorecasting Models 5 «12% 6 9c ere ssi 2) = ene ile oe) ens 1:4) reatment of Wincentainty, crass <a cise oie overs lee ool alle hp sects (ene 2. RESIDENTIAL INPUT ASSUMPTIONS DiMA FAOUSING StOCK: vyreelte oe) ticiichs o) setterie 1 =) cltelte el arelte lol) elle oe) o) 9 etd 2.2 Electric Heat Market Share ........... cece cee escreceees 23 Other Appliance:Market Shares 2. 5 oi. 6 oo nie 6 + ole 0 0 eietbas 24 BOUL Valves so cnye ciatvise te ocidle el ciielo e sence ot) se CO a sek er 25 Technical Parameters: ic <5 55006 3 0 oo eo oo oe oo 0 viele « ote) Hie 2:6) Economic) Variables oie eo oss oo 015 im 9 9) oie ie ssi susits Oe] somes 2-1 environmental Hactors) «6s 3 sw mos oo se ws ss ls cy ss see 2:8: Model-Galibration, «. scicc eo ccs 6 0 5 ci0 cc © & sie We) Gl cues 010) cere one 3. COMMERCIAL INPUT ASSUMPTIONS 35 \Commerciall Floor Space =< aysirele els tolls) as sie eked el a) atts ction eestor iene kas 3.2 Electric Market Shares .......... 00. cee eee cece eee eee eees BS VEUL Values) smc oo aces eles a icles) ae eto oieic © oie ene et 3.4 Technical Parameters .......... 0... c eee ee ee eee ens 3.5 Economic Variables ..........cccccccccssecsssssscseces 3:6 Model Calibrations... 5). 1. «eters © 0 9) vile os 9 i © Se) nw © wile ais eee as 4. Evectriciry DEMAND PROJECTIONS 4.1 Treatmentof Uncertaintyar ta. < <0 sisi 6 ois o fcc oo 09 dues 2 toi 4.2 Residential Sales’ jan sacks «0 sm wma as a's 5 Hos ww Hse 6 3 wiouere 4-3) Commercial) Sales) 5 sis 3 aye © 4 seire me) ello fo stress a) mi ouseris gence’ lait 44 Industrial Load: (5 setigs co. aise ris 9 aiaile oi 0) su51G) aici syisis| oo, 4) otletie e esha atte 4.5 Street Lights, Office Use, Losses ........... cee eee eee eee eee 4.6 Total Enerpy Requirements: <5 5 5 sis 5 6) ecie oo cere p% aici we © 0) oun 4:7 Sensitivity*to! Gas Penetration «61s. 3 6 ee > 9 shane om etrvite ore 6 II 4:8) Beak Loadiee fo) oro cnet Worsrcle ribs egclio le © 1-0) lo is)» Sire site 9) sletlageire =)onelie\re 4.9 Comparison With Econometric Results ..............200 ee APPENDIX A: REGRESSION ANALYSIS OF RESIDENTIAL CONSUMPTION APPENDIX B: RESIDENTIAL TECHNOLOGY CURVES APPENDIX C: END Use Mopet SUMMARY OUTPUT APPENDIX D: INDUSTRIAL SCENARIOS APPENDIX E: Loap FActTors APPENDIX F: RESIDENTIAL SURVEY SUMMARY RESULTS APPENDIX G: CALCULATION OF PROJECTED GAS CONVERSIONS APPENDIX H: COMMENTS AND REPLIES APPENDIX I: ADDITIONAL FAIRBANKS LoaDs: MILirary INSTALLATIONS AND THE UNIVERSITY OF ALASKA APPENDIX J: ADDITIONAL LOADS SERVED BY THE NORTHEAST INTERTIE This Page Intentionally Left Blank List OF FIGURES igure) 1.1: Railbelt Electricity Use meses oleae alae leadeneteite eiela) a evelieyetiel ailelte ter ot Figure 1.2: 1987 Railbelt Residential and Commercial Electric Sales by End Use . igure) 2.4: Projected Housings Stocks aieis alee elon a smelelole eo eretelieasveltelo| eyed seis altos Figure 2.2: Effect of NAECA standards on Railbelt Appliance Efficiency ...... Figure 2.3: Retail Residential Electric Prices by Region and Case ........... Figure 2.4: Retail Residential Gas and Fuel Oil Prices, All Cases ........... Figure 2.5: Relative Residential Fuel Prices: Electricity, Gas and Oil......... Figure ss 1s1987) Rail Delt iR1OOLstocksis ele tel elate/ ale) shore fa) a) elicieliolieke) ob sts hel cls| ial stelieliciel a BIgureis 24H 98 /ANCHOragey HLOOLSLOCK dare ieiraley olistaatlatt ae otolinton sy aielletielietieliialieviev May igure! 33711987 Bairoanks) PIOOrStOCK a ataleirel dltey ctiol oy otto} enoioliou oro) susiouollor ol cnsllolisi.c loi BIGULC) 9.451198 7) INCTAl BLOOLStOGKs Isis tae) atellettel alle o}ioltatel oi of elketiay/o) tor oy itelfoycotvol ey eile liattcl oa Bigureis!5:H198 7MatSupelOOrstoCKs eee rae fe nere a a atvelta relatos Alera ae aHal etre Ea Figure 3.6: Comparison of Building Type Composition across Regions ......... igure) 3:7; Rrojected| Railbelt -Piaorstock, Pylori rel o/elsieeifaleits) sve leo) so) et alee ley ells igure! 3.8: bloorstocktProjections| by REGION Mersenieieaiee ieee ele eicrereieiee c Figure 3.9: 1987 Anchorage Commercial Electric Sales Structure ............ Figure 3.10: 1987 Fairbanks Commercial Electric Sales Structure ............ Figure 3.11: 1987 Kenai Commercial Electric Sales Structure .............. Figure 3.12: 1987 MatSu Commercial Electric Sales Structure .............. Figure 3.13: Railbelt Lighting Energy by Lighting Technology Type .......... Figure 3.14: Projected Retail Commercial Electric Prices by Region .......... Figure 3.15: Retail Commercial Natural Gas and Fuel Oil Prices, 3 Cases ...... RigucersulGs Required baybacks/beriodsell| Me aeiaeiely aldol Me Ae vals ated aie ate Figure 4.1: Year 2010 Railbelt Sales by End Use; Res / Comm, 3 Cases ........ Figure 4.2: Residential Sales by End Use, Anchorage & Fairbanks, MID Case ... Figure 4.3: Residential Sales by End Use, Kenai & MatSu, MID Case ......... Figure 4.4: Residential Sales Growth Decomposition, All Regions, MID Case ... Figure 4.5: Residential Electric Heat Shares, Anchorage & Fairbanks, MID Case . Figure 4.6: Residential Electric Heat Market Shares, Kenai & MatSu, MID Case Figure 4.7: Appliance Market Shares, All Regions, MID Case ............... Figure 4.8: Trends in Residential EUIs, Anchorage, MID Case .............. Figure 4.9: Commercial Sales by End Use, Anchorage & Fairbanks, MID Case .. Figure 4.10: Commercial Sales by End Use, Kenai & MatSu, MID Case ....... Figure 4.11: Decomposition of Commercial Sales, All regions, MID Case....... Figure 4.12: Probability Distribution, Total Requirements, 72 Cases ........... Figure 4.13: Railbelt Energy Reqts, 3 Cases and Railbelt Reqts by Customer Class, MED Cased iraiielel terete eer cteietelelevekeietols cireteedel eee heee i tekeperectereteverels Figure/4-14:) Requirements by Region: and) |Gase ji). susie) selieietols) selene ol olleteile| |) = Figure 4.15: Requirements by Region and Customer Class, MIDDLE Case ..... Figure 4.16: Sensitivity to Gas Penetration Assumption: Kenai, MatSu ......... Figure 4.17: Railbelt Energy Requirements and Peak Load, 3 Cases .......... Figure 4.18: Regional Energy Requirements and Peak Load, 3 Cases .......... List OF TABLES Mable: {Railbelt BlectricitysWser- aac ms sete eee isis sess | eS 1-3 Table: 2:1: Railbelt Residential Housing: Stock <5 31556 «sis scl oo sieislelaclsis = 0 « 2-1 Table 2.2: Projected Occupied Housing Stock Growth Rates, All Regions and CaSeSH el rersiiele ii donereber seledellckctleledste] sieliclelalelol eke ole felebeiene (ele) st eicetels 2-4 Table 2.3: Reported Railbelt Heating Fuel Combinations .................-. 2-8 Mable: 2.4;(Gurrent Plectric/MeatiSharesei cy clic lelicle hiieciec idee ec 2-9 Table 2.5: New Construction Electric Heat Shares ...........cccecvevccees 2-9 Table 2.6: Cumulative Projected Gas Conversions ...........e0eecceeeeces 2-11 Table 2.7: Intended Replacements of Water Heaters by Replacement Fuel ..... 2-12 Table 2.8: Residential Electric Market Shares, Other Appliances ............. 2-14 Table 2.9: Miscellaneous Appliance EUI Derivation .................00005 2-16 Table 2:10: Engine Block Heater BUI Analysis ooo) .00 0 oasis ce aleicre ie wise eliele «le 2-17 Table 2.11: Anchorage Residential EUI Calibration Worksheet .............. 2-18 Table 2.12: Fairbanks Residential EUI Calibration Worksheet ............... 2-19 Table 2.13: Kenai Residential EUI Calibration Worksheet ..............6-- 2-20 Table 2.14: MatSu Residential EUI Calibration Worksheet ...............-. 2-21 Table;2215:;ResidentialaPrice Growthi Ratesic)- ofa) folclels sisieiclelo aiseiel seis cls «fe 2-24 Table 3.1: Direct vs Ratio-based Floorstock Estimates .............0000000- 3-7 Table 3:2: Commercial Electric ‘Market Shares) 215 2f0 1). 2 o.dice/ e+ sister coi) ole oe +e 3-12 Table 3.3: Measured Electric: Energy Intensities: 3-0 3 3. on ies es © 3 3-14 Table 3.4: EUI Estimates, Southcentral'Regions ...........0.2e00essccee 3-16 Table 3:5: EUL Estimates, Fairbanks Region’. 4.0%. 3 oi ain ie @ ol lacks = oie 3-17 Table 3.6: Sample Lighting Tradeoff Elasticity Calculation ................. 3-25 Table 3:7: Commercial! Discount Rates) <5 3-16) siloteke oto olloste lee sis cite tales alee tel | a. 3-29 ‘Table: 3:8: MISC. equipment*growth Tates) pac. :oeye ie eo oar ie) of soiree e)er eiisiiers ~ oi ei 3-30 Table 3.9: Development of Commercial Control Total Sales Data ............ 3-31 Table 4.1: Joint Probabilities for Employment/Households and Energy Price SCOMATIOS spereh oro afer sietororosteke rotor see ier oy elisiekoohal elie els) stele) oo elsielis whores sii 4-2 Table 4.2: Critical Assumptions behind the LOW, MIDDLE, and HIGH forecast CASES irerelleieReliclo roll oisneeietouetielelokelsietie= oho) ole foro oiler sacle lel = telraewelin eleoiistiels 4-5 Table 4.3: Railbelt Residential Forecast Summary, Low Case .............-55 4-8 Table 4.4: Railbelt Residential Forecast Summary, MIDDLE Case ............ 4-9 Table 4.5: Railbelt Residential Forecast Summary, HIGH Case .............. 4-10 Table 4.6: Railbelt Commercial Forecast Summary, LOW Case ...........--- 4-19 Table 4.7: Railbelt Commercial Forecast Summary, MIDDLE Case ........... 4-20 Table 4.8: Railbelt Commercial Forecast Summary, HIGH Case ............. 4-21 Table 4.9: 1987 Railbelt Industrial Electricity Use .............0000eeccees 4-25 Table 4.10: Projected Industrial Demand by Region and Case ..............- 4-27 Table 4.11: Street light, Office, and Loss factors by Utility ..............00- 4-28 Table 4.12: Electric Demand Growth Rate Summary ............000e00eee 4-34 Table 4.13: Effect of Gas Penetration Assumption on Sales, Relative to MIDDLE ase reel srele ele ieleaelel ike keto rouete ecole teloeebcuetcuerorsiatele oleate teers 4-35 Table 4.14: Econometric Results vs ISER Results .......... 200 e ee eee eens 4-40 Table 4.15: Synthetic Regressions on Forecast Results ...........0eeeee eee 4-41 EXECUTIVE SUMMARY There is a 66% chance that Railbelt demand for utility-supplied electric energy will grow at an average annual rate of between .2 and 1.3% during the period 1988-2010. There is a one in six chance that annual demand growth will be less than .2%, and a one in six chance that it will exceed 1.3%. Figure I shows total projected energy requirements for representative LOW, MIDDLE, and HIGH cases which have probabilities of being exceeded of 85, 50, and 15% respectively. Railbelt Total Energy Requirements Low, Middle, High Cases GWh 5000 4000} 3000} 2000 1000 | HAUL | Tee eee 1980 1985 1987 1990 1995 2000 2005 2010 — Low —~Middle —High Actual Figure I: Railbelt Total Energy Requirements These probabilities are a function of critical assumptions about the future levels of: e Households and employment Energy prices Consumer discount rates Technological change Natural Gas availability The end use structure of demand does not vary noticeably among the LOW, MIDDLE, and HIGH cases and does not change markedly during the forecast period. Figure II shows the structure of residential and commercial demand in 1987 and in 2010 under the three cases. Table I presents projected Railbelt demand by customer class and end use for the MIDDLE case projection. ES - 1 Railbelt Residential Sales, 2010 Low, Middle, High Cases GWh 2000 1500} 1000; | a ———<—— P —_— — 1987 Sales 2010 (Low) 2010 (Mid) 2010 (Hi) MMM Hest WATR (J FRIG FREZ COOK (J pry LITE {i misc Railbelt Commercial Sales, 2010 Low, Middle, High Cases GWh 2500 2000 1500 1000 500 0 1987 Sales 2010 (Low) 2010 (Mid) 2010 (Hi) MMM HEAT COOL [) vent WATR COOK [_] REFR [ MISC | Figure Il: End Use Structure of Year 2010 Demand, Three Cases ES - 2 PROJECTED RAILBELT ELECTRICITY REQUIREMENTS MIDDLE Case (GWh) Class / End Use 1987 1990 1995 2000 2005 2010 RESIDENTIAL Heat 249 222 LOS 180 185 201 Water 195 185 176 173) 182 199 Frig 154 148 147 150 156 172 Freez 82 81 83 84 86 93 Cook 62 62 67 71 77 87 Dry 99 98 102 106 Lis 130 Lite 164 164 175 194 221 258 Misc 275 273) 291; 327 379 448 RESIDENTIAL TOTAL 1281 1234 1236 1285 1401 1588 COMMERCIAL Space Heat - 48 46 45 44 38 41 Cooling 66 66 64 67 67 74 Ventilation 218 215) 212 213 220 242 Water Heat 33 34 35 38 40 44 Cooking 40 41 46 33 59 66 Refrigeration 182 179 185 oT 197 217 Lighting 724 706 698 690 728 796 Miscellaneous 175 182 216 259 304 358 COMMERCIAL TOTAL 1485 1469 1501 1555 1653 1837 INDUSTRIAL TOTAL 256 244 252 263 270 278 STREET LIGHTS/ PUBLIC AUTHORITY/ OFFICE USE/ DIST'N LOSSES 285 280 285 296 318 355 TOTAL REQUIREMENTS 3305 3225 3271 3395 3641 4053 PEAK LOAD (MW) 586 571 579 601 645 718 A EM EP ELLE SLL ELT ET aT EET, LEE EIRENE OD EGA CARN EG He CTR oN aE ec i E.R EN Table I: Railbelt Energy Requirements by End Use, MIDDLE case ES - 3 The variation in year 2010 sales across the three cases measures 23% of the MIDDLE case level and is attributable mainly to differences in the number of households and the size of the commercial floorstock. In all three cases, demand growth occurs in two relatively distinct phases: Phase 1 (1988 through ~ 1995). During this period demand is flat or falling because of three mutually reinforcing factors present in the LOW, MIDDLE, and HIGH cases: e« Employment and households are growing slowly (between 0 and 1% annually); e Electricity prices are rising rapidly (~2% annually); e Federal appliance efficiency standards and normal appliance turnover are reducing the stock average electricity use per appliance for residential water heaters, refrigerators, and freezers by ~1% per year. e The natural gas distribution system reaches Homer, Big Lake, and several other parts of the Matanuska-Susitna region, further depressing the demand for electricity’. Phase 2 (~1995 through 2010). During this period demand grows at between .7 and 1.7% annually due to the following factors present in the LOW, MIDDLE, and HIGH cases: e« Employment, households, and commercial floorstock are growing at 1.2-2.4%, about twice the annual rate of the early 1990s; e Electricity prices are essentially flat, growing at an annual rate of ~.3% Under the assumption of constant future load factors, peak load grows at the same rate as energy requirements. Figure IV shows Railbelt peak load and energy requirements by customer class; Figures V and VI display this information by region.” As Figure V shows, regional demand follows the same general growth pattern under the low, medium and high case assumptions. Demand on the Kenai is subject to a greater range of uncertainty because of the volatility of industrial load which accounts for about 30% of sales in the region. A summary of the annual growth rates of key variables for the Railbelt and its four regions under all three cases is presented in Table II. Since peak load grows at the same rate as energy requirements, its growth may be read off column 1 of this table. ‘The gas system does not expand in all of the 72 cases considered in this analysis, but each of the three representative cases chosen from the overall probability distribution contains this event. *The MatSu region corresponds to the MatSu Borough, as opposed to the MEA service territory. >We define the Kenai region as the sum of the Homer Electric and Seward Electric service territories. This definition corresponds very closely with the boundaries of the Kenai Peninsula Borough. For intertie modeling purposes, it may be appropriate to treat the service territory of Chugach Electric from Indian south as part of the Kenai load center. This retail load (Indian, Girdwood, Portage, Whittier, Hope, Cooper Landing) amounted to 34.6 GWh in 1987, or 2.0% of Anchorage region sales. CEA retail load south of Indian can therefore be reallocated from the ISER Anchorage region to the ISER Kenai region by shifting 2% of forecast Anchorage sales to forecast Kenai sales during all forecast years. ES - 4 Railbelt Total Energy Requirements Low, Middle, High Cases GWh 5000 4000; 3000; 2000 1000 go Mee 1980 1985 1987 1990 1995 2000 2005 2010 — Low Middle High Actual Railbelt Energy Requirements by Class Middle Case GWh 5000 4000 3000 2000 1000F oe a 1987 1990 Figure II: Railbelt Peak Load and Energy Reqts by Customer Class, MIDDLE case ES - 5 9 Sa. AI 21ns1y ase pue uolsoy Aq peoy yeog pue sjuswosmbey ASi0uq Total Energy and Peak Load, Anchorage Low, Middle, High Cases Energy, GWh +300 1500} 4200 1000+ 500} 4 100 a et epee e NepeoRE A poe eg 1987 1990 1995 2000 2005 2010 Low ——Middle — High Total Energy and Peak Load, Kenai Low, Middle, High Cases Energy, GWh Peak, MW 600 500 400 300 200 100 Oa eee ee 1987 1990 1995 2000 2005 2010 —Low —~Middle —*High Total Energy and Peak Load, Fairbanks Low, Middle, High Cases Energy, GWh Peak, MW 1000 800 150 600 400 400 50 200 lic Rihnchleklh cd Bind bdo bacliak ° 1987 1990 1995 2000 2005 2010 —Low ——Middle High Total Energy and Peak Load, MatSu Low, Middle, High Cases Energy, GWh Peak, MW 400 300) 200 100 oO 1987 1990 1995 2000 2005 2010 —+ Low ——Middle — High L- Sd ISPD AIAAINW ‘SSID JowoisnD pue uowoey Aq sjuourosnboy ASI0uq :A oINSLy Anchorage Energy Requirements Middle Case GWh 3000 2500 « WY 2000 OOK KOMI 1500 1000 500 (9 Ces 1987 1990 1995 2000 2005 CJres Z2Acomm ([_Jjinn WWSorHer Kenai Energy Requirements Middle Case GWh ey 1987 1990 1995 2000 2006 2010 Cres ZAcomm [Jinn SQortHER Fairbanks Energy Requirements Middle Case GWh WS — CO 600 RA aA Ce ee a ee 1987 1990 1995 2000 2005 fa ey CJres LAcomm [_jinn KWSorter MatSu Energy Requirements Middle Case GWh SS WS SWISS 50 ° gS peep ola peg pe Se 1987 1990 1995 2000 2005 (“lres LAcomm [_jinn SWorter Total |-------- Residential ------- [|-oreere- Commercial -------- | Electric Electric Occupied Retail Electric Total Retail Energy sales housing electric sales floor electric Industrial Requiremts units price stock price Sales RAILBELT 1987-1995 LOW -0.6% -0.8% 0.5% 1.7% 0.1% 0.7% 0.5% “4.9% MIDDLE 0.0% -0.5% 0.8% 2.3% 0.1% 0.7% 1.2% -0.2% HIGH 0.5% -0.2% 1.2% 2.8% 0.7% 1.4% 1.8% 2.5% 1995-2010 LOW 0.7% 0.8% 1.4% 0.3% 0.7% 1.1% -0.2% 0.2% MIDDLE 1.4% 1.7% 2.2% 0.5% 1.4% 1.9% 0.0% 0.6% HIGH 1.7% 2.0% 2.5% 0.8% 1.6% 2.2% 0.3% 1.3% 1987-2010 LOW 0.2% 0.2% 1.1% 0.8% 0.5% 1.0% 0.0% -1.6% MIDDLE 0.9% 0.9% 1.7% 1.1% 0.9% 1.5% 0.4% 0.4% HIGH 1.3% 1.2% 2.1% 1.5% 1.3% 1.9% 0.8% 1.7% ANCHORAGE 1987-1995 LOW -0.3% -0.7% 0.3% 2.3% 0.0% 0.5% 0.3% 0.7% MIDDLE 0.0% -0.3% 0.7% 3.1% 0.1% 0.6% 1.1% 1.3% HIGH 0.5% 0.0% 1.1% 3.6% 0.7% 1.3% 1.7% 6.0% 1995-2010 LOW 0.6% 0.7% 1.4% 0.5% 0.7% 1.1% -0.4% 0.0% MIDDLE 1.5% 1.6% 2.3% 0.8% 1.4% 2.0% -0.1% 1.0% HIGH VeTs 1.9% 2.6% 1.1% 1.6% 2.3% 0.3% 1.4% 1987-2010 LOW 0.3% 0.2% 1.0% 1.1% 0.4% 0.9% -0.2% 0.2% MIDDLE 1.0% 1.0% 1.7% 1.6% 1.0% 1.5% 0.3% 1.1% HIGH 1.3% 1.2% 2.1% 2.0% 1.3% 1.9% 0.8% 3.0% FAIRBANKS 1987-1995 LOW 0.8% 1.0% 1.0% 0.1% 0.6% 1.2% 0.1% 0.8% MIDDLE 0.7% 1.1% 1.1% 0.8% 0.1% 0.9% 0.8% 0.9% HIGH 1.5% 1.4% 1.5% 1.2% 0.8% 1.6% 1.3% 4.6% \ 1995-2019 LOW 0.8% 1.2% 1.5% -0.1% 0.7% 1.2% -0.1% 0.0% MIDDLE 1.3% 1.7% 1.9% 0.1% 1.1% 1.6% 0.2% 0.7% HIGH 1.7% 2.1% 2.4% 0.4% 1.4% 2.0% 0.4% 1.3% 1987-2010 LOW 0.8% 1.1% 1.4% 0.0% 0.6% 1.2% 0.0% 0.3% MIDDLE 1.1% 1.5% 1.6% 0.4% 0.7% 1.4% 0.4% 1.3% HIGH 1.6% 1.9% 2.1% 0.7% 1.2% 1.9% 0.8% 2.4% KENAI 1987-1995 LOW -3.1% -1.0% 0.6% 1.3% 0.6% 1.2% 1.4% -10.3% MIDDLE -0.7% -0.8% 0.8% 1.6% 0.6% 1.1% 1.8% -1.3% HIGH 0.0% -0.7% 0.9% 1.8% 0.9% 1.5% 2.0% 0.4% 1995-2010 LOW 0.6% 0.7% 1.2% -0.0% 0.7% 1.1% 0.0% 0.4% MIDDLE 1.0% 1.4% 1.7% 0.1% 1.0% 1.7% 0.1% 0.4% HIGH 1.3% 1.6% 2.0% 0.3% 1.1% 1.8% 0.3% 1.1% 1987-2010 LOW -0.7% 0.1% 1.0% 0.4% 0.6% 1.2% 0.5% -3.4% MIDDLE 0.4% 0.6% 1.4% 0.6% 0.9% 1.5% 0.7% -0.2% HIGH 0.8% 0.8% 1.6% 0.8% 1.0% 1.7% 0.9% 0.8% MATSU 1987-1995 LOW -2.1% Sake 0.5% 1.9% 0.2% 0.7% 1.9% 9.6% MIDDLE -1.8% -2.8% 0.8% 2.3% 0.2% 0.7% 2.3% 15.3% HIGH “1.3% -2.3% 1.3% 2.6% 0.7% 1.5% 2.6% 15.3% 1995-2010 LOW 0.8% 0.7% 1.6% 0.1% 0.8% 1.4% 0.1% 3.5% MIDDLE 2.0% 2.1% 2.6% 0.2% 1.6% 2.6% 0.2% 3.9% HIGH 2.3% 2.4% 2.9% 0.4% 1.8% 2.7% 0.4% 7.0% 1987-2010 LOW -0.2% -0.6% 1.2% 0.7% 0.6% 1.2% 0.7% 5.5% MIDDLE 0.7% 0.4% 2.0% 1.0% 1.1% 1.9% 1.0% 7.7% HIGH 1.0% 0.7% 2.3% 1.2% 1.4% 2.3% 1.2% 9.8% Table II: Electric Demand Growth Rate Summary ES - 8 Technical Conventions The following conventions are employed in this report. They are generally introduced in the text. Dollars All dollars are 1987 real (adjusted for inflation). Units and Unit Prefixes Btu British thermal unit Ft2 Square foot G Giga = 10° GW __ Gigawatt = 10°W = 1 million kW = 1 thousand MW GWh_ Gigawatt-hour. Convenient unit of electric energy, = 1000 MWh h hour k thousand kW kilowatt = 1000 W, = ~.7 Horsepower. A unit of power. kWh _ kilowatt hour, = 3413 Btu kV kilovolt. 115 kV is the dividing line between distribution and transmission. M million (except applied to natural gas where 1 MCF = 1 thousand cubic feet) MW Megawatt = 10°W WwW watt = 1 Joule/second. Fundamental unit of electric power. Vv Volt yr year Jargon ach air changes per hour AFUE average fuel use efficiency. coincidence a factor between zero and one which measures the ratio of the combined peak load of two utilities (at the time such combined load is at a maximum) to the sum of their individual system peaks which may occur at different times. EI Energy Intensity, = energy consumption of a given fuel per Ft2 for a building or end use combination, generally used for electricity, as kWh/Ft2/yr. When summarizing a building population, equal to EUI * Market share. EUI _ Energy Use Index = energy consumption per Ft2 for an end use/fuel combination. Only electric EUIs are discussed in the text, measured in kWh/Ft2/yr. EF Energy Factor. A measure of energy efficiency whose units vary with the appliance being described. For refrigerators and freezers units are Ft3-Day/kWh. For heating equipment, EF is expressed as average fuel use efficiency of AFUE. SWEF Shipment-Weighted Energy Factor. The weighted average EF for a stock of appliances where the weights are the number purchased of each different EF value. 1. INTRODUCTION 1.1 Purpose and Structure This report presents end use forecasts of utility-supplied’ electric energy and peak power demand for the Alaska Railbelt and its four constituent regions: the Anchorage Borough, the Fairbanks and Southeast Fairbanks Census Areas, the Kenai Peninsula Borough’, and the Matanuska-Susitna Borough.’ It is one of several analyses funded by the Alaska Power Authority as part of its intertie feasibility assessment. Our characterization of today’s Railbelt electricity market is based in large part on one residential and two commercial sector end use surveys which we conducted during the winter of 1988. This report also makes direct and substantial use of economic and demographic projections recently completed by ISER (Goldsmith 1988) for the Power Authority. These projections have been previously circulated for public comment and are not reviewed or amended here. While several past studies have investigated the residential sector in some detail, this report presents the first end use load forecast ever published for the Railbelt commercial sector. Results include estimated Railbelt floorstock by building type, region, and vintage and of energy use per square foot by building type and end use. Many results are presented graphically both because of the volume of numbers generated and because the relative relationships among inputs and the relative changes we forecast are far more robust than any of the absolute levels. In this forecast we project the quantity of electricity demanded under the assumption that no additional State government intervention in the demand side of the energy marketplace. We call this quantity the market-driven quantity. Since our forecast includes response to both price changes and federally-mandated appliance efficiency standards, it is not a "frozen technology" analysis. Nor is it any sort of “least cost" or "technical potential" analysis. Although the end use models employed are based in part on engineering relationships, they also employ substantial economic logic about how real people behave. In particular, the models use estimated consumer--as opposed to social--discount rates. This report is organized as follows. In the remainder of this chapter we present a "snapshot" "Chugach Electric Association (CEA), Anchorage Municipal Light & Power (MLP), Golden Valley Electric Association (GVEA), Fairbanks Municipal Utility System (FMUS), Homer Electric Association (HEA), Seward Electric System (SES), and Matanuska Electric Association (MEA). We do not forecast load for the University of Alaska Fairbanks or the Military in either Anchorage or Fairbanks, all of whom are primarily self generators. *For model calibration purposes we define the Kenai region as the sum of the Homer Electric and Seward Electric service territories. This definition corresponds very closely to the boundary of the Kenai Peninsula Borough. The discrepancy arises from the settlements of Hope and Cooper landing, which are served by Chugach Electric. For intertie modeling purposes, however, it may be appropriate to treat the service territory of Chugach Electric from Indian south as part of the Kenai load center. This retail load (Indian, Girdwood, Portage, Whittier, Hope, Cooper Landing) amounted to 34.6 GWh in 1987, or 2.0% of Anchorage region sales. CEA retail load south of Indian can therefore be reallocated from the ISER Anchorage region to the ISER Kenai region by shifting 2% of forecast Anchorage sales to forecast Kenai sales during all forecast years. *This definition of the MatSu region does not correspond to the service territory of the Matanuska Electric Association, which includes the Eagle River section of the Anchorage region. A technical memorandum is available from ISER which allocates Eagle River sales to the MatSu region to estimate MEA service territory load. 1-1 This report is organized as follows. In the remainder of this chapter we present a "snapshot" of 1987 electricity demand and introduce the basic concepts underlying end use forecasting models. In chapters 2 and 3 we develop the major input assumptions which drive the residential and commercial end use models. In chapter 4 we summarize the results of the end use forecasts for representative low, middle, and high outcomes and derive corresponding total energy requirements by adding estimates of industrial and street light sales, utility office use, and distribution losses. Finally, we assess historical load factors and project peak load under the assumption of constant regional load factors. Boxed text is used to present case examples and may be skipped without loss of continuity. Several appendices provide additional support for input assumptions, output detail from the end use models, and a summary of the residential end use survey results. 1.2 1987 Railbelt Electricity Demand It is most useful to base our discussion of end use modeling concepts on current data. In 1987, the seven Railbelt utilities required ~3400 Gigawatt Hours (GWh) of electric energy to meet their customers’ needs. Figure 1.1 and Table 1.1 show this total broken down by customer class and type of system loss.’ Total Railbelt Electricity Use 1987 Uses of Utility-Supplied (Utility Supplied) Electricity 1987, Gwh Total Railbelt, Gwh Residential <a =~ % Transepsee Street Lights & 49 Drom nage Public Authorities cone hia Distribution Losses & 225 "388 Office Use Transmission Losses 82 TOTAL 3373 OTOL: OED OS PT tb PRE EI | SE, Figure 1.1: Railbelt Electricity Use Table 1.1: Railbelt Electricity Use This study focuses on how electricity is used (end uses) rather than who uses it (customer classes). We define eight end uses in each of the residential and commercial sectors (see box). Figure 1.2 shows residential and commercial sales broken down by these end uses. We developed the breakdown from engineering estimates and end use survey data. Of the residential end uses, space heat is far and away the most important for those customers actually using electric heat. However, the electric share of the Railbelt residential space heating market is currently ~ 17 percent’ and has been declining for several years. We estimate that 19 percent of total 1987 residential electric sales went for space heating. Less electricity is now used for heating than for miscellaneous appliances. On the commercial side, however, lighting is not a competitive end use with regard to fuel choice. The importance of lighting energy per user translates directly into a 49 percent share of lighting in commercial electric sales. “These figures are approximate and derived from utility sales data as follows. Estimated residential demand served through commercial class meters is moved from commercial class sales data to residential. Estimated industrial demand is broken out of commercial sales data, since it is not reported separately. Distribution losses are from utility data. (For CEA we use "retail" losses per Cost of Service data). Transmission losses are estimated as CEA’s residual generation after removing all sales and retail losses, plus 9% of MLP sales on the Anchorage-Fairbanks intertie. Son the basis of heating services provided. This is lower than the share of electric customers using electricity for heating, since many of these people use multiple heat sources. See chapter 2. 1-3 my! v asc) pug Aq soyeg onary [eoJoWWOD pur [eNUEpIsoY HOqIeY L86I :ZT FNSLy Residential and Commercial Sales by End Use, Railbelt 1987, GWh Cook Water Water 40 33 195 7 Frig 154 Cool 66 Heat Freezer 48 82 Misc 175 Lite 164 Residential 0 ial 1281 ommercia 1485 Source: Railbelt Utility Data, weather-adjusted 1.3 End Use Forecasting Models In contrast to aggregate econometric models which treat electricity as a pure commodity such as milk or concert tickets, end use models treat the demand for electric energy as a derived demand which depends explicitly on the stock of electric appliances, their inherent efficiency, and the intensity with which they are used. The models used in this study are stock-flow models of the building and appliance capital stocks coupled with behavioral equations which simulate consumer choice in response to changing economic variables. Total sales are approached from the "bottom up" by aggregating over end uses and building types. The entire energy market is modeled. Competition between electricity and other fuels is a central determinant of total electric load. End Use Model Classifications Residential Commercial End Uses End Uses HEAT : Heating HEAT : Heating WATER : Hot Water COOL : Space Cooling FRIG : Refrigeration VENT : Ventilation FREZ : Freezer WATER : Hot Water COOK : Cooking REFR : Refrigeration / Freezing DRY _ : Clothes Drying COOK : Cooking LITE : Lighting LITE: Lighting MISC’ : Miscellaneous MISC _: Miscellaneous Building Types Building Types SINGL : Single Family SMO _— : Small Office (<20,000 Ft2) MULTI: Multi Family (2+ units) LGO _ : Large Office MOBIL: Mobile Home RES _ : Restaurant LGR _ : Large retail (>20,000 Ft2) SMR _ : Small Retail GRC _ : Grocery WRH_ : Warehouse CAR _ : Auto Service LDG _ : Lodging MED: Medical SCH — : School COL _ : College? ASB _ : Assembly MISC _: Miscellaneous VGT > Vacant 'Miscellaneous includes (residential) small appliances, headbolt heaters, roof heat tape, waterbed heaters, saunas, and jacuzzis; (commercial) office equipment, vending machines, automobile heater outlets. College includes UAA but not UAF. UAF is primarily a self-generator. 1.3.1 Basic Concepts We model end use energy demand using the following terms to represent "stylized facts". These concepts grossly simplify the structure of a complex market. However, they create a consistent framework for analysis which can be populated by available data. 75) Floorstock measures the size of the market for energy services. Commercial floorstock is measured in square feet (Ft2) for 15 separate building types. Residential floorstock is measured in housing units of 3 types. Market Share is the fraction of the total market for a given end use that is actually served by a given fuel type. In the residential sector, the total market is defined to include all households and the residential market share is equal to the product of the more familiar "saturation rate" and "fuel mode split" terms.* Commercial market shares are defined relative to the total market penetration of the end use. Since this penetration is generally 100%,’ commercial market share is generally analogous to fuel mode split. Not surprisingly, the importance of market share for a given end use/building type combination in determining total sales depends on the end use’s importance in building sales and the building type’s importance in total sales (see box). Energy Use Index (EUI) is a measure of consumption per appliance per year for a given end use and fuel. In the commercial sector we measure EUIs as energy per Ft2 of floorspace. EUI values are fuel-specific and end use- specific. For residences, electric EUI is expressed as kWh per appliance. For lights and miscellaneous appliances, electric EUI is measured on a per- household basis. Energy Intensity (EI) is a summary measure of energy consumption of a particular fuel per unit of floorstock. It is defined to be the product of market share and EUI. An overall EI can also be computed as the sum of Els for all Market Share, EUI and EI: An Example Suppose there are 100 houses in the Railbelt and that 40 of them have electric water heaters, 50 have gas water heaters, and 10 have no water heaters. Suppose that each electric water heater uses 5000 kWh/yr, while each gas water heater consumes 20 Mcf of gas per year. Then: © The electric water heat EUI equals 5000 kWh/house/yr ¢ The gas water heat EUI equals 20 Mcf/yr ¢ The electric market share of the water heat market = 40/100 = 40% ¢ The gas market share of the water heat market = 50/100 = 50% © The electric water heat energy intensity (EI) = (40%) * 5000 = 2000 kWh/house/yr © The gas water heat (EI) = (50%) * 20 = 10 Mcf/house/yr end uses. Overall Intensity is a useful concept because it can be derived from a sample of existing data: one simply divides the sum of billed kWh by the sum of all the of Ft2 served by the meters in the sample. Residential EI is equivalent to the familiar concept of use per customer." Utilization Rate is a measure of the frequency and duration of equipment use normalized to equal 1.0 in 1987. Utilization rate can change due to both lifestyle °In general, market share collapses to one or the other of these concepts: In noncompetitive end uses (lights, frig, etc.) mode split =1 and share equals saturation, while in competitive uses (heating, cooking, etc.), saturation generally equals 1 and share equals mode split. In some end uses, such as clothes dryers, our concept of market share corresponds to neither saturation nor fuel mode split. But it is always equal to the product of the two. It is modelled this way to allow direct computation of consumer movements between fuels and between ownership and nonownership of an appliance. "The exception being air conditioning. ®Interpretation of residential use per customer estimates computed from reported utility sales data is complicated by Alaska’s current high rate of residential vacancy. 1-6 changes such as thermostat setback and technical fixes such as thermostat timers. Conceptually, utilization embodies all changes in consumption during the time period when the appliance stock is fixed. In practice, it is often impossible to draw a clear line between utilization changes and equipment changes. The concepts of Market share, EUI, and EI have both stock and flow interpretations which are crucial to the evolution of the energy market. At any moment, the stock average values of these terms describe the structure of the market and lead to the level of total sales. All three terms are also used to measure the attributes of the flow of new capital (houses, buildings, and equipment) into the energy market. When used to describe the attributes of new equipment, the terms are known as marginal shares, marginal EUIs, and marginal Els. For example, the average electric share of the Anchorage clothes dryer stock is 71%. This measures what people own (a stock concept). The initial marginal share used in our forecast is 68%. This is an estimate of what people are buying (a flow concept). The difference between average and marginal shares and EUIs is generally the chief force acting on electric sales per customer. As every baseball player knows, if current performance falls short of average performance, the average will soon fall. Similarly, if no one builds electrically heated houses for 20 years, average electric heat market share will be pulled down toward zero.’ 1.3.2 Central Energy Equation The terms just described are combined to form a central energy equation which is at the core of every end use model: Sales;,, = Floorstock,, * Electric Market Share,, (1.1a) * Energy Use Index,., * Utilization Rate, where This equation says that in year n, sales of fuel i to building type j for end use k are equal to the type j floorstock times the fuel i market share for end use k times the use per appliance times the utilization rate of end use k appliances using fuel i. (1.1a) can be rewritten in terms of electric intensity EI by substituting the definition of EI for the terms (market share) * EUI: Sales, = Building Stock, * Energy Intensity, * Utilization Rate, (1.1b) (kWh) (Ft2) (kWh / Ft2) Once sales by building type, end use, and fuel are computed, total sales are simply aggregations over the appropriate index variable. For example, if electricity is indexed by ds 5 i cae F Provided, of course, that there is some new construction occurring in the overall housing market. 1-7 Once sales by building type, end use, and fuel are computed, total sales are simply aggregations over the appropriate index variable. For example, if electricity is indexed by i=1 and single family homes by j=1, then electric sales to single family homes are written: Sales,,, = > Sales... (1.2) k The subscripts in the central energy equation point up the level of detail maintained by end use models. This detail is the chief advantage of the end use framework. Electricity use can change because of shifts in: Floorstock type distribution. Grocery stores use 5-10 times the electricity of warehouses. Market share within a building type. The de facto moratorium on electric heat in Fairbanks has caused a marked drop in measured use per customer during the last decade. Technical efficiency (EUI) of specific end use equipment. Federal law mandates a ~20 percent reduction in kWh per new refrigerator starting in 1991. Utilization rate of equipment. The Anchorage School District cut its ventilation end use electric consumption by >20% by using energy management systems to "put its buildings to bed at night." Central Energy Equation: Example Consider the use of headbolt heaters by Fairbanks residential customers. There are ~25,000 customers. End use survey data indicate that 31% of all customers have 1 heater and an additional 39% have 2. The total market share is Share = 31 + (2*.39) = 1.09 On average, every house has 1.09 heaters. The EUI is a function of weather; under the assumptions of table 2.9, calculated EUI is 554 kWh per heater per year. Assume Utilization = 1.0. Then, total sales to the residential sector for the headbolt heater end use are: SALES = 25,000 * 1.09 * 554 * 1.0 = 15.1 GWh/yr floor- share EUI Util- stock ization In this example the energy intensity (EI) = 1.09 * 554 = 604 kWh/household. Notice how EUI is a technological parameter, while EI is a summary Statistic since nobody has exactly 1.09 heaters. Through the central energy equation, end use models distinguish between these effects and keep separate track of their influence on total electric sales. 1.3.3 Model Logic All end use models use some sort of central energy equation. In this sense they are accounting machines. The models used in this study” are significantly more sophisticated than this, however. In both of them, the energy equation is a final step of arithmetic. The bulk of the models’ algorthims are devoted to determining the components of (1.1a): market share, technical efficiency (EU), and utilization rate. Only floorstock is an exogenous input. °The residential model is an enhanced version of the Lawrence Berkeley Labs Residential Energy Model. It has been re-christened the Alaska Residential Energy Model (AKREM). AKREM is written in FORTRAN 77, runs on an IBM-compatible microcomputer, requires a single ASCII input file, is in the public domain, and is available without charge from ISER. The commercial end use model is EPRI's COMMEND-PC 3.0. COMMEND is written in C, runs on an IBM/AT-compatible microcomputer, is copyrighted by EPRI, and is available free to EPRI members or for a license fee to others. 1-8 The Importance of Market Share in Total Sales is a function of several factors, as shown by the central | energy equation. Neglecting Utilization, we can rewrite (1.1a) for a particular fuel (electricity) as: Gy =F,*S,*E, ; (k=end use, j=bldg type) (1) where G =sales, F = floorstock, $= market share, and E = EUI. The following propositions follow from either calculus or common sense: e First, for any fixed k and j, dG/dS = F*E. The sensitivity of changes in sales to changes in market share is a function of EUI level. Changes in electric heating shares are more important than changes in clothes drying shares. ¢ Second, for k,j fixed, the elasticity of G with respect to S is fixed at one. That is, a given relative change in S is far more important if the level of S is high to start with. Uncertainty in estimates of the gas cooling share is unimportant because the level of the share is close to zero. ¢ Third, consider aggregation to building total sales figures: G, = 2 Gy (2) dG,/dS, = E (dG,/dS,.)*(S;/G,) = Ex * (Sy/G;) = (E,*S,) / G; = G,/G; (3) That is, the elasticity of building sales with respect to one end use’s market share equals the end use’s share of building sales. ¢ Finally, consider aggregating across building types to arrive at total sales figures: G=35*G, (4) Algebra similar to (3) quickly shows that the elasticity of total sales with respect to market share of one end use in one building type is equal to: (dG/dS;)*(Six/G) = (F\/F)*(G,/G) (5) That is, the importance of a single market share to total sales is a function of the end use’s share in building sales times the building’s share in total sales. In computing these components, the end use model subroutines function as simulation models with both engineering and behavioral aspects. A very rough sketch of the algorithms looks like this: ¢ Market share (%) for each fuel/building type/end use combination is determined by evaluating a set of econometric choice equations. These equations are initially calibrated to reproduce initial estimates of actual marginal market shares (what people are buying and building). Marginal market shares evolve throughout the forecast in response to levelized operating and capital costs of available appliances as well as income. These life-cycle costs depend in turn on expected fuel prices, discount rates, and the technical efficiency of the equipment. "Detailed descriptions of model logic may be found in (residential) Hirst 1978 and (commercial) McMenamin 1988??, McMenamin 1987. 1-9 EUIs (kWh/Ft2 or kWh/appliance) are essentially determined by minimizing the life cycle cost of providing the end use service. This is accomplished by choosing an optimal point on an engineering curve which relates equipment cost to energy use. The optimal point is (partly) a function of the prevailing consumer discount rate. These discount rates are input to the model based on empirical obervation of past equipment purchases (Ruderman 1984). The high level of these disount rates attests to the fact that considerations other than life cycle cost minimization play a large role in consumer decisionmaking. For some commercial end uses, intertia factors also are used to model customer resistance to change. Marginal EUIs evolve throughout the forecast in response to changing prices and technological possibilities. Average EUIs are determined by tracking the flow of appliance retirements and purchases. Utilization Rate (%) is calculated by applying short-run elasticity coefficients to changes in electricity price. ‘ Thermal Interaction algorithms account for the effects of an increase in the load of one end use on the load of another. For example, interaction coefficients adjust the heat load upward (or the cooling load downward) if the lighting load drops. These interactions are only modelled in the commercial sector. Vintage effects are tracked and accounted for. Equation (1.1) yields correct sales calculations only when its components reflect stock average values for market shares and EUIs. When new equipment EUI and market share values differ from stock average values, the appliance stock age distribution and rate of turnover can be an important determinant of total energy sales. Our end use models are initialized with regional age distributions for buildings and appliances derived from the end use surveys. Together with retirement functions, these initial age distributions determine the flow of retiring units out of the equipment stock. As new equipment replaces old, the new EUI and market share values are "rolled in" to the averaging process, and the equipment age distribution is updated. 1.4 Treatment of Uncertainty 1.4.1 Use of Probability Tree Analysis Alaska’s volatile economy, energy price uncertainty, rapid technological change, and an immature gas distribution system all make the future level of electricity demand quite uncertain. We address this uncertainty explicitly through the use of a probability tree analysis. Our approach follows that used by Goldsmith (1988, chapter 3) to develop employment/household forecasts for input to this study. In consultation with the Power Authority, we developed alternatives for the following sets of critical assumptions assumed to affect demand: H: households and employment (3 alternatives) P: energy prices (3 alternatives) D: consumer preferences (discount rates) (2 alternatives) T: technological change (2 alternatives) G: natural gas market penetration (2 alternatives) 1-10 1 i i mL ei i a Refrigerator Age Distributio Refrigerators Anchorage Region The efficiency (measured in refrigerated Yee" Of Purcnase ° PS ——_ioO cubic foot-days per kWh, also known as 7 ae Energy Factor) of new refrigerators 06 improved at an average annual rate of 4.5 re percent between 1972 and 1986, for a 1980 ete ‘Average '= 1980 total gain of 80% (relative to 1972). Hence the current stock average efficiency is a strong function of the age distribution. By applying a_ historical vector of efficiency levels to the age 1970 distribution of Anchorage refrigerators (right), we calculated the average efficiency of the 1987 stock to be 5.6 Ft3- Bes TCI GTa Tos NoCoe TIGR CLS day per kWh. This implies that Fraction of Stock Anchorage refrigerators consume about 93 Gwh of electric energy per year = 13 percent of residential consumption. 1975 1965 ° Suppose, however, the average year of purchase had been 1974 instead of 1980. Then the stock average efficiency would be only 4.55 and total 1987 consumption would be 23% higher at 112 Gwh. Federal Standards mandate that refrigerators sold beginning in 1991 must have an EF of 7.5. This implies the appearance of a 30% gap between new units and stock average units. How fast will the new efficiency level pull up the average level? The answer depends on the rate of retirement of existing stock (and on the share of new housing units in total housing units.) However, this linkage between age structure and speed of stock turnover is less significant than that between age structure and initial stock average efficiency. To continue the example above, suppose once again that the average year of purchase had been 1974. Holding the initial efficiency level constant, the effect of the faster turnover caused by an older stock is to reduce year 2000 consumption from 93 to 90 Gwh - a 3% decline. Because end use models keep explicit track of the appliance age distribution, they are useful for exploring the effects of appliance age structure on total sales. nnn eeeeeeeeeeeeeeeeeeeeeeeeeeeeeOEOEOEeEeEeEeEeEeEeEeEO | These alternatives were combined to form a projection tree describing all possible combinations of assumptions about H,P,D,T, and G. Conditional probabilities were assigned to each branch of the tree. Denoting a specific combination of particular assumptions by the lowercase bundle (h,p,d,tg), we can write the probability that any such particular combination will come to pass as: Prob(h,p,d,t,g) = Prob(h) * Prob(p, given h) * Prob(d, given p and h) * Prob(t, given p and h and d) * Prob(g, given p and h and d and t) (1.3) There are (3*3*2*2*2)=72 possible combinations (h,p,d,t,g). We computed regional load forecasts under a representative sample of these 72 "states of the world" and interpolated the rest. By assigning probabilities to each combination of assumptions according to (1.3), UES 6 The differences between critical assumptions are summarized at the start of chapter four. In chapters two and three, as model inputs are reviewed, we refer to these critical assumptions where appropriate. It is important to keep in mind that neither the low, middle, and high employment/household projections nor the low, middle, and high energy price assumptions are necessarily linked with the LOW, MIDDLE, and HIGH load forecast cases. While it is true that high employment growth causes higher demand, there are several ways in which high demand can be realized. We try to keep the distinction between the input projection cases and the results projection cases clear by reserving capital letters for the results of this study. 1.4.2. Risk Uncertainty, "True" Uncertainty, and Conservatism Frank Knight’s (1921) classic distinction between risk uncertainty and true uncertainty applies with some force to a load forecasting exercise. Essentially, Knight defined "risk uncertainty" as the inability to know in advance the realized value of a random variable” about which we do know the underlying probability density. A good example of this is our inability to know whether a fair coin will come up heads or tails upon being tossed. In contrast to this notion, Knight suggested that true uncertainty is our inherent inability to comprehend the very nature and type of future events, let alone their probabilities. Our probability tree does an excellent job of dealing with risk uncertainty. It performs less well in handling Knight’s "true uncertainty". We may have missed an important set of branches altogether, and even if all the branches are correct the probabilities may be wrong, since the price of oil is not half as well behaved as a fair coin or pair of dice. Of course, all load forecasting methods are subject to these limitations, not just end use models. A pure econometric model would fail completely to account for the effects of federal efficiency standards, while an end use model is ill-equipped to handle major new end uses that might be captured by an econometric income elasticity term. In light of the fundamental, "true" uncertainty about future demand growth which remains after the most carefully specified probability tree has been constructed, we follow the practice of employing conservatisms at several stages of the analysis. As used here, the word "conservatism" denotes a deliberate bias of unknown but generally small size which is adopted because one believes that the costs of being wrong-on one side of the actual outcome exceed the costs of being wrong on the other. In this study, the conservatisms we employ are biases toward higher load projections. This is the conventional practice in the utility industry, which has a legal requirement to serve its customers and a social obligation to keep the lights on. The term random variable is used per mathematical definition: a mapping from possible states of the world to the real number line. 1-112) 2. RESIDENTIAL INPUT ASSUMPTIONS 2.1 Housing Stock Residential energy use is largely determined by the size and composition of the residential housing stock. A model which estimates housing stock based upon the outputs of Goldsmith’s (1988) projections of employment and households was constructed to project housing stock under three employment/households scenarios: Low (85% chance of being exceeded), Middle (50%), and High (15% chance of being exceeded). 2.1.1 Current Housing Stock The composition of the residential housing stock for 1988 is presented in Table 2.1. These estimates are based on the best information available in the summer of 1988, but are subject to some error due to the transitional nature of the Railbelt housing market at the time. RAILBELT RESIDENTIAL HOUSING STOCK Thousands in mid 1988 SINGLE MULTI MOBILE TOTAL SECOND FAMILY FAMILY HOME HOMES ANCHORAGE occupied 36.317 = 35.314 5.680 77.311 vacant 2.734 7.286 1.420 11.440 total 39.051 42.600 7.100 88.751 0.506 GREATER occupied 15.511 6.713 2.276 24.500 FAIRBANKS vacant 0.990 2.391 0.569 3.950 total 16.501 9.104 2.845 28.450 0.195 KENAI occupied 9.313 3.086 2.301 14.700 PENINSULA vacant 0.594 0.110 0.575 1.279 total 9.907 3.196 2.876 15.979 2.438 MATANUSKA = occupied 9.677 1.227 1.225 12.129 SUSI TNA vacant 2.419 0.815 0.346 3.580 total 12.096 2.042 1.571 15.709 4.129 TOTAL occupied 70.818 46.340 11.482 128.640 vacant 6.737 10.602 2.910 20.249 total 77.555 56.942 14.392 148.889 7.268 These housing unit figures have been adjusted for consistency with electric utility customer records and consequently do not conform to the actual housing stock counts within each region. Greater Fairbanks includes the Fairbanks Borough and the Southeast Fairbanks Census Area. Matanuska Susitna (MatSu) refers to the Borough, not including Eagle River. Calculations to allocate Eagle River variables to MEA service territory are available from ISER. Table 2.1: Railbelt Residential Housing Stock 2-1 2.1.2 Housing Stock Projections The stock of residential housing units changes over time due to demolitions and new additions, both of which occur at different rates for the different types of units. In the near term the stock of units falls because new additions are essentially zero in the presence of excess supply in all markets. As additional households increase demand, the existing "excess" vacant stock is occupied before new housing units are constructed. "Excess" vacancies are defined as the increment above the long-term equilibrium vacancy rate. This equilibrium rate accounts for the normal turnover of the occupied housing stock which is relatively high in the Railbelt due to the mobility of the population. Vacancies .are disproportionately concentrated in the less desirable housing types--the multifamily and mobile home units. New construction occurs only after all excess vacancies have been eliminated, either through demolitions or growth in demand. Once construction starts, we assume that a specified composition of housing types will be maintained. However, our model allows for the marginal composition to be different from the 1987 average, thus introducing a source of change in the composition. The housing stock begins to increase at different times in different markets and under different assumptions about economic growth, but in general it is several years before there are significant additions to the housing stock with the exception of second homes. The stock of second homes is assumed not to have excess vacancies and to increase as a function of population. Figure 2.1 displays projected occupied housing units by type while Table 2.2 shows occupied housing stock growth rates. During the projection period there is little change in the composition of the total (physical) housing stock. However, the composition of the occupied stock changes while vacant units are filled up at differential rates (see box). Housing Market Dynamics are clearly shown in the MatSu panel of Figure 2.2. The white bar depicts multifamily homes. Between 1987 and 1990, there is a “flight to quality" as people take advantage of the high vacancy rate and low prices among single family homes. As the figure shows, the number of single family households increases even as the total number of households falls. Between 1990 and 1995, total demand for housing grows. The flight to quality stops because we have assumed that a core group of multifamily households remains. Occupied single family stock grows until 1995, at which time single family stock is fully occupied. Between 1995 and 2000, the large number of excess multifamily vacancies soaks up all increases in demand. By our assumption, no new construction is called forth while excess vacancy remains in the multifamily stock. Therefore, the natural process of demolition causes the occupied single family stock to decline slightly. Finally, after year 2000, all excess vacancy has been eliminated and new construction begins in response to further increases in demand. The occupied stock then grows smoothly with housing types built in fixed proportions. We have assumed for simplicity that throughout this entire housing stock scenario involving single- and multifamily houses and households, the mobile home population is an essentially separate market segment. yooisg BuisnoyY powoloig :[°7 sms Anchorage Occupied Housing middle case thousands 80 60 40 20 our 1987 1990 1995 CJ singte-Famity — [_] Multi-Family Mobile Homes Kenai Occupied Housing middle case thousands SS eg 1990 > 1995 2000 CJ singte-Famity —[_] Multi-Family Mobile Homes Fairbanks Occupied Housing middle case thousands o bat aot 1987 1990 1995 2000 2005 (J singte-Famity ([_] Multi-Family Mobile Homes MatSu Occupied Housing middle case thousands a Ses ger pe peg tg oO 1987 1990 1995 2000 2005 C_] singte-Famity (_] Multi-Family Mobile Homes ia tel a a PROJECTED OCCUPIED HOUSING STOCK GROWTH RATES (Annual Rate) Single Multi Mobile Family Family Home ANCHORAGE 1987-1995 Low 0.0% 0.6% 1.1% Middle 0.0% 1.4% 1.1% High 0.5% 1.6% 1.3% 1995-2010 Low 1.5% 1.4% 0.7% Middle 2.8% 1.9% 1.5% High 3.2% 2.2% 1.7% 1987-2010 Low 1.0% 1.1% 0.9% Middle 1.8% 1.7% 1.3% High 2.2% 2.0% 1.6% FAIRBANKS 1987-1995 Pr Low 0.1% 3.0% 1.3% Middle 0.2% 3.1% 1.3% High 0.5% 3.5% 1.5% 1995-2010 Low 1.6% 1.7% 0.7% Middle 2.0% 2.1% 0.9% High 2.5% 2.6% 1.3% 1987-2010 Low 1.0% 2.1% 0.9% Middle 1.3% 2.4% 1.1% High 1.8% 2.9% 1.3% KENAI 1987-1995 Low 0.7% -0.7% 1.8% Middle 0.7% 0.4% 1.8% High 0.9% 0.3% 1.9% 1995-2010 Low 1.2% 1.2% 0.9% Middle 1.8% 1.7% 1.4% High 2.0% 2.0% 1.7% 1987-2010 Low 1.0% 0.5% 1.2% Middle 1.4% 1.3% 1.6% High ic 1.4% 1.8% MATSU 1987-1995 Low 1.6% -15.9% 1.4% Middle 1.6% -7.9% 1.4% High 1.6% “1.7% 1.4% 1995-2010 Low 0.9% 13.5% 0.8% Middle 2.0% 9.4% 1.9% High 2.5% 6.1% 4% 1987-2010 Low 1.1% 2.2% 1.0% Middle 1.9% 3.0% 1.7% High 2.2% 3.3% 2.0% mime iemnlhimie iinet Table 2.2: Projected Occupied Housing Stock Growth Rates, All Regions and Cases 2.2 Electric Heat Market Share Electric heat market share continues to be an important determinant of total residential sales due to the very large consumption per house in this end use, as well as significant share levels for some housing types in some regions. 2.2.1 History The mid-1970s pipeline construction boom and the early 1980s State spending boom put an economic premium on the availability and first cost of housing in Fairbanks and Southcentral Alaska. Not surprisingly, during these periods significant numbers of electrically heated buildings were constructed. In Fairbanks, a moratorium on electric space heat has been (effectively) in place since the mid-1970s. This factor, combined with a substantial increase in electricity prices and a shift to oil in existing buildings, has resulted in a steadily declining heating market share over the past ten years, reflected in dramatically decreased Kwh per customer. In Southcentral Alaska, the electrically heated share of new residential construction appears to have peaked during the early 1980s building boom. During the past 4 years the gas system has been extended into unserved pockets of the Anchorage market and major new areas of the MatSu valley. Since the gas system reached Palmer and Wasilla in 1984, ~3500 residential and 870 commercial customers have hooked up. Enstar Natural Gas Co. estimates that, as a historical rule, when gas service has reached a subdivision, 75% of newly served houses have converted during the first year, followed by 50% of the remaining stock during successive years. 2.2.2 Current Market Trends’ The entire Fairbanks electric heat market is probably in long-run equilibrium. Even the extension of the gas pipeline would not significantly affect electric heat sales, given today’s low market shares.’ The Anchorage single family heating market is now in equilibrium for new construction, with gas service available to 98 percent of the service territory. Among houses already built, thousands have converted to gas from other fuels during the past 5 years. In the multifamily market, with a 24% electric heat share,’ there appears to be some continuing conversion activity in two market areas. First, in several large apartment complexes we contacted, ‘Data on conversions was provided by Enstar Natural Gas. Co., personal communication, February 1988. >This analysis of heating market structure is based on: (1) Interviews with Enstar Natural Gas Co., Mountain Alaska Energy, MatSu Borough, Kenai Peninsula Borough, Cities of Homer and Seward (2) Detailed review of Platting maps for the Kenai and Matsu Boroughs. °GVEA’s 1987 Power Requirements Study notes (p. 47) that the number of residential customers recording more than 4000 Kwh of December consumption fell from 2450 in 1975 to 721 in 1980 to 192 in 1986. These data do not capture residential use in multifamily buildings served through commercial rate class meters. “As we discuss below, these shares are derived from a survey of occupied units. The share of the total stock including vacancies would probably be higher. landlords pay heat or plan to start offering paid heat after conversion as a marketing strategy. Similarly, in the condominium market, two of the three major secondary lenders (MGIC, FNMA) regard conversion to gas as a valuable boost to the unit’s salability and are converting projects as part of many of their sales. In both of these market segments, the current housing recession "cuts both ways" in motivating conversions: higher quality units are hard to sell without first converting to gas, while lower quality units do not justify the investment in conversion. Enstar Natural Gas Company is not currently planning to serve the Turnagain Arm area. They did look into using the existing LPG pipe from Elmendorf to Whittier to serve Girdwood and Whittier, but the military wouldn’t sell or lease the line. Mountain Alaska Energy has also Market Value of Electric Heat Alaska’s current housing market oversupply has caused a "flight to quality" by consumers. In a recent regression analysis of a large sample of AHFC-financed transactions, Berman (1988) estimated the amount by which the presence of electric heat affects the selling price of a home. His analysis controlled for the house’s age, size, features, and location, and excluded areas of the Anchorage Hillside where gas is not available. The results are expressed as the percentage difference in sales price attributable to electric heat: ; % Difference T-stat -13% 1.26 -16% 2.12 -15% 187 Market ANC/MAT Single Condo Single Fairbanks With average sales prices in the $80-100,000 range for single family homes, the percentage adjustments above clearly exceed the cost of conversion in single family homes. It may be that the presence of electric heat is correlated with other undesirable house attributes not captured by the regression. In condominiums the data are inconclusive. considered serving Girdwood, but currently finds the market too small to support the cost of the line, considering construction problems of a narrow right of way and extensive bedrock. On the Kenai Peninsula, Enstar estimates that it is continuing to hook up 50 percent of the remaining non-gas housing stock per year in their service territory, which extends south from Sterling as far as Soldotna and Kenai and bypasses Seward. Mountain Alaska Energy has applied to the Alaska Public Utilities Commission for a permit to construct a new gas pipeline which would serve Homer and territory north to Anchor Point. In the MatSu Valley, the gas system is still catching up with the residential housing stock. Twenty three pockets of unserved subdivisions representing 1,427 homes have been organized into proposed Local Improvement Districts (LIDs) within close reach of existing gas lines. Three of these districts have been approved by voters. We estimate there are ~300 homes close enough to these LIDs to also receive service with the LIDs. Several relatively large unserved areas in the MatSu Valley may get gas service within the next few years: e Big Lake and Houston (1600 improved parcels)* will probably be first, because one adopted LID would bring a distribution line very close to Big Lake and because the residents are more actively pursuing gas service. Enstar would run the main transmission line down the highway, and reinforce it with perhaps several transmission lines, including 5 Improved parcels have some type of building on them. 2-6 the Hollywood Road line. The Houston high school, one mile from Big Lake, uses propane, and can be easily converted to gas. e The Butte/Lazy Mountain area (1200 improved parcels) is less likely to be served because of the extra $100,000 cost of crossing the Matanuska River. Proper protection of Salmon spawning areas would further increase construction costs. The area has a very low density compared to other valley service areas. e Meadow Lakes (1500 parcels) is unlikely to get service unless settlement density increases, although a proposed LID would service a small part of it. e The Independence area (374 parcels) would only get service if the proposed Hatcher Pass ski resort is built. Enstar’s preliminary cost estimate ranges from $1-3 million, depending on the size of the ski area. Developers are considering using gas to generate electricity for the resort. 2.2.3 Multiple Heating Fuel Use The amount of electricity which gas conversions may save depends critically on the current structure of the heating market, since conversions from propane, oil, or wood save no electricity, while conversions from partial electric heating save some fraction of the "all- electric" amount. Using end use survey data, we were able to quantify both the extent and nature of mixed-fuel heating. One important survey result is that there are very few all-electric single family heat users left in the Railbelt. Most such households that use electricity do so in combination with other fuels. To prevent this fact from seriously biasing our analysis, we sorted out all possible combinations of heat sources and determined how many customers used each. Table 2.3 summarizes this breakdown for electric heat customers. Regression analysis (see Appendix A) showed that a primarily electric user used 60 percent as much as an electric only customer. For partially electrically heated houses, the figure was 40 percent. Using these weights, we constructed electric heat market shares on an energy basis. That is, we determined the number of "all electric'-equivalent households® as: HAE = H, + .6*H,, + .4*H, + 33*H. (2.1) where HAE = # of “all electric" households which would consume the same energy as the actual population now does; = # of observed, actual, all-electric houses # houses primarily electric with a secondary fuel; # houses with nonelectric primary source and secondary electric; H.,. = # houses reporting three or more fuels, none primary (Upper case subscripts denote a primary fuel while lower case denotes a secondary fuel) He Hs Ax. SThis concept is exactly analogous to "Full-time-Equivalent" employment statistics. 2-7 Railbelt Residential Heat Fuel Combinations (Responses to ISER 1987 End Use Survey) Fuel Mix ANCHORAGE FAIRBANKS KENAI MATSU (PRIMARY : secondary) number percent | number percent | number percent | number percent Single Family total 301 100 111 100 67 100 59 100 nonelectric heat 262 87 9 86 40 60 31 53 ELECTRIC only 6 2 1 Hl 4 6 6 10 GAS:electric 17 6 2 3 1 2 OlL:electric 8 v 1 1 1 2 ELECTRIC:gas 3 1 ELECTRIC: wood : 3 1 1 1 7 -10 5 Db ELECTRIC:oil 1 2 WOOD :electric 1 0 4 4 8 12 11 19 gas:electric:wood 7 2 1 2 oil:electric:wood 2 2 2 3 5 5 oil:electric:other wood:electric:other 2 1 3 4 1 2 Fuel Mix ANCHORAGE FAIRBANKS KENAI MATSU (PRIMARY: secondary) number percent | number percent | number percent | number percent Multi family 179 100 37 100 11 100 1 100 nonelectric heat 126 70 30 81 @ 64 4 36 ELECTRIC only 38 21 4 11 ) 27 5 45 GAS:electric it 4 1 9 OlL:electric ELECTRIC:gas 3 2 ELECTRIC: wood ‘ 1 ELECTRIC:zoil 2 1 1 9 WOOD:electric 1 9 gas:electric:wood 2 1 oilz:electric:wood 3 8 oil:electricz:other wood:electric:other Fuel Mix ANCHORAGE FAIRBANKS KENAI MATSU (PRIMARY : secondary) number percent | number percent | number percent | number percent Mobile Home 44 100 14 100 15 100 4 100 nonelectric heat 37 84 8 57 13 87 ELECTRIC only GAS:electric 4 9 OlL:electric 1 25 ELECTRIC:gas 3 7 ELECTRIC: wood ELECTRIC:oil WOOD:electric 1 25 gas:electric:wood oil:electric:wood 6 43 1 25 oil:electric:other 1 t wood:electric:other 1 7 f 235 Note: In the Fuel Mix listings, uppercase names denote a primary fuel. Source: Responses to ISER end use survey, November 1987 Table 2.3: Reported Railbelt Heating Fuel Combinations The remainder of our analysis is conducted in terms of these “all electric'-equivalent market share numbers. It is important to keep in mind that these numbers are: Always less than the number of customers who "use electric heat," hence less than the apparent heat shares determined in some past Railbelt end use surveys.’ Sensitive to the weights used in their construction, since relatively few customers are unequivocally all electric. 2.2.4 Electric Heat Share Model Inputs 1987 Electric Heat Shares All-electric Equivalent, % The residential end use model requires as inputs current estimates of the stock average electric heat share, the electric share of replacement equipment, and the ANC FBX = KEN MAT electric share of new construction. The x . A ingle 6 4 21 24 latter two variables are then projected by Multi 24 9 28 47 5 2 5 20 the model each forecast year based on the Mobile starting values. Our esImates OID y>E>E>EE>>——————EESESEE=s average electric heat market shares Table 2.4: Current Electric Heat Shares calculated as above are presented in Table 2.4. These are the base year average market shares input to the end use model. We assume an infinite lifetime for electric baseboard heating equipment. That is, the equipment lasts as long as the house it is installed in. We also assume that households will never switch to electric heat when an existing fossil unit wears out. These two assumptions, well-supported empirically, imply that the electric share of replacement heating equipment is zero for forecasting purposes. We thus model the decision to switch from electricity to fossil fuel as a pure "conversion" decision rather than a "replacement" decision. This formulation effectively takes the determination of the rate and timing of gas conversions outside the normal fuel choice equations of the end use model for existing houses. Exogenous treatment of conversions is called for by the disequilibrium that still exists in the Southern Railbelt, due to recent building booms and continuing expansion of the gas delivery system. New Home Electric Heat Shares All-Electric Equivalent, % Estimating the initial values for electric heat market shares of new construction is difficult because the houses built during the last three years are not a representative acl OK KEN AT sample of the stock that will be constructed _ single 5 4 15 15 once the current backlog of vacancies is re- — Mul ti 10 9 15 20 Mobile 10 2 20 10 occupied. Most of the houses built during this time have been high-end custom construction jobs. Almost none are peer eeeeeneeseenebeed nemmecemnenensasnenmndemmenionealidinaetantemenianemmenciemenasiaeammmmetioeeiamsamendiadl Table 2.5: New Construction Electric Heat Shares "Homer Electric used a similar weighting scheme in a regression analysis of heat consumption from their 1987 member survey. HEA used weights of 1 for "electric only” respondents and 5 for anyone who answered that they used both electric heat and some form of "alternate" energy (not gas or oil). Our survey data indicate that a substantial number of Kenai homes do in fact heat with electricity and gas/oil; they may have been coded as "electric only" under the HEA protocol, which may be one reason HEA’s regression analysis showed the "solely electric"-equivalent heat consumption to be only 7074 Kwh per year, which is at the low end of the plausible range. 2-9 electrically heated.* Also, little if any new construction is projected before 1995 under any of the housing stock projection cases. In making the judgments on these parameters, we assumed that when construction does resume in the mid-1990s, there will be some continuing long-run demand for electric heat. Electric heat can pay in homes that are little used (such as vacation homes), and it offers the consumer certain quality advantages over fossil fuels. The initial new construction heat shares chosen are a conservative combination of the stock average and the very recent past data. In Fairbanks, where we believe the market is now in long run equilibrium, the new construction shares are set equal to the current average. The future values of new home electric heat shares were chosen exogenously consistent with the gas conversion case being modeled (see next section). These future shares are presented in detail in Appendix C. 2.2.5 Projected Gas Conversions We developed a spreadsheet model to project the number of electrically space heated residential units which will convert to natural gas under two distinct sets of assumptions about gas distribution system growth and conversion activity. This model produced vectors of conversions for each of the Southern Railbelt regions under each of the following alternatives. The results are shown in Table 2.6 and the calculations in Appendix G. e Base Penetration: No significant extension of the gas distribution system, but conversions of existing heating systems to gas continues at historical rates. Specifically: Anchorage Borough a. Conversions occur at a rate based on recent Enstar data reflecting the fact that virtually the entire Anchorage Bowl is currently served and the assumption that existing multifamily units do not convert to gas. A core of electric and fuel oil use remains. , b. Gas service is not extended into Girdwood or Whittier. Kenai Peninsula Borough a. Conversions occur at the same rate as in Anchorage. b. Gas service is not extended into Homer or other new areas not currently served. Matanuska/Susitna Borough a. Conversions in this new or "immature" market continue at the annual rate of 25% of the remaining stock accessible to gas until a "core" residual of electric/wood space heating remains--consistent with the pattern in the "mature" Kenai Peninsula gas market. b. Gas service extends into "LID" subdivisions on the assumption that the residents desire the gas and conversions occur at the same rate as in "a". ®Ron Watts, Anchorage building inspector, via Peter Poray, CEA, 15 August 1988 letter; Bob Klein, HEA, personal communication, 2 August 1988; Our analysis of AHFC transactions data showed that of ~120 AHFC-financed homes built in 1986 and 1987 throughout the Railbelt, only 1 was electrically heated. This structure was on the Kenai Peninsula, where gas may not have been available. 2-10 High Penetration: Anchorage Borough a. Double the amount of conversions occur as in alternative 1., reflecting some conversion of multifamily units from electric to gas heat. b. Gas service is extended into the Girdwood area in 1990. Conversions occur at the same rate as in the Matanuska/Susitna Borough. Kenai Peninsula Borough a. Conversions occur at the same rate as in Alternative 1. b. Gas service is extended to Homer in 1990. Conversions to gas occur at half the rate as in the Matanuska/Susitna Borough. Matanuska/Susitna Borough a. Conversions of gas-accessible stock occur as in Alternative 1. b. Gas service extends during the next 5 years into major areas not currently served--Meadow Lakes, Big Lake, and Butte. Conversions occur at the same CP rate as "a". Cumulative Number of Gas Conversions |---- Base Penetration ----- | |---- High Penetration ----| Cumulative Conversions as of: 1990 1995 2000 2010 1990 1995 2000 2010 Anchorage Total 250 556 854 1428 436 1611 2696 4562 Single 187 417 6401071 109 403 674 1141 Multi 31 69 107 179 284 1047 1752 2965 Mobi le 31 69 107 179 44 161 270 456 Kenai Total 148 258 310 357 337, 1198 = 1441 1541 Single 92 160 192 221 209 743 893 955 Multi 30 52 62 71 67 240 288 308 Mobile 27 46 56 64 61 216 259 277 MatSu Total 693 1238 1496 1639 693 1647 2141 2378 Single 534 953 1152 1262 534 1269 1648 1831 Multi 90 161 194 213 90 214 278 309 Mobile 69 124 150 164 69 165 214 238 LLL ALE LILI LIE LDL LELLD DSL LEE DOL IED EDE II EE DELEON TELL OLE TE LEI DILLEL LLL ECD ELLE: Table 2.6: Cumulative Projected Gas Conversions 2-11 2.3 Other Appliance Market Shares Table 2.8 displays the market share estimates input to AKREM (the residential model) for end uses other than heat. Three different sets of shares are needed. First, the current average market shares are computed directly from the end use survey. Second, AKREM allows for separate shares for replacement equipment in existing houses. Third, there are new house equipment shares for appliances installed in new houses. All three sets of share values evolve through the forecast period. (The detailed results are shown in Appendix C). Both sets of marginal shares (replacement equipment and new construction) evolve in response to changing economic forces. The average shares evolve in response to the flow of new equipment into. the stock and in direct response to conversion activity. Since we have no comprehensive data on the current levels of marginal market shares, we generally used today’s average market share levels as an analytical starting point for determining the marginal values. We adjusted the average shares to reflect market trends revealed by the end use survey or our interviews with appliance dealers. These adjustments are described below. 2.3.1 Water Heat Water heater fuel choice is closely linked to space heat fuel choice. To simplify the bookkeeping involved in accounting for conversions, we tied a water heater conversion to every space heat conversion. That is, we assumed that all electrically heated homes have electric water heaters, and that all such water heaters are converted when the heating system is converted. This linkage probably overstates the degree of actual water heater conversion, since (1) some electrically heated homes may already have wood or propane water heat, and (2) not all water heaters will be converted along with the heating system due to long plumbing distances to the water heater and/or inadequate venting possibilities. Given this overstatement, we felt that downward adjustment of the marginal market shares for replacement equipment would constitute double counting of conversion activity. We therefore left marginal replacement shares equal to average shares. End use survey respondents did indicate a significant intent to convert from electric to gas water heating within the next three years. Table 2.7 displays these results. Intended Replacements of Electric Water Heaters Within Three Years By Replacement fuel and Region Region ANC FBX KEN MAT # % # % # % # * Plans within the next three years no replacement planned 55 7 51 79 48 73 43 78 plan to replace with electric 7 10 10 15 9 14 4 7 plan to replace with gas/oil 12 16 4 6 9 14 8 14 Table 2.7: Intended Replacements of Water Heaters by Replacement Fuel The survey also showed that almost no one plans to switch fo electricity from another fuel. These results should be interpreted with some caution, however, since many people replace electric water heaters on a "distress" basis. We therefore used these responses as general guidance in adjusting new housing market shares downward to approximately half their average values. 2.3.2 Refrigerators The survey showed that there are an average of 1.03 refrigerators per household. This average value was not changed. 2.3.3 Freezers Survey respondents indicated some plans to buy freezers during the next three years. However, these plans are consistent with expanding personal income, the effect of which is already captured within the model. We left the marginal shares equal to the average values determined by the survey. 2.3.4 Stoves The only adjustment made to the stove shares was a downward adjustment to the Anchorage figure for new construction single family and multifamily homes. The single family share was dropped from .71 to .5; multifamily from .85 to .75. This adjustment is based on current sales data obtained from Anchorage appliance dealers’ which shows the gas share of current sales to be between 40 and 50 percent. In the other regions, data are far more inconclusive, although two dealers claimed there is a boom on in gas appliance sales in the MatSu and Kenai regions. None of the replacement equipment shares are adjusted away from the average due to the high cost of plumbing gas to the stove in existing houses. 2.3.5 Clothes Dryers We adjusted current average dryer electric shares downward by 10-15% to arrive at initial marginal shares in the southern regions. This adjustment reflects increased availability of gas and sales data suggesting that 30-80% of current dryer sales are gas. Dryer fuel choice is not tied directly to space heat conversion scenarios. This lack of a direct linkage reflects our understanding that many gas dryers and stoves are currently plumbed to propane. The forecast values for dryer shares are presented in Appendix C and in Figure 4.7 of chapter 4. The figure shows the evolving value of the average shares. *stolts, Spenard Builders Supply, United Building Supply, Personal Communication and sales data, 12 September 1988. 2-13 Residential Electric Market Shares (Percent of Total Households) AVERAGE End Use Anc Fbx Kenai Matsu Hot Water Single 14 34 56 46 Mult 27 28 33 50 Mobi le 41 48 55 65 Refrigerator Single 103 103 103 103 Mult 103 103 103 103 Mobi le 103 103 103 103 Freezer Single 57 63 74 83 Mult 57 63 74 83 Mobi le ai 63 7 83 Cooking Single 71 67 52 68 Mult 85 100 3 100 Mobile 18 44 10 19 Dryer Single 76 82 67 82 Mult 56 51 64 25 Mobile 83 85 62 49 MARGINAL: Existing Stock Replacements MARGINAL: New Construction End Use Anc Fbx Kenai Matsu Hot Water Hot Water Single 14 34 56 46 Single 5 a 30 23 Mult 27 28 33 50 Mult 10 14 30 25 Mobi le 41 48 55 65 Mobile 41 24 30 32 Refrigerator Refrigerator Single 103 103 103 103 Single 103 103 103 103 Mult 103 103 103 103 Mult 103 103 103 103 Mobi le 103 103 103 103 Mobi le 103 103 103 103 Freezer Freezer Single 57 63 74 83 Single 66 63 7% 83 Mult 57 63 7% 83 Mult 66 63 7% 83 Mobi le 57 63 74 83 Mobi le 66 63 74 83 Cooking Cooking Single 71 67 52 68 Single 50 67 52 68 Mult 85 100 73 100 Mult re) 100 7% 100 Mobile 18 44 10 19 Mobile 18 44 10 19 Dryer Dryer Single 68 82 54 66 Single 68 82 54 66 Mult 51 51 51 20 Mult 51 51 51 20 Mobile 7% 85 50 39 Mobile 74 85 50 39 Table 2.8: Residential Electric Market Shares, Other Appliances Bod 2.4 EUI Values 2.4.1 EUI Estimation Procedure Current residential EUI values were developed from a number of sources including: e National appliance data developed for the U.S. Department of Energy by Lawrence Berkeley Laboratory; e The AKWARM heat loss model and data sets developed for the State of Alaska’s thermal standards analysis; e Vintage Profiles of the Railbelt appliance stock derived from the ISER end use survey; e Usage data from the end use survey. Because the Railbelt housing stock is known with reasonably good precision, we treated the EUI values as the main calibration levers of the residential model. The numbers were developed on an interactive spreadsheet with the twin goals of (1) maintaining a reasonable foundation in engineering and end use data and (2) accurately reproducing control data on total sales by region. Electric heat EUIs are difficult to develop because of the multiple heating fuel phenomenon described in section 2.2.3 and because of the lack of good data on thermal integrity. As stated in section 2.2.3, our analysis was conducted on an "all-electric equivalent" basis. We used an initial estimate of consumption derived directly from the AKWARM thermal model using average housing sizes of 1700 Ft2 (single), 1000 Ft2 (multi, mobile)"* and the proposed state thermal performance standards for heat loss per Ft2, which are generally believed to be comparable to the current stock average of electrically heated construction. These initial estimates of 18,508 kWh/yr for the southern regions agreed almost perfectly with the Figure of 18,511 kWh/yr determined from regression analysis of the end use survey data (see Appendix A). During calibration, Fairbanks values were adjusted up to reflect a colder climate, then down to reflect better insulation levels and to retain consistency with overall sales data. Kenai numbers were adjusted down and MatSu numbers were adjusted up to be consistent with overall sales data. These heat EUIs are substantially less than the Figures used in some previous Railbelt electric demand forecasts (Scott 1983 suggested 40,000 kWh for Anchorage, 53,000 for Fairbanks; Goldsmith & Huskey 1980 suggested 32,000). Water Heat EUIs were developed by using the demand Figure of 18 gal/day/person developed by Lawrence Berkeley labs from a database of 11,000 metered water heaters, and applying the appropriate household size (by housing type) and water inlet temperatures. Refrigerator and Freezer EUIs were developed from nationally reported consumption values (Geller 1988). We adjusted these down to account for Alaska’s younger (hence more efficient) appliance stock. The final values used are consistent with our calculated Railbelt stock average efficiency of 5.61 (Energy Factor) and with the Berkeley/DOE technology curve for refrigerators (described below). These house sizes were determined from the Anchorage Property Appraisal Tape (N=59,000). They agree well with residential end use survey data. 2-15 Cooking and Drying EUIs were taken from Geller 1988. The Anchorage and Fairbanks values were adjusted down to reflect the increased number of meals eaten out in these regions as determined from the end use survey. Lighting numbers are taken from Geller 1988, the AKREM default data base, and simple engineering calculations. Although Alaskans receive the same amount of total annual sunlight as the rest of the U.S., most of our late night (summer) light arrives while we are asleep. In contrast, our daytime (winter) darkness is felt during waking hours. This imbalance implies a higher Alaska demand for lighting. EUIs for the Miscellaneous End Use were developed by spreadsheet analysis of survey data, reproduced in Table 2.9. Our separate spreadsheet analysis of engine block heater use (Table 2.10) derives regional EUIs less than half as large as those used by Scott (1983) and Goldsmith (1980). Tables 2.11 through 2.14 present the final residential EUI numbers in a calibration worksheet framework. Miscellaneous End Use Consumption Regional Estimates |- kWh / Appliance -||------- Saturations ||---- kWh per customer ----| Appliance kWh or kW * Hrs Anc Fbx Kenai Fbx Ken Mat Note: fan/pump 400 1.200 1.200 1.200 480 480 480 1 clocks 60 1.000 1.000 1.000 60 60 60 Iron 120 1.000 1.000 1.000 120 120 120 Dishwash 150 0.810 0.520 0.600 78 90 108 1 Microwave 30 0.820 0.810 0.810 24 24 25 1 Stereo 1 1.000 1.000 1.000 75 75 75 TV 300 1.300 1.700 1.700 510 510 540 1 Toaster 20 1.000 1.000 1.000 20 20 20 Vacuum 30 1.000 1.000 1.000 30 30 30 Washer 60 0.880 0.840 0.860 50 52 55 1 Computer 20 0.250 0.240 0.210 0. 5 4 4 Bolt Htr [See detailed calculations Table 2.10] 108 604 110 114 2 Jacuzzi 1600 0.042 0.022 0.015 0.047 67 35 24 75 is Sauna 1600 0.046 0.028 0.029 0.047 7% 45 46 i is Waterbed 1200 0.260 0.240 0.221 0.310 312 288 265 372 1,4 Heat Tape 400 2 200 0.025 0.006 0.030 0.009 10 2 ne 4 1 Total Misc. kWh: NSO heal 1923 ei er, sessssscsc= piissssssscssssssssssssssssssssssssss2se=s2=====: Notes: 1 Saturations from ISER 1987 End Use Survey 2 See Separate table for Block Heater Analysis 3 kWh per yr from Battelle, 1983 4 kwh per yr Estimate from ISER regression compares with 1440 engineering analysis eeEaesezessccesces SEE RT ee Ri ROOT TENS OMI cit EE REE TAD: Table 2.9: Miscellaneous Appliance EUI Derivation ERE SEDI SITS SAEED TL LETTE TEE LL, SOE CTA I LS EG TON HI is * OY Engine Block Heater Analysis Ownership (1) fraction owning: 1 0.25 0.31 0.16 0.31 2 0.19 0.39 0.25 0.24 Total Saturation: 0.63 1.09 0.66 0.79 Use patterns among users: 1: used if <0 (-5 av) 0.28 0.52 0.31 0.45 2: used if <20 (10 av) 0.72 0.48 0.69 0.55 Weather Data (2) hours < -5 183 1571 183 183 hours < 10 905 2684 905 905 Engineering Assumptions: Hours enabled = .5 * weather group 1 92 786 92 92 group 2 453 1342 453 453 Duty Factor: group 1 0.5 0.5 0.5 0.5 group 2 0.4 0.4 0.4 0.4 Conclusions: Time on: group 1 46 393 46 46 group 2 181 537 181 181 average 143 462 139 120 Capacity (kW): 1.2 1.2 Tee 1.2 Consumption (kWh): 172 554 167 144 (per device) Consumption per 108 604 110 114 household Notes: 1 Ownership and usage patterns from ISER end use survey. 2 Weather Data from NWS Bin Data for Eielson and Elmendorf AFB ATE Et SER AL cA TATA SEALE AEC ETT ALITTLE AEB EP TIOIBS RAE Nit aT se NAM ERE Table 2.10; Engine Block Heater EUI Analysis Qe TL A AT TA AE ELL IEEE ODOT LITLE ELE BA IE EE | SARA TS ETL TIERCE: Final Residential EUIs Region: ANC Total Sales Calculated: 689 Total Sales from Data: 670 Current House House Market EUI Sales End Use Type Stock Share (kWh/yr) GWh HEAT Single 36,3 0.06 18429 40.2 Multi 35.3 0.24 10000 84.8 Mobile Saf 0.05 10840 3.1 WATR Single 36.3 0.14 5300 26.9 Multi 85.3) 0.27 4770 45.5) Mobile 57 0.4 4180 9.5 FRIG Single 36.3 OS: 1200 44.9 Multi B53 1.03 1100 40.0 Mobile S57 103 1000 5.59 FREZ Single 36.3 0.54 1000 19.6 Multi 35.3 0.54 1000 19.1 Mobile aud 0.54 1000 Sal COOK Single 36.3 0.71 650 16.8 Multi 35r75) 0.85 650 19.5 Mobile Saf 0.17 650 0.6 DRY Single 36.3 0.75 1100 30.0 Multi 35:..3 0.56 1100 21.8 Mobile Oi 0.79 1100 4.9 LITE Single 36.3 1 1500 Das 5 Multi 55.3 1 1000 33.3 Mobile our 1 1000 Sis MISC Single 36.3 ZL 1950 70.8 Multi 96.8 A 1950 68.9 Mobile a7 1 1950 io VACANT Single Pe 1 600 1.6 Multi 7.2 1 600 4.4 Mobile 1.4 1 600 0.9 Table 2.11: Anchorage Residential EUI Calibration Worksheet 2-18 Final Residential EUIs Region: FBX Total Sales Calculated: 219 Total Sales from Data: 207 Current House House Market EUL Sales End Use Type Stock Share (kWh/yr) GWh HEAT Single ESS) 0.04 20000 12.4 Multi 6.7 0.09 10000 6.0 Mobile z.5 0.02 10000 0.5 WATR Single 13.5 0.34 5302 28.0 Multi 61.7 0.28 4770 9.0 Mobile 2.3. 0.47 4183 405) FRIG Single 5155) | 1200 18.6 Multi S.7 a 1100 7.4 Mobile 2.3 1000 243 FREZ Single 15.5 0.63 1000 9.8 Multi 5.7 0.63 1000 4.2 Mobile 1.2 0.63 1000 1.4 COOK Single MS AD) 0.67 650 6.8 Multi S.7 1 650 4.4 Mobile 235 0.44 650 O.7 DRY Single a wae 0.82 1100 14.0 Multi SL? 0.51 1100 3.8 Mobile Zea 0.85 1100 aos LITE Single 13.3 se 1500 2325 Multi 47/ 1 1000 6.7 Mobile 2.3 1 800 1.8 MISC Single 13.5 1 2000 3130 Multi 6.7 a 2000 13.4 Mobile aa 1 2000 4.6 VACANT Single LO) 1 600 0.6 Multi 2.4 1 600 Le Mobile 0.6 1 600 0.3 Final Residential EUIs Region: KENAI Total Sales Calculated: 165 Total Sales from Data: 153 Current House House Market EUI Sales End Use Type Stock Share (kWh/yr) GWh HEAT Single 3.3 03,3) 10000 27.9 Multi Sea 0.28 7500 625 Mobile iad) 0.05 8000 0.9 WATR Single a2 0.56 5302 Q2ad Multi Se 0.33 4770 4.9 Mobile 2.3 0.55 4183 a. FRIG Single 953) 0.96 1200 10.7 Multi aa 0.96 1100 ae Mobile 28) 0.96 1000 22 FREZ Single 9.3 0.73 1000 6.8 Multi aun 0.73 1000 ao Mobile 23) On73) 1000 ee COOK Single 9.3 0.52 800 3.9 Multi 2 O73 800 1.8 Mobile 23) (nak 800 On2 DRY Single One 0.67 1100 6.9) Multi SL 0.64 1100 2.2 Mobile 23) 0.62 1100 1.6 LITE Single Pe 1 1500 14.0 Multi pM 1 1000 Sak Mobile z.3 1 800 D8 MISC Single 9.3 Z 1980 18.4 Multi 8.1 al 1980 6152 Mobile 23 i; 1980 4.6 VACANT Single 0.6 ak 600 0.4 Multi 0.1 i 600 0.1 Mobile 0.6 a 600 0.3 Final Residential EUIs Region: MAT Total Sales Calculated: 182 Total Sales from Data: 192 Current House House Market EUI Sales End Use Type Stock Share (kWh/yr) GWh HEAT Single 9.7 0.24 23000 bane Multi 12, 0.47 16000 9.2 Mobile ae, O72) 12000 2.9 WATR Single OT 0.46 5302 23.6 Multi r2 025 4770 219 Mobile Le 0.64 4183 3.3 FRIG Single 9.7 a%03 1200 127.6 Multi 2 1303, 1100 1.4 Mobile T.2 1.03) 1000 Lo FREZ Single 9.7 0.83 1000 8.0 Multi ia 0.83 1000 170) Mobile Lae 0.83 1000 1.0 COOK Single 957 0.52 800 4.0 Multi 132 0.73 800 0.7 Mobile Te2 0.1 800 0.1 DRY Single 9.7 0.82 1100 8.7 Multi det 0.25 1100 0.3 Mobile a2 0.49 1100 OL7 LITE Single SAA b 1500 14.5 Multi Rak a 1000 ian ® Mobile 2 1 800 1.0 MISC Single 9.7 aL 2400 23.2 Multi aly a 2400 2.9 Mobile ae 1 2400 239) VACANT Single 2.4 1 600 Lo) Multi 0.8 i: 600 035) Mobile 0.3 1 600 0.2 Table 2.14: MatSu Residential EUI Calibration Worksheet 2.5 Technical Parameters 2.5.1 Technology Curves AKREM uses a technology curve approach to determine the EUI values for new residential equipment. For each end use, a functional relationship is estimated between equipment price and energy consumption. This relationship is used in every year of the forecast to estimate the economically optimal EUI value, given the current year’s energy prices. Thus, the EUIs chosen for new equipment evolve through the forecast period. These projected EUI values are shown in Appendix C. The technology curves were re-estimated for this project by James McMahon and are presented in Appendix B. Two sets of technology curves are used as branches of the critical assumptions tree. The Base set is the one presented in Appendix B. A High set is also used in which the marginal cost of energy efficiency is reduced by 20 percent. Mathematically, this is accomplished by increasing the curvature of the technology curve. 2.5.2 Federal Energy Efficiency Standards The National Appliance Energy Conservation Act of 1987 sets minimum efficiency standards for appliances. The standards were agreed to by appliance manufacturers and conservation advocates and passed by the U.S. Congress in October 1986. In March 1987 they were signed into law as an amendment to the Energy Policy and Conservation Act. NAECA defines performance standards and design requirements for refrigerators, freezers, room air conditioners, central air conditioners, heat pumps, water heaters, furnaces, dishwashers, clothes washers and dryers, direct heating equipment, kitchen ranges and ovens, and pool heaters. The standards apply to appliances manufactured on or after January 1, 1990. For Refrigerators and Freezers, NAECA sets minimum efficiency levels of 13.82 EF for freezers and 7.52 EF for refrigerators. These levels are 17 and 10 percent below 1985 shipment-weighted EF values, and 28 and 26 percent below the 1987 Railbelt average EF values. For Electric Water Heaters, NAECA mandates an 88 percent average fuel use efficiency (AFUE). This is 5 percent below 1985 new appliance levels and 7 percent below the estimated Railbelt 1987 stock average value of 82. NAECA also sets several standards relating to gas furnaces and ranges which are included in the AKREM model. The NAECA standards are input to AKREM at their legal minimum levels. They are a lower bound on efficiency and EUIs are free to evolve towards higher efficiency levels if dictated by economics. This treatment of the standards is a significant conservatism since it is highly likely that manufacturers will sell, and that some consumers will buy, appliances of higher efficiency than the minimum level. Since no one can buy below the minimum, the actual average efficiency of new equipment will be above the NAECA level. This has certainly been true in California, where the marketplace average efficiency is ~7% above 2 - 22 the required minimum set by state standards (Geller 1986). It is also possible that under the provisions of the bill the refrigerator and freezer standards will be upgraded before 1995. Geller (1986) considers this possibility to be "highly likely."" No such changes are assumed for this forecast. Figure 2.2 shows the effect of NAECA on efficiency levels relative to both the 1988 Railbelt stock average efficiency and to the 1985 average efficiency of new shipments. NAECA Efficiency Standards Effect on Railbelt Efficiency AFUE (Watr) Product : Case 0 20 40 60 80 100 = ma ies 1 WATR: AK AVG } WATR: 1985 NEW } WATR: NAECA } FRIG: AK AVG LMM FRIG: 1985 NEW LAAT FRIG: NAECA UA FREEZER: AK AVG FAA FREEZER: 1985 NEW + 3 FREEZER: NAECA = : T r T T T T T 1 0 2 4 6 8 10 12 14 16 EF (Frig/Frez) See text for sources Figure 2.2: Effect of NAECA standards on Railbelt Appliance Efficiency 2.6 Economic Variables 2.6.1 Energy Prices We projected residential fuel prices under Low, Middle, and High crude oil price scenarios for use in the critical assumptions probability tree (see chapter 4). The various crude oil price scenarios drive the price of natural gas used for electric generation. Figure 2.3 shows projected retail residential electric prices by region under the three scenarios. The figure “Geller’s reasoning is that there will be significant pressure on DOE to upgrade the standards by a rulemaking in order to prevent California and other states from adopting their own even more stringent standards. It was the inconsistency among state standards that led manufacturers to support NAECA in 1986. . shows that the dispersion across possible crude oil cases is relatively small compared to the general upward price trend. We computed these prices as follows:” e We projected the cost of gas to the utility by applying the terms of the recently negotiated contract between Enstar and Marathon Oil Co. to our three crude oil price scenarios. e We projected non-fuel costs by relying on recent or in-progress utility power requirements studies. We. were able to use a financial forecast by Chugach Electric, based on a detailed generation expansion plan, as the basis for wholesale price projections for the southern Railbelt utilities. e Regional electric prices are a sales-weighted average of constituent utility prices. As Table 2.15 shows, prices grow rapidly between 1987 and 1995 Residential Electricity Price Growth Rates a ie recently negotiated gas 1987 $, MID Case Projections contracts between Marathon Oil MTA ANCL — Fe KEN AL Retr Company and its wholesale 1987-1995 3.1K 0.8% 1.6% 2.3% 2.3% customers take effect and as the 1995-2010 0.8% 0.1% 0.1% 0.2% 0.5% Bradley Lake hydro project 1987-2010 1.6% 0.4% 0.6% = 1.0% 1.1% comes on line. Thereafter, prices AA AE A at ATCA IE TAS ELT TIE YP EL IE TIT grow with the world price of Table 2.15: Residential Price Growth Rates crude oil. Natural Gas and Fuel Oil prices were also projected consistent with the three crude oil scenarios. Figure 2.4 displays these prices.” In Figure 2.5, we show relative residential fuel prices for delivered heat by assuming furnace efficiencies of 80% for gas furnaces and 75% for oil. The figure shows that while electricity prices rise relative to natural gas prices in Southcentral Alaska, they fall relative to oil prices in Fairbanks. For a detailed description of the computations, see memo of 8/23/88 from S. Colt to R. Emerman titled "Retail Electric Price Projections," available from ISER. 2 Survey Data showed that prices on the Kenai Penninsula were approximately 2 cents lower than prices in the other three regions. 2 - 24 St-72 ANCHORAGE 1987 $ per KWH Year — Low — Medium —* High KENAI 1987 $ per KWH ase_ pue uolsey Aq soolg oDa[q [eNUSpIsoy [Ieloy :¢7 omMsLy Year —— Low -—+—Medium —* High Residential Electricity Price Projections 1987 $ per KWH 0.14 0.12 hy tg gp - pee tte te ee ae i See 0.08 0.06 0.04 0.02 (ee eee eee ee 1987 Year —— Low —+~—Medium -—*~ High 1987 $ per KWH 0.14 ap 0.125 T O11 0.08 0.06; 0.04; 0.02 (1) ae a a SS a ee et ee el 1987 Year Residential Gas Price Anchorage, Kenai, MatSu 1987 $ per Mcf 5.00 4.00F 3.00 F 2.00 1.00 F 0.00 SS ey 1987 1990 1995 2000 2005 Residential Oil Price All Regions 1987 $ per Gallon 1.60 1.40 1.20 1.00 0.80 0.60 + 0 Ahm eee ee 0.20 |-———— 0.00 22 tt 1987 1990 1995 2000 2005 Figure 2.4: Retail Residential Gas and Fuel Oil Prices, All Cases 2 - 26 Relative Residential Fuel Prices Middle Case Projections ANCHORAGE a5 1987 $/MMBTU useful heat 30 25 207 15> 10F 5& 0 1 1987 1990 1995 2000 2005 2010 . Year —~ Electricity —°~ Natural Gas FAIRBANKS i 1987 $/MMBTU useful heat BOE pe eet 25> 20-7 15> VO i ett 5/ e7 7990 — "3995. i 2000 ~ ‘2005 - "2010 Year —— Electricity —*~ Fuel Oil Po Figure 2.5: Relative Residential Fuel Prices: Electricity, Gas and Oil 2-27 2.6.2. Consumer Discount Rates AKREM uses average, observed consumer discount rates in determining the efficiency of new appliances. These rates were calculated by Ruderman et al (1987) by determining what discount rates must have been in order to induce the appliance purchases which actually occurred during the early 1980s, assuming consumers were following a life-cycle cost minimization rule. The resulting rates are quite high, generally falling in the 50-150 percent range. These rates should not be thought of as actual financial criteria which consumers apply when choosing appliances. Instead, they can be regarded as a "reduced form" description of all the noneconomic factors which determine appliance choice, as well as imperfect information about costs and performance and the perceived risk involved with an illiquid investment in energy efficiency. The discount rates can also be thought of simply as calibration constants which are adjusted so that the observed data (appliance purchases) fits the theoretical model of life-cycle cost minimization. 2.6.3 Exogenous Demand Growth Increasing House Size. Over the past 30 years the average size of a new Anchorage house has increased at an average annual rate of ~1.8%."* We assume that the average size of new houses will continue to grow at an average rate of .6% per year throughout the forecast period. This growth is modeled as an exogenous increase in the size of all types of housing. Miscellaneous Equipment Growth. We assume 1.3% annual growth in miscellaneous end use demand to reflect our basic ignorance of new uses of electricity which may become available. 2.7 Environmental Factors In this section we discuss two environmental issues which we initially felt might affect the demand for residential electricity. We conclude that no ad hoc adjustment to the inputs described above is called for. 2.7.1 Indoor Air Quality” Air exchange through infiltration accounts for a significant fraction of home heating costs. The widespread installation of insulation to cut these energy costs has increased concern over indoor air quality. However, in most cases, neither weatherized conventional nor superinsulated new houses in Alaska require additional devices to offset indoor air pollution. When required, these devices consume minimal amounts of additional electricity. “As determined from simple semi-log regressions of house size on time, controlling for the 1980s “petrodollar b boom" with dummy variables. Data from Anchorage Property Appraisal Files. N ranges from 458 (Triplexes) to 42,120 (Single Family). R? > .98 in all cases. SThis section is based on interviews with Don Markle, Alaska Cooperative Extension Service; Norman Bair, DCRA; Adams Morgenthaler & Co.; and reference to EPRI 1986. 2 - 28 ASHRAE’s” 1981 standards recommend a minimum of 0.5 air change per hour (ach) for residential buildings. For commercial buildings it is 4 ach, but this includes inside as well as outside air. In 1979, the U.S. Department Housing and Urban Development (HUD) adopted residential standards of 0.5 ach plus natural ventilation through windows that equal 1/20 of the floor area of each room. If there are no extenuating circumstances (see below), 0.5 ach is still considered sufficient. In fact, ASHRAE may soon lower their standards to 0.35 ach. Canada’s standards are currently set at 0.35 ach; Sweden’s at 0.25 ach. In a conventional house, indoor air pollution is marginal, unless there is a specific problem like radon or an extremely high level of pollutants such as fiberglass, formaldehyde, or particle board. Weatherization doesn’t usually make a house tight enough to create a problem. In tighter houses with new construction techniques, levels as low as 0.1 ach have been measured. However, heating and ventilation systems in new houses are now designed together for both structure and situation so that no additional ventilation should be needed. Designs can be site-specific, depending on whether occupants have plants, smoke, boil food, or work with fiberglass. Retrofits for indoor air pollution usually consist of a series of small fans. Air-to-air heat exchangers, which move air more slowly than a forced-air furnace and do not create drafts, meet the ASHRAE minimum standards and are adequate for most indoor air pollution problems. Ventilation systems without heat recovery may not be effective in cleaning out pollutants, since it is difficult to match exhaust with intake ventilation. Fully integrated, ducted heat recovery systems do both together. Most houses also have point-of-usage fans in the bathroom and kitchen; volume can be increased to deal with greater amounts of point-source pollutants. In new construction with good designs, specific pollution sources such as radon can also be dealt with in construction, usually by isolating the house from the ground. When a problem is discovered after construction, retrofit applications such as fan systems designed specifically for radon mitigation must be installed. Installation is very expensive--the ground must be depressurized so that radon is sucked out before it reaches the house. After installation, the system uses two low-powered fans which have energy requirements equal to a 60 watt heat exchanger: the same consumption as a standard light bulb. The electricity consumption of indoor air pollution mitigation devices ranges from 200 to 1000 kWh per year, depending on house size, weather, and number of occupants. A generally accepted average figure is 500 KWh/yr for a 60 watt air exchanger left on continuously. With the increasing demand for heat exchangers, several companies are coming out with more efficient designs. One company is working on a 30 watt machine. In summary, while additional mitigation devices have nontrivial electricity consumption, it is unlikely that they will contribute substantially to net demand because residential retrofit demand for mitigation devices is correlated, almost by definition, with significantly reduced heat energy consumption. If even 5-10% of the houses so retrofit have electric heat, the savings from weatherization cancel out the aggregate demand from fans. Since we are conservatively assuming that zero weatherization activity takes place during the forecast '6The American Society of Heating and Refrigeration Engineers. 2 - 29 period, it is not appropriate to count as additional demand what is essentially a side effect of retrofit activity. 2.7.2 Restriction of Wood Stove Operation” Wood stoves are an attractive heat source used heavily in combination with electric heat, especially in areas not served by gas. Restrictions on wood stove use in the southern Railbelt because of air pollution problems could increase electricity demand. In 1987, the U.S. Environmental Protection Agency issued new ambient air standards for particulate matter 10 microns and smaller, the PM10 regulations. Cities found in violation must formulate implementation plans. Eagle River is in category 1--in violation of the new levels. Fairbanks is in category 2--a "wait-and-see" area where the potential for future problems exists if the population grows. The Municipality of Anchorage began additional monitoring in Eagle River to see what restrictions will be needed to meet PM10 standards. It is unlikely that wood stoves will be targeted, since violations usually occur in the spring and fall, not the winter. The most likely sources of high PM10 levels is fugitive dust from roadways, open lots, or river beds. The Matanuska-Susitna Valley was not found in violation of PM10 regulations. The area may have been exempted because of the rural fugitive dust policy, an informal ruling that rural areas shouldn’t be subjected to the same standards as cities. This is being challenged now, and the outcome could have an effect on the valley’s compliance status, but not in the near future. EPA recently completed New Source Performance Standards (NSPS) for wood stove manufacturers. This will add to the cost of stoves, and so may affect their economic attractiveness, but the standards militate against further restrictions on use. By 1990, all stoves must meet stack emissions standards. Outside of EPA regulations, there is little likelihood of restrictions from complaints. The state only monitors when there are complaints, or when they are made aware of certain industry and wood stove density changes. In summary, it is unlikely that wood stove restrictions will lead to a significant shift toward electric heat during the next ten years. By that time, the stove stock will have turned over enough that NSPS-regulated equipment will form a significant fraction of the total. Even so, the prospects for stove restrictions are one reason we have retained new house electric heat shares substantially greater than zero. 2.8 Model Calibration Tables 2.11-2.14, presented above, indicate the extent to which the adopted EUI values reproduced control sales totals. At the top of each, the "calculated total" is shown next to the control sales data. Overall, this discrepancy was less than 7%. In a final calibration process we computed constants which forced calculated 1987 sales to be equal to control "This section is based on interviews with Randy Poteet, Anchorage DHHS; Debra Williams, American Lung Association; Bill McClarence, DEC; Diane Sutherland, EPA. sales totals adjusted for commercial meter sales and weather. These factors are loosely analogous to the constant term in a regression equation. They also reflect the consumption of second homes, which are not modeled explicitly. The calibration constants are computed as: (Weather-normalized residential sales plus res. sales through comm. meters) (Calculated sales from end use model) The values of these adjustment factors are: e Anchorage: 1.03 e Fairbanks: 0.96 e Kenai: 0.97 e MatSu: tei, We chose to apply these final adjustment factors directly to the end use forecast of total sales after the raw output was generated by AKREM. This decision reflects our ignorance about the sources of the discrepancy between calculated and actual 1987 consumption. The discrepancy could be due to a combination of many factors. This Page Intentionally Left Blank 3. COMMERCIAL INPUT ASSUMPTIONS The commercial’ sector is far more heterogeneous than the residential, and hence far more difficult to characterize via summary statistics. Buildings vary in size by a factor of ~100, are put to hundreds of different uses, and contain a multitude of electric equipment which defies characterization as a set of appliances. While we use the same analytical approach in both sectors, it must be recognized that the point estimates used to represent the "stylized facts" of the commercial market are forced to summarize far more variation than their residential counterparts. 3.1 Commercial Floor Space We used city and borough property appraisal files to compile an inventory of the Railbelt nonresidential building stock, classified by size, age, and space type. 3.1.1 Current Floorstock Composition Figure 3.1 shows the 1988 regional breakdown of Railbelt floorstock. Figures 3.2 through 3.5 show our estimates of the 1987 Railbelt floorstock by region and building type. Figure 3.6 affords a comparison of the commercial floorstock composition across regions. It shows the percentage of each building type relative to regional total floorstock, with all four regions clustered together for comparison. For example, Fairbanks has 36% of its commercial space devoted to warehousing and storage buildings, reflecting its status as a regional distribution center. Anchorage, meanwhile, has twice the relative office space of its sister regions, reflecting its role as a financial, personal, and legal services hub. Figure 3.1: 1987 Railbelt Floorstock 1987 Railbelt Floorstock Total = 81.5 Million Ft2 ANCHORAGE 54.5 FAIRBANKS 13.2 Since our floorstock estimates are based on an actual building count, we know they are low. Without an independent control total number, however, it is difficult to assess the degree of the undercount. To check the reasonableness of the total floorstock numbers for each region, we compared our direct estimates to those obtainable from national ratios of Ft2/Population and Ft2/Employee taken from the 1983 NBECS data.’ Table 3.1 presents these comparisons. ‘A more accurate term would be nonresidential buildings, since the stock includes government, education, and public assembly buidings. The term commercial is taken from the utility rate class to which most of these buildings belong for billing purposes. "DOE's Nonresidential Buildings Energy Consumption Survey is a comprehensive assessment of ~5000 buildings, completed in 1979 and again in 1983. NBECS ratios by building type can be found in McMenamin 1988. 3 = Building Type Composition Anchorage Region, 1987 Building Type Small Office KA Large Office XA. xi NW Restaurant \\ Large Retail AA) Small Retail Kl Grocery \ Warehouse {\\\\) \\ Car service Lodging \\\ Medical \ School K\Wmm|) College \ Assembly \ Miscellaneous XW Vacant AM ; \ L 0 2 4 6 8 10 12 14 Million Square Feet Figure 3.2: 1987 Anchorage Floorstock Building Type Composition Fairbanks Region, 1987 Building Type Small Office A Large Office Restaurant Large Retail AD Small Retail Grocery Warehouse \ \\ \\ Car Service Lodging Medical School \ \\ * College F Assembly \\ Miscellaneous Vacant \ ; | ! i 0 1 2 3 4 5 6 Million Square Feet * UAF not included Figure 3.3: 1987 Fairbanks Floorstock Building Type Composition Kenai Region, 1987 Building Type Small Office { Large Office \ Restaurant { Large Retail Small Retail KM Grocery \ Warehouse UN AN A) \\ \\\ Car Service WI Lodging AAI Medical N School \\ ANT College - Assembly \ Miscellaneous \\ Vacant \ | 0 0.5 1 1.5 2 2.5 Million Square Feet Figure 3.4: 1987 Kenai Floorstock 3-4 Building Type Composition Matsu Region, 1987 Building Type Small Office Large Office Restaurant Large Retail Small Retail Grocery Warehouse Car Service Lodging Medical School College Assembly Miscellaneous Vacant < Xl a Z ZG \\\ AU ZZ ZL AAA A \ UII) BPeaew'’ gaa FA ZZ ZA _ UI 0.4 0.6 0.8 1 1.2 Million Square Feet Figure 3.5: 1987 MatSu Floorstock 1.4 Percentage of 1987 Regional Floorstock | by Building Type Building Type Small Office Large Office Restaurant Large Retail Small Retail Grocery Warehouse Car service Lodging Medical School * College Assembly Miscellaneous Vacant 0% 10% MMM Anchorage Kenai * UAF not included pd 1 20% AW dd | 30% Percentage Fairbanks Matsu Figure 3.6: Comparison of Building Type Composition across Regions 3-6 40% Overall, our estimates seem to 5 : , ISER Floorstock Estimates vs Estimates from National Ratios lie within the plausible range. The Anchorage floorstock National Ratios: 190 Ft2/Person 619 Ft2/Employee estimate is highest when based ' Fs ANC FBX KEN MAT on population, while the MatSu,. <0 <0 <0) jw os cwtpttbicecccce secon, poeee estimate is highest when based Railbelt Values en a a is = : ‘opulation on population. We would expect Total Employment 116 37 13 BG these results since the MatSu Civilian Employment 105 30 13 6.6 region i rth room el is a Anch bed "0 Estimated Floorstock from: community for Anchorage. On Population 44.1 15.2 7.6 7.0 the Kenai, the close proximity of _ Total Employment = 71.8 22.9 8.04.1 both ratio estimates to our direct CiViUt eri Employment au i6e-0) “= 16-6005” 6-08 os estimate is reassuring, especially vs ISER Direct Estimates 55.0 13.2 7.9 4.9 since this is the One area WhOSC a Statistics are uncomplicated by Table 3.1: Direct vs Ratio-based Floorstock Estimates the presence of significant military or student populations. The least satisfying results of this check are in Fairbanks. However, it must be remembered that our estimates do not include UAF floorspace, while the population and employment figures do encompass UAF. If we add the estimated 1.6 million Ft2 of UAF floorstock to our estimate of 13.2, the result compares well with the population-based estimate of 15.2. (An additional .1 -.2 million Ft2 could conceivably be added to represent the Southeast Fairbanks Census area, for which we had no direct floorstock data.) Although no one knows the exact degree of coverage which assessing files provide, the Fairbanks data appears to include significantly less of the non-profit and government buildings than do other regional files. We have made no attempt to adjust any of the floorstock estimates since we have no firm basis for doing so. The potential floorstock undercount is just one of many potential sources of inaccuracy which was considered during the model calibration process described in section 3.6. 3.1.2 Treatment of Vacancy An abnormally large portion of the commercial building stock is now vacant and requires special treatment to avoid a low forecast of future sales’. To accomplish this we employ a 15th building type to represent excess vacant space. The initial proportion of excess vacant commercial property, P,, was calculated (by region) as one-half of one minus the ratio of 1987 to 1985 employment in the relevant sector: P, = 5 * (1- (Ep / Es)) (3.1) >Carl Macmanus, Fairbanks North Star Borough, Personal Communication, 30 August 1988. “Such an underforecast would occur if current electric consumption control totals were used to calibrate a model based on the physical stock of buildings rather than the occupied stock. As the economy begins to grow, the physical stock of buildings will remain constant for several years as vacancies are reduced. A model based on physical stock would thus predict no consumption increase, while in fact demand could be growing rapidly. The error so induced can be shown to be proportional to the total size of the building stock; hence it grows in absolute magnitude as the forecast progresses. 3-7 For example, if (Es / Es) = .90, excess vacancy equals 5%. The assumed "inertia factor" of .5 reflects our judgment that occupied space is "sticky downward": firms cannot effortlessly shut down space as employment drops. We computed excess vacant space levels for office, retail, restaurant, warehousing, and auto service space types. For all other space types we assume that current vacancies are at their equilibrium level. As employment increases, new construction is called forth. There is no pool of vacant stock to absorb growing demand for occupied space. 3.1.3 Floorstock Projections The demand for commercial floor space within each space type and region is assumed to be a function of one of the following economic or demographic variables--wage and salary employment, tourism employment, population, or federal civilian employment. As demand increases the excess stock of that category of space is returned to the occupied stock. Only when the entire excess stock of that category of space has been utilized will new additions be built. The demand for space of a particular type is independent of the total amount of excess floorstock since a building designed for one type of use is assumed not to be convertible to another use. As a consequence of this assumption, the total commercial building stock grows at the start of the forecast even though there is excess commercial floor space of some types projected for a number of years in all the cases. Table 4.12 shows the growth rates of commercial floorstock. Floorstock projections also appear as part of the projection results in tables 4.6 through 4.8. Railbelt Floorstock Projections Low, Middle, High Cases Million Ft2 140 120 100 40 20 | 9 Vi. A bah nt 1 1 = | 1987 1990 1995 2000 2005 2010 —- Low ——~Middle —High Figure 3.7: Projected Railbelt Floorstock 3-8 g'¢ ans] Anchorage Floorstock Projections Low, Middle, High Cases Million Ft2 100 ry Aq suonsefoig yo01s100],4 40} 20+ 0 Pg ge 1987 1990 1995 2000 2005 2010 o g. —Low — Middle —* High 5 Kenai Floorstock Projections Low, Middle, High Cases Million Ft2 eS NM Nf tht 1987 1990 1995 2000 2005 2010 —-Low —~Middle — High Fairbanks Floorstock Projections Low, Middle, High Cases Million Ft2 25 10F 5+ [ea YY EW Er 1987 1990 1995 2000 2005 2010 —-Low —-Middle High MatSu Floorstock Projections Low, Middle, High Cases Million Ft2 10 4h 2h ee ee eee 1987 1990 1995 2000 2005 2010 ——Low — Middle — High 3.2 Electric Market Shares In today’s Railbelt commercial market, the only truly competitive end uses are space and water heating. While gas cooling technologies do exist packaged in rooftop multi-zone air handling units, we found no field evidence of these units in Alaska. Two large Anchorage building complexes have gas-fired absorption chillers, but these sites are highly unrepresentative of the building population. Electricity’s high cooling market share is hardly surprising, given the low number of cooling degree days in the region. Cooking is another end use for which gas and electricity can both be used. However, we know of no empirical evidence that commercial cooking equipment choice is sensitive to relative fuel prices. In most commercial kitchens, gas is used on ranges, ovens, and steamers, while electricity powers deep fat fryers, grills, and other specialized means of delivering heat. We therefore model the cooking end use as noncompetitive.° 3.2.1 Average and New Equipment Fuel Shares Table 3.2 shows our estimates of stock average and new equipment heat and hot water electric shares by building type and region. The new equipment shares are initial values; the COMMEND model uses them to calibrate a set of choice equations. The new equipment share values are then re-computed each year to reflect changes in life-cycle costs. We collected electric market share data through our on-site commercial survey (N=135, Ft2=6 million) and a brief mail survey (N=596, Ft2=9.5 million). The individual data records were combined into area-weighted average shares by region and building type’. With 4 regions and 14 building types, coverage of individual cells (region/building type combinations) was somewhat thin. However, we elected to use the disaggregated data rather than go through another complicated weighted averaging process and throwing away some information about the differences in shares across building types. There is even less data to support an estimate of new commercial building electric heat shares than there is in the residential sector. Adams Morgenthaler & Co. suggests that the current share is "less than one percent"; however, the amount of recent new commercial construction is so small as to be an unreliable guide for the future. In estimating new building heat shares we have again taken the conservative position that there will continue to be some long-run demand for electric heat. Especially in small buildings, the capital cost of a gas installation can be greater than a first glance might indicate. For example, current building code requires that a separate space be set aside in the building for any heating system of greater than 400 kBtu/h. Our calculations based on the estimated costs of heating systems indicate that the cost of this boiler room space exhibits tremendous economies of scale and can be prohibitive for small enough buildings. Commercial clothes drying is another potentially competitive end use; it is too insignificant to warrant separate classification in the model. “We have a strong hypothesis, as yet formally untested, that electric heat share is strongly and negatively correlated with building size. In larger buildings, commercial customers can take advantage of scale economies in the capital and design costs of fossil heating systems. In addition, the absolute payoff from attention to this cost area becomes large enough to justify management’s attention to a life-cycle cost approach to heating system choice. If our hypothesis is true, it implies that electric market shares of commercial heat energy are substantially lower than electric shares of commercial heat customers. 3 - 10 ~~ As the end use breakdown of sales (Figure 1.2) suggests, the importance of the heat share parameter is far less important in the commercial sector than in the residential. No matter what fuel is chosen to run the heating system, electricity does much of the actual heating in many buildings through lights and office equipment. We also project a continuing significant electric share of the commercial hot water equipment market. Many office and retail buildings use small point-of-use electric water heaters in lavatory areas with low water demand because they avoid the need for both gas piping and exhaust stacks, both of which can be very expensive to install relative to the hot water demand served. It may be be the case, however, that while electric hot water heaters hold a large portion of the market in terms of capacity or square feet served, their share of delivered hot water energy is appreciably lower. To the extent that this phenomenon is a function of building type, it is automatically controlled for in the COMMEND model. 3-11 ATSC TLE | STL ENE oc ‘eALERTS aI UST ETT LOTS aT I PTET AES I IP SCE RET SAS Commercial Electric Market Shares (Percent of Total Square Feet) HEATING HOT WATER Building Type Anc Fbx Kenai Matsu Building Type Anc Fbx Kenai Matsu Small Office Small Office Average 5 1 15 4 Average 17 79 ot 17 New Equipment 5 1 5 4 New Equipment 17 79 57 17 Large Office Large Office Average 2 1 15 4 Average 13 43 57 17 New Equipment 2 1 15 4 New Equipment 13 43 57 Az Restaurant Restaurant Average 2 6 14 3 Average 9 38 17 33 New Equipment 2 6 10 3 New Equipment 2 38 17 33 Large Retail Large Retail Average 18 0 15 3 Average 43 100 50 4 New Equipment 5 0 10 3 New Equipment 43 100 50 4 Small Retail Small Retail Average 26 0 15 3 Average 47 14 50 4 New Equipment 5 0 10 5 New Equipment 47 14 50 4 Grocery Grocery Average 0 0 0 0 Average 27 100 0 0 New Equipment 0 0 0 0 New Equipment 27 100 0 0 Warehouse Warehouse Average 8 0 15 0 Average 32 26 53) 25 New Equipment 8 0 10 0 New Equipment 32 26 53 25) Car Service Car Service Average ic 0 0 0 Average 24 86 20 100 New Equipment c 0 0 0 New Equipment 24 86 20 100 Lodging Lodging Average 0 50 20 0 Average 4 82 100 0 New Equipment 0 10 20 0 New Equipment 4 82 100 0 Medical Medical Average 0 0 20 0 Average 0 0 100 0 New Equipment 0 0 10 0 New Equipment 0 0 100 0 School School Average 0 0 10 0 Average 0 10 100 0 New Equipment 0 0 10 0 New Equipment 0 10 100 O° College College Average 0 0 10 0 Average 100 100 100 100 New Equipment 0 0 10 0 New Equipment 100 100 100 100 Assembly Assembly Average 0 24 14 50 Average 9 7% 54 100 New Equipment 0 24 10 50 New Equipment 9 7% 54 100 Miscel Laneous Miscellaneous Average 5 0 0 100 Average 25 100 30 100 New Equipment 5 0 0 100 New Equipment 25 100 30 100 Vacant Vacant Average 5 1 15 4 Average New Equipment 5 1 15 4 New Equipment === =: =: ssesscssssssss=s==: 87 Floorstock Weighted Average 87 Floorstock Weighted Average Average ri 4 13 Tr Average 24 42 61 18 New Equipment > 1 14 4 New Equipment 17 79 66 ait LLL D ELLE LLL AL, ELLE LTE TOOL: TELE ET EET ELLE OI DIODE A LAID LD LLIN ELA LLL L LEAT NEDA Table 3.2: Commercial Electric Market Shares 3.- 12 “- 3.3. EUI Values Energy Use Index values (EUIs) measure annual energy use per square foot of floor space for a particular space type, end use, and fuel. Along with floorstock and market share, the EUI values form the core of the COMMEND model. Together with capital cost and tradeoff data, discussed below, EUI estimates define the engineering boundaries within which the model operates. Commercial EUIs are difficult to develop, for two reasons: Because they are end-use specific, EUIs cannot be measured from market data such as utility bills. They can only be measured by installing expensive metering equipment.’ Because they are measures of energy use--as opposed to power demand--EUIs cannot be easily developed from a simple inventory of nameplate kW ratings such as that provided by a thorough on-site survey. 3.3.1 EUI Estimation Procedure We used a combination of the following data and techniques, listed in order of importance, to develop the Railbelt EUI estimates used in COMMEND: on-site survey data mail survey data engineering calculations existing estimates from other regions of the United States building simulation modeling educated guesswork We began the EUI estimation process by computing average energy intensities (overall building kWh/Ft’) for each building type from the on-site and mail survey data. These data are useful starting points because they are the only control totals available at the building type level. (The next available level of control totals is aggregate total sales.) Table 3.3 presents these EI estimates. Calculations based on the on-site survey data records were performed to determine the likely distribution of the total EIs across end uses. The survey provided direct measurement of installed kW by end use. Technical utilization rates were judgmentally estimated for both occupied and unoccupied time periods. For some end uses, such as lighting, this was a straightforward exercise as the survey provided operating hours data. For others, such as hot water, the load factor during operating hours is far less than one. In these cases we used engineering judgment and refered to metered end use studies such as Cleary 1986. To compute the EUI values for noncompetitive uses (all except space and water heat), we applied the distribution from step 2 to the Els determined in step 1. "Several commercial end use metering projects have been undertaken during the past 5 years. Last year Bonneville Power spent over 1 million dollars metering 10 commercial buildings. See Synergic Resources, 1986 for an overview of metering projects and Cleary, 1986 for an in-depth look at metered data from Seattle. aie 13 << SSSSS Estimated Electric Intensities for the Railbelt Region WITD ONSITE AREA STD AVG CALC ANNUAL 289,584 8.7 2,148,386 119,818 8.6 . 461,300 9.2 22.5 23,511,017 1,043,858 3.9 22.1 20.2 120,100 26.6 28.6 1,417,141 49,616 15.3 33.1 43.9 61,000 2.3 15.6 9,365,821 602,192 7.5 15.5 31.4 84,600 8.5 17.2 488 , 809 28,497 11.4 14.8 17.7 5,200 7.0 41.6 3,676,670 88,298 20.4 42.0 70.4 476,052 7.9 7.1 1,106,042 154,837 4.5 7.8 22.5 55,640 11.4 17.6 503,842 28,665 13.4 17.5 13.5 28.7 14,130,142 492,550 4.1 28.7 20.6 130,754 5.1 17.8 562,276 31,532 ERR 20.4 13.4 691,400 4.5 9.5 5,887,011 617,225 1.5 9.4 15.0 27.2 2,514,740 92,355 ERR 27.2 27.7 3,065,928 185,884 7.9 37.6 15,214,235 404,806 4.4 33.6 42.9 MSC 5.7 153,379 26,832 2-2 5.7 vcT 11.9 16,126 1,350 ERR 11.9 1.6 Terms: EI Electric Intensity = ANNUAL surveyed Kwh / total surveyed AREA N Useable sample size after elimination of outliers ANNUAL sum over sample points of measured Kwh AREA sum over sample points of measured or reported Ft2 STD Standard deviation of individual building Els about building-type average WTD AVG Weighted average of Onsite and Mail results; Onsite weighted by 2 to reflect better data quality ONSITE CALC Engineering calculation of EI performed off ONSITE survey data records and averaged. SRS ALAR TP PI LS I aD a aT a Sra Ui SR aE SSE TATE Table 3.3: Measured Electric Energy Intensities 4. EUI values for heat were developed from building simulation analysis of prototype buildings. Since we could not simulate every building type, we used the simulations as a benchmark to which we could scale up the relative EUI values from national data sets developed by COMMEND users.’ Water heat EUIs were taken directly from the "COMMEND-cold climate" series developed by COMMEND users. 5. We made several adjustments to the base EUI estimates for Fairbanks. We increased heating, cooking, and misc. EUIs by 10%, 30%, and 1 kWh/Ft2 respectively to account for colder weather (tempered by better insulation), lack of gas, and head bolt heaters. The cooking EUI adjustment for Fairbanks was based on the fact that we modeled cooking as a noncompetitive end use, but derived the base EUI for cooking from survey data from sites at which some gas was generally used. *McMenamin 1988, p. 5-23. 3 - 14 How many kWh/yr do "One square foot of headbolt heater" consume? The onsite survey showed an average of 1.4 watts headbolt heater capacity per Ft2 floorstock. Weather Service BIN data show 2259 hours below 0°F in Fairbanks. Assuming 50% of the heaters are on during these hours each operating with a 50% duty cycle implies annual consumption of .8 kWh per Ft2. 6. Finally, we considered the relationship of new building EIs to the stock average. Simple regressions of the form log(EL,) = C, + a;*t, where t = year built and j indexes building types, (3.2) were performed on the on-site survey data for each separate building type j and for the pooled sample of all building types taken together. The estimated a, gives the trend rate of change in energy intensity since 1970. The pooled sample trend coefficient was -.0135, implying an average annual decrease in EI of 1.35% since 1970. Since market shares appear to have risen and fallen during this time period, we felt that the data supports a modest reduction in EUI values for new buildings. We performed this reduction on an end use basis, increasing the EUI in the cooking and miscellaneous end uses to reflect obvious trends in office and kitchen automation. 3.3.2 EUI Estimates Tables 3.4 and 3.5 show the final EUI estimates input to COMMEND. Multiplied together, the estimated electric market shares and EUI values developed above yield a complete set of electric Energy Intensity estimates broken down by end use, building type, and region. These four "market snapshots" are presented in Figures 3.9 through 3.12. They represent the distilled knowledge gained from the end use survey data. 3 - 15 (Kwh/Square Foot/Year) Energy Use Index Values (EUIs) Southcentral Regions AVERAGE Misc Hot Water Cooking Refrig Lighting Vent Heating Cooling Building Type MANTDHARAONK NON Mace esommeconion =MNOTOKNNDONK LO on soromeraosvndoon mMownKotraraNnannte ScaedrdnidN-ddo-do eoncon-eoommanece SSoNDoK- CONDO SGOGS eoomanamooNnNmNMOAY eeNooroemNeeeoo anm+nowvananonown enter Koen ONAN NHN AMNQROKMDOOCONN SONK- OK CONC OYMeO econmmyeywnomtoooH SAAD UNANSAAAD 2 oo ~~ 88 ss © 3 om OD ° ee Cee oO ee GOO D> >a OObteersio~ ws 2 CoeoCa~ aQ—¥ ~YRO~GEHK-LOVEWE S=>Ov aod = O-GoG coe ce £20990 SOGELC aR SO a SesuhGGFOIEVCKEDS NEW EQUIPMENT Misc Hot Water Cooking Refrig Lighting Vent Heating Cooling Building Type ONTNOTONANDON AH MaeKe nue sme ooncon OO ay he Pe Baie ie Ov co Novowvandaonancooe NONOAHNTA INNA ScNccagNoNKSGoK-ScG eoncoonoonme4noo SSOadSo-coONDSOOCGCS ecoomanmoOCNNMOCA+ HHeNooK-oOeKMNK KK woenrnarnhnwrononnan ener ome oe nN tome nwoonmwnowoman Sore oer cONDOMAK OS QAOMNNHNHE MAM DAD NN Onna ONMONNAN Small Office Large Office Restaurant Large Retail Small Retail Miscellaneous Vacant Grocery Car Service Lodging Warehouse Medical School College Assembly Table 3.4: EUI Estimates, Southcentral Regions 3 - 16 Energy Use Index Values (EUIs) Fairbanks Region (Kwh/Square Foot/Year) AVERAGE Misc Hot Water Cooking Refrig Lighting Vent Heating Cooling Building Type Small Office Large Office Restaurant " MONNTODAKRONK NONO+ I o n nC 2 nS om = ° S wacon MOWnKM ODM ANNAN HK SocoaoKe qi NONeKooK-co ecokhoovoromnwaco oo S Ooco-comecCCCCSG COOCMANDMOONDNMDNOA+ reNooroemnNee-oo am+thownananonown Hunter Koen ONIN HN KMNQDOMKMDOCOCOMNN SCONK OK COND OYMeO AMNAKRDAAADN AWA O ao s wo wo wna 3 KH aaand - Car Service Lodging Miscellaneous Vacant Small Retail Medical Large Retail Grocery Warehouse School College Assembly NEW EQUIPMENT Misc Hot Water Cooking Refrig Lighting Vent Heating Cooling Building Type Small Office Large Office Restaurant ONT NOTONANDONANH YMNAMOK YK OK NCS QYON nnn hh nr anqganhr x cas} WAGON ANAW e Cae Nnonoante+aa yan SooNoo e NON-cooK-Gco eomoowonrnvenha00 oo a SSo-ScoOmMeGGGCCCSG eeoeomaNDmaOCOMMAMODY+ HeNooKoOeMNe-K-H-oo eqeanavwhenwononnn HYMN or oe nN tome noonwnowonn SCoOerK- or ooONCOMNKO wYnK CCANAWDDODDWOO BHAGOMNDOLSODON Large Retail Small Retail Grocery Miscellaneous Car Service Lodging Medical Vacant Warehouse College Assembly School Table 3.5: EUI Estimates, Fairbanks Region 3-17, BUILDING TYPE Small Office Large Office Restaurant Large Retail Small Retail Grocery Warehouse Car Service Lodging Medical School College Assembly Miscellaneous Vacant Average Electricity Intensity Estimates Anchorage Region, 1987 \ MMM HEAT 4] COOK 20 30 El ESTIMATES (KWH/SQ FT/YR) END USE cooL | _ REFR Figure 3.9: 1987 Anchorage Commercial Electric Sales Structure 3-18 VENT CITE WATR J MISC Average Electricity Intensity Estimates Fairbanks Region, 1987 BUILDING TYPE Small Office Large Office Restaurant Large Retail Small Retail Grocery Warehouse Car Service Lodging Medical School - College : || Assembly Miscellaneous N Vacant 1 ! 1 ! 0 10 20 30 40 50 El ESTIMATES (KWH/SQ FT/YR) END USE MMM OHEAT coo. |_] vent ZZ watrR REFR LITE Ii] misc Figure 3.10: 1987 Fairbanks Commercial Electric Sales Structure 3-19 Average Electricity Intensity Estimates BUILDING TYPE Small Office Large Office Restaurant Large Retail Small Retail Grocery Warehouse Car Service Lodging Medical School College Assembly Miscellaneous Vacant Kenai Region, 1987 0 10 20 30 40 El ESTIMATES (KWH/SQ FT/YR) END USE Mm Heat) (W Coot COOK REFR Figure 3.11: 1987 Kenai Commercial Electric Sales Structure oF 20 VENT LITE Willa ‘NTR MISC Average Electricity Intensity Estimates Matsu Region, 1987 BUILDING TYPE Small Office Large Office Restaurant Large Retail Small Retail Grocery Warehouse Car Service Lodging Medical School College Assembly Miscellaneous Vacant | 1! i | 10 20 30 40 El ESTIMATES (KWH/SQ FT/YR) END USE MMM OHEAT coo. [|] vent ZZ water REFR 3 LITE (0 misc Figure 3.12: 1987 MatSu Commercial Electric Sales Structure 3-21 3.4 Technical Parameters 3.4.1 Technical Tradeoff Elasticities The possibilities for saving electricity by increasing the efficiency of new equipment are characterized in COMMEND by a set of technology curves. These curves are described for each end use by three parameters: ¢ Base EUI level, equal to the EUI as developed in section 3.3 above ¢ Upper and lower percentage bounds on achievable efficiency improvements e Technical tradeoff elasticity, equal to the percentage decrease in EUI from a 1% increase in capital cost. The energy efficiency of new equipment, in both existing buildings and new buildings,’ is determined in COMMEND by choosing points from these curves. For simplicity and conservatism, retrofits of equipment before their normal retirement and replacement are not allowed in COMMEND. The points chosen are a function of equipment life-cycle cost.” In this section we briefly discuss several commercial sector technologies which form the foundation for these technical tradeoff relationships. The actual parameters are taken from EPRI’s COMMEND default data set which has been under development for the past 5 years. Two extremely detailed and up-to-date assessments of individual efficiency opportunities are available in Barakat Howard and Chamberlin, 1988, and McDonald, 1987. It is important to note at the outset that two of the most common opportunities for saving electricity--air conditioning improvements and efficienct heat pumps--are unlikely to signifcantly affect Railbelt load. Cooling load is low in Alaska and heat pumps are generally believed to be uneconomical for the foreseeable future, for two reasons. First, Alaska’s annual ambient temperature regime guarantees that heat pump performance will be "poor to nil" just when the heat is most needed, implying that other heat sources would be necessary. Second, our reduced need for cooling greatly reduces the amount of heat pump capital cost which could be charged against the cooling cycle. Heat can be saved through improvements to the thermal envelope and attention to energy management. Assumed electric tradeoff elasticity: 1.0. Gas and Oil elasticities are higher (equal to 1.5) reflecting the oportunity for furnace efficiency gains. Cooling energy can be saved by the use of economizers to back out direct chilling and improved glass reflectance, in addition to more efficient chillers and compressors. "The current version of COMMEND (version 3.0) does not distinguish strongly between replacement equipment for existing buildings and new equipment for new buildings. If the inertia parameters (discussed in section 3.5.4) are greater than zero, however, there is a difference in the equipment EUI that gets installed. The strength of the difference is set by the interia parameter. '°COMMEND does not simply minimize life-cycle cost, since empirical observation has for a long time shown that electric consumers do not fit into life-cycle cost-minimizing straightjackets. In a nutshell, equipment choice equations are calibrated to reflect initial observed EUI levels, and future changes in EUI levels are computed based on changes in LCC from the base level and an assumed elasticity of responsiveness to life-cycle cost. For details see McMenamin, 1987. 3 22 Economizers are required in new buildings under ASHRAE standard 90, which the State of Alaska is considering adopting as a building standard for state-funded buildings. Current research is progressing rapidly on variable speed drives and improved heat exchangers, each of which could further increase the stock average efficiency of active cooling equipment by ~25% (Chiu, 1987). Some concern has been raised over the possibility that future efficiency gains in this area will be attenuated by limitations on the use of chloroflourocarbons. Only large centrifugal chillers currently use the chemical of concern, CFC-12. Small building HVAC systems which form the vast majority of Railbelt installations use the far less harmful and unregulated CFC-22. (Statt 1988). Assumed electric tradeoff Elasticity: 1.5. Ventilation energy can be saved most easily through energy management systems and the use of Variable Air Volume (VAV) systems, which respond to changes in the heating and cooling load by changing the volume of conditioned air, instead of changing the temperature of a fixed volume of air. Within the VAV technology class, significant efficiency gains can be realized through the use of adjustable speed drive motors instead of the more traditional inlet vane controls." These drives have only recently become available, so their use is not reflected in base EUI calculations. VAV installations currently make sense only in buildings larger than 10,000 Ft2 or 10,000 cubic feet per minute (cfm). Adams, Morgethaler & Co.” notes that in some cases their apparent savings are not realized because people demand a certain minimum volume airflow (~.5 cfm). However, VAV systems are currently installed in several larger Anchorage buildings and should continue to experience cost reductions and efficiency improvements with overall technical progress on variable speed drives. Assumed tradeoff elasticity: 1.5. Water heating efficiency can easily be improved in the electric configurations commonly seen in commercial settings by installing higher levels of insulation and water heat traps, since both of these measures directly reduce the standby losses which consume substantial energy in these low use settings. Assumed elasticity: 1.25 for electric systems; 1.5 for fossil, again reflecting possible improvements to combustion efficiency. Possibilities for improvements in commercial Cooking equipment are more limited than in other end uses. Better oven insulation, the adoption of the bi-radiant oven, and the use of higher conductivity cooking vessels can improve the efficiency of this end use. Finally, since we are modeling the cooking sector as if it had a 100% electric share (see section 3.2), the possibilities for switching to gas are partially subsumed by this parameter. Switching to gas does involve a higher up-front capital cost in exchange for saved electricity, so there is nothing conceptually wrong with this construction. Assumed tradeoff elasticity: .8. Dramatic improvements in Refrigeration have recently been effected in new Anchorage supermarket installations. Savings of ~30% relative to stock average existing installations have been effected through a combination of parallel rack compressor systems with floating head controls and efficient defrost systems. These improvements add ~25% to the cost of the compressor unit, but the incremental cost measured relative to the entire installation "Graham AC adjustable Fan Drive” product information supplied by Adams Morgenthaler & Co. shows a typical VAV/variable speed drive system using 1/2 the kWh of an inlet vane VAV system and 1/3 the kWh of a constant volume system. The Graham drives are distributed locally. 12 etter from David Crews, PE to Steve Colt, 16 August 1988. 3 - 23 (including cases) is only 3-5%”. These figures imply tradeoff elasticities of between 1.2 and 6.0 under favorable conditions. For conservatism, we retain the COMMEND default value of 1.0. Lighting electricity consumption can be cut in half in typical standard flourescent installations such as those that dominate the Railbelt. Figure 3.13 shows the 1988 distribution of Railbelt interior lighting energy by technology type. High efficiency flourescent lamps currently provide only 10% of the flourescent light energy. Interior Lighting Energy Railbelt Commercial Sector Lighting Type Quartz/Incandescent Fluorescent Hi-Efficiency Fluor. Mercury Vapor Metal Halide High Press. Sodium 4 1 i 4 0% 20% 40% 60% 80% 100% % of Interior Lighting Energy Figure 3.13: Railbelt Lighting Energy by Lighting Technology Type Lighting energy savings can be realized in a variety of ways through the careful combination of lamps, ballasts, fixtures, and controls. The real world abounds with unique savings opportunities and unique constraints. As an example of the latter, the fixed size of lamps and fixtures can impose "integer constraints" on installations in small spaces. Design calculations could call for 3.5 fixtures in the space, but the choice is between 3 and 4. To ensure that the tradeoff elasticity reflected reality with reasonable accuracy, Adams Morgethaler & Co. completed actual engineering design calculations of lighting cost and energy useage for a variety of actual building applications, using prevailing Alaska labor costs and materials prices. We then computed the technical tradeoff elasticities implied by these calculations. A sample calculation performed for a 1200 Ft2 small office is shown in table 3.6. On line 1 we compute the base system costs for a standard lamp, standard ballast, Mike Andrews, Operations Manager of Hussmann Refrigeration Anchorage, kindly provided data on this installation. 3 - 24 standard fixture installation. On line 2 a 34 watt wattmiser lamp is substituted for the standard 40 watt lamp. On line 3 a wattmiser plus lamp is substituted. The implied tradeoff elasticities (measured relative to the base system) are 5.2 and 7.2. In general, tradeoff elasticities fell between .5 and 5. Of course, only the "best buys” in systems should be considered part of the technology curve, while this exercise considered a mechanical combination of different possibilities. n COMMEND there is also a necessary averaging over various component lifetimes which renders the analysis less transparent. We are confident that the assumed tradeoff elasticity = 1.5 is a conservative estimate of the possibilities for lighting efficiency improvements. Total Implied Direct Direct Direct Labor Install Tech Watts Lamps Materials Install Cost Cost Cost System Elast Per Per Per Per Per Per Per EUI From Fixture Ballas Lamp Lamp Fixture Fixture Fixture Lamp Lamp FT2 Kwh/FT2 Base ee a Pasa a eee eee 2x4 Lens STD F40 CW 40 | 4 48.50 |i ee si i “es | NA 2x4 Lens STD F40 CW WM 34 4 48.50 52.30 2.70 1.00 1.74 5.95 5.23 2x4 Lens STD 40 CW WMPLUS 32 4 48.50 52.30 2.95 1.00 1.75 5.32 °7.25 Table 3.6: Sample Lighting Tradeoff Elasticity Calculation Miscellaneous End Uses have an assumed tradeoff elasticity of 1.0. 3.4.2 Energy Management Systems Energy Management and Control Systems (EMCS) are a new technology which shows signs of rapid market growth in new construction. Even in retrofit situations, typical EMCS systems using direct digital control technology (DDC) can achieve 10-20% savings on lighting and HVAC energy at a cost of ~$1.50/Ft2.". The Anchorage School District installed EMS equipment throughout its system during 1985-86 which contributed to a decline in electricity and gas consumption of ~20%." Several large Anchorage buildings have EMS systems controlling both HVAC and lighting. In one Anchorage case, a large medical facility was constructed without DDC control, but retrofit within one year of construction. New buildings greater than 50,000 Ft2 can design in DDC systems at an incremental cost that approaches zero as building size approaches 100,000 Ft2. These strong economies of scale derive from the increased complexity of the required status quo in large buildings. While EMCS systems are claiming a substantial share of the new construction market, they can be costly to retrofit and are not likely to penetrate the existing building stock. They are “adams Morgenthaler & Co.estimate of actual costs incurred for past installations in typical school and office installations conducted during the past two years, including "soft" costs of engineering and design. Steve Mullen, ASD, personal communication, and ISER analysis of ASD utility bills for a random sample of school buildings. 3-25 inherently difficult to model in the COMMEND framework because their cost and performance is a function of the type of HVAC and lighting hardware already in the building, its layout (open vs. closed space), and the operating hours of the building. EMCS systems are implicitly addressed in our forecast because several of the newer, larger buildings in our on-site survey use them. Hence, they contribute to the decline in EI over time, which we observed in regression analysis (section 3.3.1), and are incorporated in the lower EUI values used as inputs for new construction. 3.4.3 Federal Ballast Standard Beginning in 1990, Federal standards promulgated under the National Appliance Energy Conservation Act of 1990 prohibit the sale of standard magnetic ballasts for fluourescent lights. The imposition of the standard is modeled in COMMEND as a one-time downward shift in the technical tradeoff curve to position the "base EUI" at a point reflecting the savings induced by the standard. The shift is 8.8%, calculated as S * F * I where S= percent savings = .11; F = fraction of interior lighting energy consumed by flourescent systems = .84; I = fraction of lighting energy that is interior lighting = .96. 3.5 Economic Variables 3.5.1 Energy Prices Three levels of potential electricity, gas, and oil prices were developed for use in the critical assumptions probability tree (section 4.1). The projections are based on the Low, Middle, and High crude oil price scenarios developed by ICF (1988) and adopted by the Power Authority. The electric price forecasts shown in Figure 3.14 were developed by the same method used for residential prices (section 2.6.1)."° Figure 3.15 shows retail commercial natural gas and fuel oil prices. Price growth rates may be found in Table 4.12. For price projection details see "Retail Electric Price Projections" memorandum of 19 August from S. Colt to R. Emerman, distributed to Railbelt Utilities and available from APA or ISER. 3 - 26 “LEAS uolsey Aq sadig dDI[q [BIOISWUIWIOD [IRIEeY pojoefolg :pT"¢ aN3Ly 0.14 0.12 0.1 oO Commercial Electricity Price Projections ANCHORAGE 1987 $ per KWH eR eR eR 1987 1990 2000 Year —— Low -—+~Medium ~—* High KENAI 1987 $ per KWH I 87 1990 1996 2000 Year —— Low —+~ Medium —*— High 1987 $ per KWH FAIRBANKS 1987 TT ow: 1987 $ per KWH —- Medium —* High MATSU 0.1 0.08 0.06 0.04 0.02 oO 1987 1990 Commercial Gas Price All Regions 1987 $ per Mcf 0.00 1987 1990 Commercial Oil Price All Regions 1987 $ per Gallon 1.60 1.40 1.20 1.00 0.80 "0.60 0.40 0.20 0.00 1987 1990 Figure 3.15: Projected Retail Commercial Natural Gas and Fuel Oil Prices, 3 Cases 3 - 28 3.5.2 Discount Rates COMMEND employs a distribution of real consumer discount rates to represent customer preferences toward capital/operating cost tradeoffs. These rates correspond to required payback periods of ~1.5-5 yrs on projects with lifetimes of 5-30 years. They are consistent with both national level survey data” and our Railbelt Commercial Discount Rates used in COMMEND Population Fraction 25 2 -25 “2 Discount Rate 15% 25% 33% 50% ALLELE LAOS OIE TS PE LEG OD EOE AOC HR Table 3.7: Commercial Discount Rates commercial mail survey. Figure 3.16 shows the distribution of required payback periods obtained from the mail survey. The frequency distribution comprises 144 numerical responses. In addition, 281 respondents” did not mark payback as a basis for decisions and 123 said energy investment decisions were not made within their firm or business. These Statistics strongly echo nationwide findings that many firms which pay electric bills are divorced from decisions about energy efficiency investments and that many consider electricity to be a trivial factor cost not worthy of management’s scarce time. Required Payback Periods Figure 3.16: Required Payback Periods. Railbelt Commercial Customers Responses Percent 50 - 30 40> r 25 30 5 + 20 20 4 15 10 1075 — m 5 0 T T T T T T T z T T T 0 1 2 3 4 5 6 7 8 10 15 Required Payback, Years ISER 1988 Commercial Sector Mail Survey Caution: See text for interpretation "eg Alliance to Save Energy 1987, California PUC 1985, Easton Consultants 1987. In one study of investments actually made, mean payback periods of conservation measures actually installed are tightly clustered around 3 years. (McMenamin 1988b) SThe survey sample was drawn from utility customer records and the survey instrument sent to commercial billing addresses with the "usual" instructions to route it to a person knowledgeable about company energy use. 3/129 3.5.3 Miscellaneous Equipment Growth It seems clear that the use of computers and other office and workplace automation equipment will continue to grow during the next two decades”. At the same time, design improvements could substantially lower the EUIs for specific machinery. Standard 1987 vintage desktop computers draw ~ 100 W, but laptop models using CMOS chips and LCD displays require only 10-30 W. It is important to remember that EUIs in this sector will be determined by national economic and electricity price trends. With electricity priced at $.12/kWh, as it is in many large urban centers, a desktop computer operating 8 h/day @ 100 watts costs ~$25/yr to run without including the cost of associated cooling which is significant in summer peaking urban centers. Certainly electricity consumption is not a major factor influencing the course of computer technology, but the elimination of heat- producing components arguably is, since heat buildup limits the durability and reliability of integrated circuits. : * TL a ET Ls: For this forecast, we conservatively assume that * * Bui ldi Rat Limit the amount of all miscellaneous equipment ss Een (not just computers) continues to grow at SMALL OFFICE 3.0% 200.0% annual rates ranging from .5 to 3% throughout = —FARSE, DFFice a the forecast period, subject to absolute caps on LARGE RETAIL 1.0% 115.0% rowth which vary by buildin 2: SMALL RETAIL 1.0% 115.0% 8 ns y 8 ‘YP GROCERY 0.5% 110.0% WAREHOUSE 2.0% 130.0% AUTO SERVICE 2.0% 130.0% i, LODGING 3.0% 150.0% 3.5.4 Consumer Inertia Seeee vomteenatedoa SCHOOL 0.5% 110.0% Consumer inertia factors are used to represent COLLEGE 0:55 eee ee ce | k f inf . * d risk ASSEMBLY 0.5% 120.0% ack of information, perceived risk, unproven MISCELLANEOUS 0.5% 120.0% technology, and force of habit in the VACANT 0.0% 0.0% commercial energy market. COMMEND calculates new building EUI values each forecast year. These values are assigned to any Table 3.8: MISC equipment growth rates buildings built in that year. However, EUIs for the replacement equipment installed in existing buildings are computed as a weighted average of the existing building EUIs and the new building EUIs. The inertia factor determines the weights: COMMEND input file XEINT.EDI, screen ED8& EUIpepiuce = (Inertia * EUI,,).+ ((1 - Inertia) * EUI,..) (3.2) An inertia value of 1.0 indicates complete resistance to efficiency changes, while a value of 0.0 indicates complete acceptance. We developed two sets of inertia factors as part of the critical assumptions about future discount rates. The Base case inertia scenario assigns a value of .5 to all building types and end uses. The High case scenario assigns a value of 0.0 to all building types and end uses. Brookhaven National Labs (1987) projects a nationwide ‘oral sales increase of 240% in this end use by 2010. 3 - 30 3.6 Model Calibration The COMMEND model produced initial forecasts significantly (~20%) less than estimated control totals in the Anchorage, Fairbanks, and MatSu regions. The initial Kenai forecast was high by ~ 10%. Further research into the Kenai building stock reduced this discrepancy to +2%. In the other regions it is impossible to determine the extent to which the discrepancy is due to inaccurate control data, inaccurate inputs, or a combination of the two. This fundamental ignorance cannot be resolved without additional research into floorstock levels, EUI values, and miscellaneous electric loads unrelated to the building stock. 3.6.1 Development of Control Totals The COMMEND calibration process consisted of two steps: (1) developing control total sales data and (2) developing adjustment factors for each region to make calculated 1987 sales commensurate with control totals. Commercial control totals cannot be developed with great precision due to the confounding influences of "nonbuilding" consumption and residential consumption metered through master meters. Very little, if anything, is known about the magnitude of either of these quantities. Our development of control totals is shown in Table 3.9. The key items of uncertainty are industrial sales and sales for residential use through master meters. Commercial Control Totals Data Item CEA MLP GVEA = FMUS HEA SES MEA 1987 Comm. sales Gwh 420 614 226 89 258 22 143 Less:"Industrial Load" 30 2 65 0 137 20 1 Equals: Bldg-based 390 612 161 89 121 2 141 Less:Residential Use 10 10 1 1 5 0 5 Exterior Lights 4 6 2 1 1 0 1 (net of street) Equals: COMMEND Load 376 596 159 87 115 2 135 Control Total ET a ae aE a TE ANT GAIT TTT Tt A EIT LEAL SNE AN ARTI Table 3.9: Development of Commercial Control Total Sales Data Industrial sales are known only for major installations. Sales to light industry are impossible to determine without a comprehensive analysis of utility customer files and additional survey work. Commercial metered sales for residential use are an equally significant and unknown quantity. We canvassed several large apartment complexes and turned up several instances of master metering, so we know the phenomenon still exists. Similarly, Railbelt utility staff are aware of single examples of master metering, including some to all-electric complexes. The one residential complex for which MLP has data consumes ~5 GWh/yr. The estimates 3-31 shown in Table 3.9 are based on the residential calibration worksheets, which set an upper bound on the amount of electric heat consumed by multifamily housing. The worksheets, in turn, are based on the residential end use survey. 3.6.2 Final Adjustment Factors The final adjustment factors needed to ensure that calculated 1987 COMMEND load was equal to the control total for 1987 are: e Anchorage 1:23 e Fairbanks 1.27 e Kenai 0.98 e« MatSu 1.15 Although the adjustments are significant for Anchorage and Fairbanks, they are not unreasonable in light of the fact that the control totals developed above are probably high while the floorstock estimates, developed by actual count, must be low. There are several places where the required adjustments could be made so that the model would "appear" to reproduce 1987 consumption exactly. Any of the following could be adjusted: (1) Control totals (2) EUIs in any building type and any end use, particularly miscellaneous (3) Market shares in any building type and any end use (4) Floorstock in any building type Additional data would allow some insight into which of these factors is the key source of discrepancy. For example, detailed data on gas sales would allow investigation of the hypotheses that the electric market shares are low. We could compare calculated gas sales from COMMEND with control totals by building type; if the former exceed the latter, we have good evidence that gas shares should be reduced and electric shares increased. Lacking such data, we applied the adjustments called for above directly to total sales figures after they were produced by COMMEND. Since all the input assumptions described above are best estimates based on the data at hand, direct adjustment of sales is the least biased method of benchmarking the model to 1987 data. As further data become available, it should be possible to move from general to specific adjustments to the input parameters. 4. ELECTRICITY DEMAND PROJECTIONS In this chapter we present projections of electricity demand by customer class, end use and region. We first discuss how the various critical assumptions influencing the end use demand forecasts were combined and probabilities were computed to produce representative LOW, MIDDLE, and HIGH projections of residential and commercial sales by end use. We then present three industrial demand forecasts consistent with the LOW, MIDDLE, and HIGH end use projections. By adding in estimated street light sales, utility premises use, and distribution losses we arrive at projected total energy requirements for these three cases. We then present projections of peak load based on historical load factors. Finally, we briefly compare the results obtained here with two recent utility load forecast results produced with econometric models. 4.1 Treatment of Uncertainty 4.1.1 Description of Critical Assumptions Railbelt electricity consumption is primarily a function of the number of households and the level of employment. The following other factors which cannot be determined with certainty were also suspected of being important determinants of future electricity consumption in the Railbelt: energy prices consumer discount rates technological change natural gas market penetration Three employment/household cases are taken directly from the ISER report on Economic and Demographic Projections for the Railbelt (Goldsmith 1988). Each one of these cases represents one-third of the total distribution of employment/houesholds outcomes. Each employment/household case was used to generate one projection of occupied housing stock and commercial floor space. These projections have been described in chapters 2 and 3. Three energy price scenarios consist of projected consumer prices for electricity, natural gas, fuel oil, and propane for each region of the Railbelt consistent with the 3 world oil price scenarios developed for APA by ICF and used as critical assumptions in the economic and demographic assumptions. These world oil price cases were the Low, Low Consensus, and Consensus which were assigned probabilities of 60%, 30% and 10%, respectively, by the APA Board. In each case the prices of the alternative energy types vary as a function of the price of crude oil. The three sets of price projections have been described in chapters 2 and 3. Each of the individual economic and demographic projections has a world oil price assumption associated with it. Consequently, it is possible to calculate the joint probability of each of the 3 energy price scenarios associated with each of the 3 employment/household cases. This calculation produces the joint probabilities shown in Table 4.1. Joint Employment/Household and Energy Price Probabilities Employment/Household Case Energy Price Level Joint NAME 1.D. — PROBA- CASE PROBABILITY Proba BILITY bility LOW 91 30 LOW APA115 33% LOW CONSENSUS -09 -03 CONSENSUS -0 0 LOW -70 23 MIDDLE APA74 33% LOW CONSENSUS 25 09 CONSENSUS -05 01 LOW +18 -06 HIGH APA10 33% LOW CONSENSUS 56 19 CONSENSUS 26 09 Table 4.1: Joint Probabilities for Employment/Households and Energy Price Scenarios Two sets of Consumer Discount Rates are modeled. The choice of new and replacement appliances is partially dependent upon the consumer discount rate. For a given discount rate a consumer seeks to minimize the life-cycle cost of the appliance being purchased or the retrofit being considered. A higher discount rate means that the operating cost portion of the life-cycle cost of an appliance will receive less weight in the choice. As a consequence, a consumer with a high discount rate will choose an appliance with higher operating costs relative to initial cost and, consequently, with a higher lifetime consumption of energy. e Alternative 1: (Probability 80%) assumes a distribution of consumer discount rates consistent with historical patterns. Observed and inferred consumer (including commercial consumer) discount rates for appliance purchase decisions are extremely high compared to rates on other investments and borrowing opportunities. As discussed in section 2.6.2, this is a reflection of some combination of uncertainty, imperfect information, and cash flow constraints. e Alternative 2: (Probability 20%) approximates a case in which consumers act as though . discount rates are lower than in Alternative 1 by about 25%. In the residential sector, this alternative is modeled by a direct reduction in the consumer discount rate. In the commercial model the alternative is modeled by reducing the consumer inertia factors (see section 3.5.4) from .5 to zero. This drop in discount rates assumes either that consumers will become more aware of the life-cycle costs associated with energy investments or that they will perceive efficiency less risky. It also allows for the possibility that institutions other than the State--the federal government, utilities, private businesses, or non-profit organizations--will institute programs which reduce perceived risk, provide better information, or reduce cash flow constraints associated with appliance purchases for residential and commercial customers. 42) Technological change can take many different forms having different implications for electricity consumption: ¢ Over time technological change may reduce the energy use associated with an appliance independent of its other performance characteristics. This would generally reduce energy use as these new appliances replaced the existing stock. This alternative is mathematically close, but not perfectly equivalent, to a reduction in the incremental cost of a more efficient appliance. (Example: refrigerators, 1973-1986) e Technological change could result in a reduction in the initial cost of an appliance independent of its other performance characteristics. This could increase or decrease energy consumption, depending mostly on the distribution of the cost reductions among models of different efficiencies and somewhat on the degree to which consumers spend the "income effect" of lower purchase price on increased usage of the appliance. (Example: color televisions: 1960-1980) e Technological change could result in the development of appliances to serve new functions. This would increase consumption. (Example: personal computers: 1979-1983) e Technological change could result in the development of new appliances for doing existing jobs. This could increase or decrease consumption. (Example: microwave ovens: 1960-present) We characterize technological change in this analysis as the first alternative: a reduction in the marginal cost of saving a kilowatt-hour of electricity by purchasing a more efficient appliance. e Alternative 1 ((Base); probability 40%) No technological change is assumed except as incorporated into national efficiency standards which are currently programed to take effect. e Alternative 2: ((High); probability 60%) Commercial Sector: Incremental appliance operating efficiencies improve according to a linear trend averaging 1% annually, with this trend temporarily overridden when federal ballast standards take effect in 1991. Residential Sector: The incremental price of more efficient appliances is reduced by 20%. Southern Railbelt Gas Market Penetration. Further penetration of natural gas in the Southern Railbelt energy market could reduce the growth rate of electricity consumption if electricity use were "backed out" by such penetration. In order to calculate the effect that this might have on Railbelt electricity use, a spreadsheet model was constructed which projected (1) the number of electrically space heated residential units which would convert to natural gas as the distribution system for gas expanded (including the backlog of potential conversions represented by all currently electrically heated residential units in areas served by natural gas), and (2) the proportion of new residential units which would choose electric space heating. This model produced vectors of conversions and proportions for each of the Southern Railbelt regions under each of the following alternatives. e Alternative 1. ((Base), probability 50%) No significant extension of the gas distribution system, but conversions of existing heating systems to gas continues at historical rates. For details, refer to section 2.2.5. ¢ Alternative 2. ((High), probability 50%) The gas system is extended to the remote MatSu Valley, Girdwood, and Homer. Conversion of multifamily buildings is accelerated in Anchorage. For details, refer to section 2.2.5. 4.1.2 Probability Tree Methodology The combination of 3 possible economic/demographic assumptions, 3 different energy price assumptions, and 2 different assumptions each for consumer discount rate, technological change, and natural gas market penetration results in a total of 72 possible sets of electricity consumption projections, with each set consisting of a projection for residential and commercial use for each of the 4 regions of the Railbelt. We generated a representative sample of these 72 sets using the end-use models and interpolated the other sets from this sample. To each set we added one of three industrial load projections consistent with the Low, Middle, or High economic/demographic case. We also added to each case an estimate of miscellaneous load--street lights and utility premises use--and distribution losses. The result was 72 projections of Railbelt electricity requirements which were ranked from lowest to highest based upon total Railbelt electricity demand in 2010. We calculated the probability of each of the 72 projections from the joint probabilities of the S assumptions which varied from case to case. The distribution of these probabilities, contingent upon the assumptions for the 5 critical inputs, is shown in Figure 4.1. After these probabilities were assigned to each of the projections the cumulative probability distribution was calculated. Three points on this distribution were chosen to represent the total distribution of possible outcomes. These cases are the following: LOW case 16.5% probability that electricity use in the Railbelt in 2010 will be less than this amount and 83.5% probability that electricity use in the Railbelt in 2010 will be greater than this amount (contingent upon the assumptions for the 5 critical inputs) MIDDLE case 50% probability that electricity use in the Railbelt in 2010 will be less than this amount and 50% probability that electricity use in the Railbelt in 2010 will be greater than this amount (contingent upon the assumptions for the 5 critical inputs) HIGH case 83.5% probability that electricity use in the Railbelt in 2010 will be less than this amount and 16.5% probability that electricity use in the Railbelt in 2010 will be greater than this amount (contingent upon the assumptions for the 5 critical inputs) The specific combinations of the 5 critical assumptions which give rise to these representative cases are shown in Table 4.2. Note, however, that any given amount of 4-4 electricity consumption could be the result of any one of several different possible combinations of the critical assumptions. For example, a given load forecast could result from either high economic growth with high gas penetration or low economic growth with low gas penetration. Furthermore, the fact that all three representative cases happen to be based on "high" gas penetration does not mean that "low" gas penetration was not considered. The representative cases summarize a distribution of outcomes which is determined by combinations of all possible values of critical assumptions. Critical Assumptions Behind the LOW, MID, and HI Forecast Cases Case Name LOW MEDIUM HIGH Case Components: Econ/Demographic Low Middle High Energy Price Low Low Middle Technological Change Base Base Base Discount Rate Hi Hi Hi Table 4.2: Critical Assumptions behind the LOW, MIDDLE, and HIGH forecast cases 4.2 Residential Sales 4.2.1 Narrative The story behind the residential sales forecasts is straightforward: e Occupied housing stock grows at between 1.1 and 2.1% per year. e Space heat market share remains steady or declines, depending on the amount of gas conversions. e Average market shares for other appliances in competitive markets generally decline as the appliance stock turns over, since the electric share of replacements is generally lower than the electric share of the average stock. In addition, the marginal (replacement equipment) shares themselves fall slightly as electric prices rise. e Average EUI values for water heat, refrigerators, and freezers fall to the level mandated by federal standards. EUI values for electric dryers and stoves remain the same. EUI values for miscellaneous use are measured on a kWh per house basis, so they increase as a reflection of the assumed exogenous increase in miscellaneous energy services per house. ¢ Utilization falls by ~5% as electric prices rise. 4-5 In essence, consumers are forced off their "demand curves" for appliances by the federal standards. They therefore do not respond further to price increases. There is no difference in purchasing behavior between those runs that feature lower projected discount rates and others in which discount rates are left at current levels. 4.2.2 Guide to Tables Residential sales summaries for the LOW, MIDDLE, and HIGH cases are presented in Tables 4.3 through 4.5. Figure 4.1 shows that the end use breakdown is similar in all three cases. Therefore, only MIDDLE case results are shown for regional sales by end use in Figures 4.2 and 4.3. Regional detail is available in Appendix C. All of the representative case results for the Kenai Peninsula reflect the extension of natural gas to Homer. A helpful way to view these results is to decompose total sales into the product of number of customers and use per customer. This decomposition is reported in Tables 4.3-4.5 and displayed on Figure 4.4. These components are reported as indices (relative to 1987=1) to allow comparison between them. These indices can also be interpreted as percentage changes. For example, Figure 4.4 shows that in the MatSu region occupied housing grows by ~58% between 1987 and 2010, while use per house falls ~30%. Together these forces cause total sales to rise ~8% from their 1987 level. We can further decompose the Energy Intensity into its components: market share and EUI. These results are presented in Figures 4.5 through 4.7 and 4.8 respectively. For numerical details, see also Appendix C. Appendix C presents more detailed results, including a breakout of new versus average market shares and EUIs. It is important to remember that these presentations are aggregated across housing types. Therefore apparent trends in shares or EUIs are often more a function of shifting housing type composition than of economic or other forces. 2000 1500F Railbelt Residential Sales, 2010 Low, Middle, High Cases 1000 | Bees: 500- 0 Railbelt Commercial Sales, 2010 Low, Middle, High Cases 2500 1000 500 2000 F 1500- GWh QW 1987 Sales 2010 (Low) 2010 (Mid) 2010 (Hi) MMM HEAT WATR (J Frie FREZ COOK (-) ory E28 Lite (i) misc GWh 1987 Sales 2010 (Low) 2010 (Mid) 2010 (Hi) COOL (“J vent lla. NTR HES LITE (i misc Figure 4.1: Year 2010 Railbelt Sales by End Use; Res / Comm, 3 Cases 4. Residential Forecast Summary Region: RAILBELT Case: LOW 1987 1988 1990 1995 2000 2005 2010 Occupied Housing Stock (000) Single 70.8 71.1 72.0 72.8 75.6 81.9 89.7 Multi 46.3 45.2 44.2 48.7 53.1 57.6 61.9 Mobile 11.5 11.5 11.9 12.7 13.0 13.5 14.3 TOTAL 128.7 127.8 128.1 134.2 141.7 153.0 165.9 New Housing Units (000) Single 0.0 0.0 0.0 0.2 a2, 1.6 2.0 Multi 0.0 0.0 0.0 0.1 0.7 120 eel: Mobile 0.0 0.0 0.0 0.0 Oor 0.2 0.2 TOTAL 0.0 0.0 0.0 0.3 Zork 2.8 S25 Adjusted Sales by End Use Heat 249 240 222 187 169 164 167 Water 195 195 186 172 168 170 180 Frig 154 52 148 144 142 142 150 Freez 82 81 81 81 13 78 80 Cook 62 62 62 64 67 71 76 Dry : 99 98 99 100 99 102 109 Lite 164 164 165 sak 180 196 215 Misc 275 273 274 284 304 336 373 TOTAL SALES 1281 1266 1235 1203 1208 1260 # 1350 Kwh per Customer 9,958 9,909 9,645 8,964 8,527 8,234 8,136 eS 1987- 1995- 1987- GROWTH RATES 1995 2010 2010 Housing Stock 0.5% 1.4% 1.1% Use per House -1.3% -0.6% -0.9% Total Sales -0.8% 0.8% 0.2% ——— INDEX VALUES (1987=1) Housing Stock 1.000 0.993 0.996 1.043 1.101 1.189 1.290 Use per House 1.000 0.995 0.969 0.900 0.856 0.827 0.817 Total Sales 1.000 0.988 0.964. 0.939 0.943 0.983 1.054 <A NRL RINNE LOE LEELA TEI ET I TIE ATE AE SESE IIE TE EI OED EL LTTE IEE, EER OEE: IEE ae aA NT DETTE Table 4.3: Railbelt Residential Forecast Summary, Low Case 4-8 LN Residential Forecast Summary Region: RAILBELT Case: MIDDLE 1987 1988 1990 1995 2000 2005 2010 Occupied Housing Stock (000) Single 10:8 71.1, 72,01, 7229 79.9 —9073=- 108.8 Multi G6n3 44°59 343.8) 51,9 57.2 | 62.7 -* 100 Mobile 15 11.5 110) 127 34 14 4 157 TOTAL 128.7 127.5 127-777 13755' 150.54 167-74. - 189558 New Housing Units (000) ‘ Single 0.0 0.0 0.0 0.4 Led) 2.6 Srl Multi 0.0 0.0 0.0 0.2 eid, Les 1.8 Mobile 0.0 0.0 0.0 Oe 0.2 0:3 053 TOTAL 0.0 0.0 0.0 0.6 Sto) 4.4 5.2 Adjusted Sales by End Use Heat 249 240 222 195 180 185 201 Water 195 195 185 176 173 182 199 Frig 154 152 148 147 150 156 172 Freez 82 81 81 83 84 86 93 Cook 62 62 62 67 71 77 87 Dry 99 98 98 102 106 TS 130 Lite 164 163 164 175 194 227 258 Misc 275 273 273 291 327, 379 448 TOTAL SALES 1281 1264 1234 1236 1285 1401 1588 Kwh per Customer 9,958 9,913 9,657 8,987 8,544 8,367 8,365 1987- 1995- 1987- GROWTH RATES 1995 2010 2010 Housing Stock 0.8% 2.2% 1.7% Use per House -1.3% -0.5% -0.8% Total Sales -0.5% Love 0.9% INDEX VALUES (1987=1) Housing Stock 15000-02991 0.993 -*1,5069= 1-169). I 301 12476 Use per House 1.000 0.995 0.970 0.902 0.858 0.840 0.840 Total Sales 1.000 0.987 0.963 0.964 1.003 1.093 1.240 i Table 4.4: Railbelt Residential Forecast Summary, MIDDLE Case 4-9 Residential Forecast Summary Region: RAILBELT Case: HIGH 1987 1988 1990 1995 2000 2005 2010 Occupied Housing Stock (000) Single 70.6 /1.b = 7250 75.0 85.6 99-5 11353) Multi 66:3 45.1 45.6 53:2 6051 68.0 75.6 Mobile SS ae Sas a 9 a 29 13 9 153 oa) TOTAL 12857-12757 129-55 14) 1 159.65 1827-20535 New Housing Units (000) Single 0.0 0.0 0.0 2.0 Sisi2, Sie 323 Multi 0.0 0.0 0.0 1.4 Les 1.9 19) Mobile 0.0 0.0 0.0 0.2 0.4 O53 0.4 TOTAL 0.0 0.0 0.0 3.6 ono 5.4 a5 Adjusted Sales by End Use Heat 249 240 226 199 194 204 219 Water 195 195 188 180 185 199 220 Frig 154 152 150 UST 159 170 185 Freez 82 81 82 85 88 93 100 Cook 62 62 63 68 75 84 93 Dry 99 98 - 99 104 112 124 138 Lite 164 164 166 179 206 239 272 Misc 275 273 276 298 346 408 469 TOTAL SALES 1281 1265 1250 1265 1364 1520 1695 Kwh per Customer 95958 9,911 9,650 8,963: 8,550 8,319 8,249 1987- 1995- 1987- GROWTH RATES 1995 2010 2010 Housing Stock Lee 2.5% Zeke Use per House -1.3% -0.6% -0.8% Total Sales -0.2% 2.0% 1.2% INDEX VALUES (1987=1) Housing Stock 1/000 10.993) 1.007 1.097 —1:°240° 1.3420 1.598 Use per House 1.000 0.995 0.969 0.900 0.859 0.835 0.828 Total Sales 1.000 0.988 0.976 0.987 1.065 1.187 1.323 Neen Table 4.5: Railbelt Residential Forecast Summary, HIGH Case 4-10 1000 800 600 400 200 GWh Residential Sales by End Use Anchorage Middle Case (49) GWh Ky 1987 1988 1990 1995 2000 2005 2010 MM Heat Water [J Frig Freezer Cook (J Dryer HEB Lite MM Mise Residential Sales by End Use Fairbanks Middle Case (49) 350 300 250 200 150 100 50 0 MM Heat Water () Frig Freezer Cook () Dryer Lite (ll) Mise Ulddddddda ol — 1987 1988 1990 1995 2000 2005 2010 Figure 4.2: Residential Sales by End Use, Anchorage & Fairbanks, MID Case 4-11 Residential Sales by End Use Kenai Middle Case (49) GWh 200 150 100 — Essssssssssssses] 490 (Eussusssseessss) Wy) 50 . 0 1987 1988 1990 1995 2000 2005 2010 MB Heat Water (5 Frig Freezer Cook CJ bryer Lite (0) Mise Residential Sales by End Use MatSu Middle Case (49) Gwh 250 200F me | | || 100F- | | | 50- UI : N \ \\ = 0 || 1987 1988 1990 1995 2000 2005 = MM Heat MM Water (—) Frig Freezer Cook () Dryer Lite (Ml Misc Figure 4.3: Residential Sales by End Use, Kenai & MatSu, MID Case 4-12 cl - > Py O1ndi.y ANCHORAGE (1987 = 1) 1.6 1.45 0.8 0.6 0.4 0.2 1.2 a 0 1 4 1995 2000 Year Isodwi0seq YIMOIDH soles [eNUapIsoy ——— House Stock —+— Kwh per House —* Sales ‘uoly KENAI (1987 = 1) 1.6 1.4 a2 0.8 0.6 0.4 ase CIW ‘SuoIsey [IV O.27 aE, 1987 1990 1995 2000 Year —— House Stock —+— Kwh per House ~*~ Sales Components of Sales Growth Residential, Middle Case FAIRBANKS (1987 = 1) 1 1 ——— House Stock (1987 = 1) 1995 2000 Year — + Kwh per House ~—*~ Sales MATSU 4 1 ——— Hause Stock 1995 2000 Year —+~ Kwh per House 25% 20% 15% 10% 5% 0% 10% 8% 6% 4% 2% 0% Average Residential Electric Market Share, Heat, Anchorage Middle Case (49) 1987 1990 MMW Single-Family 1995 2000 Multi-Family 2005 2010 Mobile Home Average Residential Electric Market Share, Heat, Fairbanks Middle Case (49) at 1987 1990 AN) Single-Family 1995 2000 Multi-Family 2005 2010 Mobile Home Figure 4.5: Residential Electric Heat Shares, Anchorage & Fairbanks, MID Case 4-14 Average Residential Electric Market Share, Heat, Kenai Middle Case (49) ZZ | M0 1990 1995 2000 2005 2010 (MW single-Family Multi-Family Mobile Home Reflects Gas Extension to Homer Average Residential Electric Market Share, Heat, MatSu Middle Case (49) a 1987 1990 1995 2000 2005 2010 Single-Family Multi-Family Mobile Home Figure 4.6: Residential Electric Heat Market Shares, Kenai & MatSu, MID Case 4-15 9L- +h Ly emn3iy asta CW ‘suoisey [Tv ‘soreys Joye sourrddy Average Residential Electric Market Shares, Anchorage Middle Case (49) Average Residential Electric Market Shares, Fairbanks Middle Case (49) 1 1 an 4 1990 1995 2000 2005 —— Water —* Cooking —— Drying Average Residential Electric Market Shares, Kenai Middle Case (49) 1 4 4 4 ze i 1 ene 1990 1995 2000 2005 —— Water — Cooking ~—— Drying Reflects High Gas Penetration 1990 1995 2000 2005 — Water Cooking —— Drying Average Residential Electric Market Shares, MatSu Middle Case (49) 1 1 4 4 1990 1995 2000 2005 —*— Water —* Cooking —~ Drying Reflects High Gas Penetration Ll-¢ ased QIW ‘esesioyouy ‘s[(jq [eNUepIsoy UI Spud] :g*p oIN31y Residential Average EUI Index Anchorage (Other Regions Similar) Middle Case (49) 198 2000 —=— Water i —*— Freezer —— Dryer i —& Misc NB: Lite and Misc EUIs are kWh per house 4.3 Commercial Sales 4.3.1 Narrative Commercial sales per square foot decline at an average annual rate of .5 percent in all three cases. This decline is due to a combination of the following factors: e New equipment heat and water heat shares shift slightly as the price of electricity relative to gas and oil shifts. In Anchorage and Fairbanks, the relative price drops and shares increase; in Kenai and MatSu price rises and shares drop. e Average heat shares are driven down by turnover of the building stock. e New buildings and equipment with lower EUIs replace old. e Federal ballast standards drive the new equipment lighting EUI down by 9 percent. The initial difference between new and average EUI also contributes to a total decline in average lighting EUI of ~12 percent over the forecast period. e Miscellaneous end use penetration grows (more equipment per Ft2) while MISC EUI declines slightly. In spite of this decline in Electricity Intensity, total sales grows in response to a floorstock growth rate of between 1 and 2 percent. Occupied commercial floorstock grows smoothly with employment. Still, some building types do not experience net additions to the physical stock until the mid 1990s. 4.3.2 Guide to Tables Tables 4.6 through 4.8 show Railbelt commercial sales summaries. Figures 4.9 and 4.10 display sales by end use for the middle case. Decomposition into sales growth components, explained above, is shown in Figure 4.11. All results for the Kenai Peninsula reflect the extension of Natural Gas to Homer. Appendix C provides regional detail and shows the evolution of new and average EUI and market share values. Additional detail is available from the authors. 4-18 Commercial Forecast Summary Region: RAILBELT Case: LOW 1987 =1988 1990 1995 2000 2005 2010 Floorstok (Million Ft2) 81.0 81.3 81.3 85.6 88.7 94.8 101.4 Adjusted Electric Sales (Gwh/Yr) HEAT 48 47 46 44 43 38 38 cooL 66 66 66 66 67 67 70 VENT 218 218 215 213 207 206 216 WATR 33) 34 34 35) 37 39 41 COOK 40 40 41 48 54 60 65 REFR 182 181 180 185 184 186 196 LITE 724 724 709 694 659 679 708 MISC LTS 176 184 216 250 290 324 TOTAL ELECTRIC SALES 1485 1486 1474 1501 1500 1563 1657 eee ta senescence senibeianttnits - 1987- 1995- 1987- GROWTH RATES 1995 2010 2010 Floorstock 0.7% 1.1% 1.0% Kwh per Ft2 -0.6% -0.5% -0.5% Total Sales 0.1% 0.7% 0.5% sw INDEX VALUES (1987=1) Floorstock 1.000 1.004 1.004 1.057 1.095 1.170 1.252 Kwh per Ft2 1.000 0.997 0.989 0.956 0.922 0.899 0.891 Total Sales 1.000 1.001 0.993 1.011 1.010 1.053 1.116 Table 4.6: Railbelt Commercial Forecast Summary, LOW Case 4-19 Commercial Forecast Summary Region: RAILBELT Case: MIDDLE 1987. 1988 1990 1995 2000 2005 2010 Floorstok (Million Ft2) 81.0 81.3 81.0 85.6 92.6 101.3 114.0 Adjusted Electric Sales (Gwh/yr) HEAT 48 47 46 45 44 38 41 cooL 66 66 66 64 67 67 74 VENT 218 219 215 212 213 220 242 WATR 33 34 34 35 38 40 44 COOK 40 40 41 46 53 59 66 REFR 182 181 179 185 191 197 217 LITE 724 723 706 698 690 728 796 MISC 175 176 182 216 259 304 358 TOTAL ELECTRIC SALES 1485 1486 1469 1501 1555 1653 1837 1987- 1995- 1987- GROWTH RATES 1995 2010 2010 Floorstock 0.7% 1.9% To Kwh per Ft2 -0.6% -0.6% -0.6% Total Sales 0.1% 1.4% 0.9% i INDEX VALUES (1987=1) Floorstock 1.000 1.004 1.000 1.057 1.143 1.251 1.407 Kwh per Ft2 1.000 0.997 0.989 0.956 0.916 0.890 0.879 Total Sales 1.000 1.001 0.989 1.011 1.047 1.113 1.237 a Table 4.7: Railbelt Commercial Forecast Summary, MIDDLE Case 4 - 20 Commercial Forecast Summary Region: RAILBELT Case: HIGH 1987 1988 1990 1995 2000 2005 2010 Floorstok (Million Ft2) 80.6 80.6 82.1 89.9 101.0 113.7 124.8 Adjusted Electric Sales (Gwh/Yr) HEAT 48 46 46 45 44 41 43 CooL 66 66 66 68 69 75 79 VENT 217 216 215 217 227 239 256 WATR 33 34 34 36 40 42 46 COOK 40 40 42 49 57 65 71 REFR 182 180 183 195 208 220 235 LITE 722 721 718 727 744 804 856 MISC 175 176 187 230 284 342 392 TOTAL ELECTRIC SALES 1483 1478 1490 1568 1672 1830 1978 ES LOS Tei 1995- 1987- GROWTH RATES 1995 2010 2010 Floorstock 1.4% 2.2% 1.9% Kwh per Ft2 -0.7% -0.6% -0.6% Total Sales 0.7% 1.6% 13%; ee ey aEE EE EyEIESESEES SSIES INDEX VALUES (1987=1) Floorstock 1,000 , 1.000. 1,019 T5115 2.253: hk Git 71548 Kwh per Ft2 1.000 0.997 0.987 0.948 0.900 0.875 0.862 Total Sales 1000 0.1997) 1.005. 1.058" 25128) 1234 12334 Table 4.8: Railbelt Commercial Forecast Summary, HIGH Case 4-21 Commercial Sales by End Use Anchorage Middle Case GWh 1400 1200 1000 800 600 400 200 0 y 1987 1988 1990 1995 2000 2005 2010 MMB Heat Cool {J vent Water Cook (] Retr Lite {ll Mise Commercial Sales by End Use Fairbanks Middle Case GWh 350 1 300 250 200 150 Es SS 100 50 a emcees 5 1987 1988 1990 1995 2000 2005 2010 MM Heat Cool [J vent Water Cook (J Retr Lite Ml Mise Figure 4.9: Commercial Sales by End Use, Anchorage & Fairbanks, MID Case 4-22 Commercial Sales by End Use Kenai Middle Case Xe 1987 1988 1990 1995 2000 2005 2010 MM Heat Cool CJ vent CJ Retr EEA Lite Commercial Sales by End Use Matsu Middle Case | La Kom 1987 1988 1990 1995 2000 2005 2010 Ml Heat Cool [J vent Water ' Cook (_) Retr M0 Mise ‘ Figure 4.10: Commercial Sales by End Use, Kenai & MatSu, MID Case 4 - 23 vo- Ilr omnsiq (1987 = 1) 2000 Year ~ 77 Floor Stock —+~KwhperFT2 ~*~ Sales KENAI (1987 = 1) 1.2 1 0.8 0.6 aseg CIW ‘suorses [Ty ‘seTes [eIoJoWUIOD jo uoNIsodwoseq 0.4 0.2 qui 1987 Year —~ Floor Stock —+— Kwh per Ft2 “~*~ Sales Components of Sales Growth ANCHORAGE Commercial, Middle Case (1987 = 1) 1.6 FAIRBANKS 1.4 1:2 1 0.8 0.6 0.4 0.2 Oo 1987 1990 —— Floor Stock (1987 = 1) 2000 Year —— Kwh per Ft2. —*~ Sales MATSU 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 1987 1990 ——~ Floor Stock 2000 Year —~ Kwh per FT2 —*~ Sales 4.4 Industrial Load 4.4.1 Current Situation The utility-served industrial load in the Railbelt is concentrated in the Kenai and Fairbanks areas and dominated by petroleum processing, production, and transportation, as shown in Table 4.9. Petroleum refining at one refinery on the Kenai Peninsula and one refinery in the Fairbanks area accounted for more than half of the industrial consumption of electricity provided by the Railbelt utilities. Additional significant consumption was accounted for by two other refineries, production facilities, and the LNG processing plant, all on the Kenai Peninsula. General industrial use in the Anchorage and Kenai areas is the next most significant use of electricity, followed by the mining of coal in the Fairbanks area and fish processing in the Kenai area. Electricity use for petroleum-related activities has a very high load factor, in sharp contrast with the very low load factor for some of the other uses, particularly fish processing, which is a very seasonal activity. NN A RS A EDGR ASS SESS DE FS REIT I a! 1987 INDUSTRIAL LOAD BY REGION AND USE Energy Demand % * Load GWh Railbelt MW Railbelt Factor’ KENAI TOTAL 158 0.62 33 0.54 Pet Processing 129 0.50 21 0.33 0.71 Manufacturing 20 0.08 0.11 0.35 Fish Processing 9 0.03 6 0.10 0.17 FRBKS TOTAL 65 0.25 13 0.21 Pet Processing 50 0.19 7 0.12 0.77 Mining 10 0.04 a 0.05 0.36 Petroleum Trans. 4 0.02 1 0.02 0.36 Construction 1 0.00 1 0.02 0.09 ANCH Manufacturing 32 0.13 14 0.23 0.26 MATSU Construction 1 0.01 2 0.03 0.10 TOTAL 256 1 62 1 0.474 'Noncoincident SA A AC EL I RM TE ITE TS LL RE EAD eA GMD BOE EE PSR NDOD EOE ein a TP Table 4.9: 1987 Railbelt Industrial Electricity Use 4-25 4.4.2 Industrial Load Projections Three projections of industrial load served by the Railbelt utilities have been used in this analysis. Each projection is associated with and consistent with one of the three economic/demographic cases. Each projection is a combination of assumptions regarding the petroleum processing load, the trend with respect to general industrial activity, and the special projects identified in the economic projection. These projections are shown by region in Table 4.10. Detailed projections by case, region, and industry may be found in Appendix D. 4 - 26 TT SE AT ES HRT I TT DEL SNE PROJECTED RAILBELT INDUSTRIAL ELECTRICITY LOAD (GWh) 1987 1990 1995 2000 2005 2010 MID CASE : KENAI 158 140 143 146 149 152 FRBKS 65 69 70 3 75 7 ANCH 32 34 36 37 39 41 MATSU 1 2 5 7 8 8 TOTAL 256 244 252 263 270 278 LOW CASE KENAI 158 140 66 68 69 71 FRBKS 65 69 69 69 69 69 ANCH 32 34 34 34 34 34 MATSU 1 2 3 4 5 5 TOTAL 256 244 172 174 176 178 HIGH CASE KENAI 158 140 163 166 189 192 FRBKS 65 69 93 100 106 112 ANCH 32 34 51 55 59 64 MATSU 1 2 5 7 11 13 TOTAL 256 244 311 327 364 380 a LTT T TA SAAT ET lOO REE ADIL PED TEETER 8 ie ti 55 ath Table 4.10: Projected Industrial Demand by Region and Case MIDDLE Case HIGH Case LOW Case Petroleum processing demand declines on the Kenai Peninsula due to planned self generation by the processors, and increases slightly in Fairbanks due to increases in capabilities in the existing refineries. There is a general growth in other components of demand consistent with the growth of the economy in the middle case economic projection including new mining activity in the Interior and both mining activity and ski resort development in the Matanuska Valley. Petroleum processing capabilities expand in the Interior in the high case, and a new gas processing facility is built on the Kenai Peninsula. Mining activity expands more rapidly in the Interior, "homeporting" becomes a reality in Anchorage, and the expansion of mining and ski resort facilities in the Matanuska Valley occurs at a more rapid pace. General growth in other components of demand follows the pattern of the high case economic projection. Petroleum processing on the Kenai Peninsula continues its trend toward self generation and Expansion in Fairbanks is at the same rate as the MID CASE. There is no expansion of mining in the Interior and no mining demand in the Matanuska Valley. General growth in other components of demand follows the pattern of the low case economic projections. aioe 4.5 Street Lights, Office Use, Losses Energy requirements for street lights, office use, and distribution losses' are projected as simple Additional Requirements, 1987 Estimated GwWh levels and % of Retail Sales percentages of total residential CEA MLP GVEA FMUS HEA = SES_—s MEA plus commercial plus industrial geieeeyputic 6.0 13.3 3.5 25.3 0.3 0.1 Ont sales. Table 4.11 shows the x 0.7% 1.7% 0.9% 18.5% 0.1% 0.2% 0.0% percentages used, by utility’. aie Be ia a cael tela Regional _ percentages _ were % 0.0% 0.5% 0.5% 0.5% 0.5% 0.5% 0.5% computed as sales weighted , “1; Losses 97 47 25 8 22 2 24 averages of utility figures. % 11.2% 6.0% 6.0% 6.0% 5.7% 6.0% 6.0% ‘for CEA factor is combined loss/use measured against retail sales Table 4.11: Street light, Office, and Loss factors by Utility 4.6 Total Energy Requirements Total energy requirements in this study are defined as’: Requirements = Retail Sales + Street lights + Office Use + Distribution Losses (4.1) Figure 4.12 shows the probability distribution of year 2010 energy requirements developed from all 72 possible cases defined by the critical assumptions discussed in 4.1. In this figure the height of each bar reflects the total probability that year 2010 requirements will fall within the interval shown on the x-axis. Each interval has the same width. Therefore, the figure is an accurate description of where the "probability mass" lies. Total area of all the bars adds to the sum of the probabilities and the total area covered by the bars to the left of any given load value on the x-axis is equal to the cumulative probability that load will be less than that amount. ‘The definition of distribution losses for CEA is complicated somewhat by their retail sales to Girdwood which flow over a transmission line. Using CEA’s cost of service definition of retail losses as a stand-in for actual losses below the 115kV level should be a reasonably accurate approximation. *Data sources include: Utility PRS documents (GVEA, HEA, MEA, MLP), APA Electric Power Statistics (FMUS,SES), and Personal Communication (CEA, September). *Requirements are defined within the utility industry as Generation + Purchases - Sales for Resale. This is a "sources of power” accounting approach. In this study we have developed the various “uses of power" from the demand side. The two definitions should be exactly equal unless purchases are measured at the far end of a transmission line. 4 - 28 saseD ZL ‘sjuawosnboy [elo ‘UoNNqsIq AifIqeqolg :Z["p 21NS1] RAILBELT ELECTRICITY USE Probability Dist’n, all Cases Probability 0.16 0.145 Oot2 iii aes Olah 0.08; 0.06; 0.047 0.02, O Requirements, yr 2010, 00 Includes Sales, Office Use, Losses Figure 4.13 shows total energy requirements for the Railbelt under the LOW, MIDDLE, and HIGH case scenarios. Average annual growth lies within a range of .2 to 1.3%. Growth in the 1995-2010 period, during which growth in housing and floorstock resumes in all cases, lies within a range of .7 and 1.7%. Also shown is a breakdown of Railbelt sales by customer class for the MIDDLE case. Requirements at the regional level are presented in Figure 4.14 and broken down by customer class for the MIDDLE case in Figure 4.15. The steep but smooth drop in sales on the Kenai Peninsula in the LOW case is the result of the assumed loss of major industrial (refinery) load in the early 1990s, smoothed by interpolation. Figure 4.15 shows the significant share of industrial load in the Kenai total. Table 4.12 presents a comprehensive Growth Rate Summary covering the Railbelt and all regions under all three projection scenarios. Railbelt Total Energy Requirements Low, Middle, High Cases g Meat) 1980 1985 1987 1990 1995 2000 2005 — Low ——Middle — High Actual Railbelt Energy Requirements by Class Middle Case GWh 0 1987 1990 2005 IND OTHER Figure 4.13: Railbelt Energy Reqts, 3 Cases and Railbelt Reqts by Customer Class, MID Case 4-31 ce- h PI p emns1y ased pur uoisey Aq sjuowrornbey Total Energy Requirements, Anchorage Low, Middle, High Cases a a aR tL 1 at 1987 1990 1995 2000 2005 — Low —— Middle = —5~ High Total Energy Requirements, Kenai Low, Middle, High Cases 600 500 400 300 200 100 4 1 I Ma dh °o 1987 2005 1990 1995 2000 —- Low —~Middie High Total Energy Requirements, Fairbanks Low, Middle, High Cases a oO 1987 1990 1995 2000 2005 ——Low —~Middle — High Total Energy Requirements, MatSu Low, Middle, High Cases oO 1987 1990 1995 —- Low — Middle High ce-p SVD AICI ‘sse[D JowojsnD pure uoisey Aq sjuowenbey :¢{p omMsny = Anchorage Energy Requirements Middle Case GWh 3000 2500; 2000 & s WG 0D VV 1500 1000 500 oui 1987 1990 Cores LAcomm [_linn Sorter Kenai Energy Requirements Middle Case GWh 600 wl . RASS 400 GQ GLH ALLA GGFFP}P97}D}79DHLHDD} 300F 100; ye i a 1987 1990 1995 2000 2005 CJres E2%comm [_linn SSoTHer Fairbanks Energy Requirements Middle Case GWh 1 ps 1990 1995 2000 2005 (lres ZAcomm [jinn Sorter MatSu Energy Requirements Middle Case GWh 50 0 eS gee 1987 1990 1995 2000 2005 Cres LAcomm [_jinn SSotHer Total _|-------- Residential ------- ||-------- Commercial -------- | Electric Electric Occupied Retail Electric Total Retail Energy sales housing electric sales floor electric Industrial Requiremts units price stock price Sales RAILBELT 1987-1995 LOW -0.6% -0.8% 0.5% 1.7% 0.1% 0.7% 0.5% “4.9% MIDDLE 0.0% -0.5% 0.8% 2.3% 0.1% 0.7% 1.2% -0.2% HIGH 0.5% -0.2% 1.2% 2.8% 0.7% 1.4% 1.8% 2.5% 1995-2010 LOW 0.7% 0.8% 1.4% 0.3% 0.7% 1.1% -0.2% 0.2% MIDDLE 1.4% 1.7% 2.2% 0.5% 1.4% 1.9% 0.0% 0.6% HIGH 1.7% 2.0% 2.5% 0.8% 1.6% 2.2% 0.3% 1.3% 1987-2010 LOW 0.2% 0.2% 1.1% 0.8% 0.5% 1.0% 0.0% 71.6% MIDDLE 0.9% 0.9% 1.7% 1.1% 0.9% 1.5% 0.4% 0.4% HIGH 1.3% 1.2% 2.1% 1.5% 1.3% 1.9% 0.8% 1.7% ANCHORAGE 1987-1995 LOW -0.3% -0.7% 0.3% 2.3% 0.0% 0.5% 0.3% 0.7% MIDDLE 0.0% -0.3% 0.7% 3.1% 0.1% 0.6% 1.1% 1.3% HIGH 0.5% 0.0% Ta1% 3.6% 0.7% 1.3% 1.7% 6.0% 1995-2010 LOW 0.6% 0.7% 1.4% 0.5% 0.7% 1.1% -0.4% 0.0% MIDDLE 1.5% 1.6% 2.3% 0.8% 1.4% 2.0% -0.1% 1.0% HIGH 1.7% 1.9% 2.6% 1.1% 1.6% 2.3% 0.3% 1.4% 1987-2010 LOW 0.3% 0.2% 1.0% 1.1% 0.4% 0.9% -0.2% 0.2% MIDDLE 1.0% 1.0% 1.7% 1.6% 1.0% 1.5% 0.3% 1.1% HIGH 1.3% 1.2% 2.1% 2.0% 1.3% 1.9% 0.8% 3.0% FAIRBANKS 1987-1995 LOW 0.8% 1.0% 1.0% 0.1% 0.6% 1.2% 0.1% 0.8% MIDDLE 0.7% 1.1% 1.1% 0.8% 0.1% 0.9% 0.8% 0.9% HIGH 1.5% 1.4% 1.5% 1.2% 0.8% 1.6% 1.3% 4.6% 1995-2010 Low 0.8% 1.2% 1.5% -0.1% 0.7% 1.2% -0.1% 0.0% MIDDLE 1.3% 1.7% 1.9% 0.1% 1.1% 1.6% 0.2% 0.7% HIGH 1.7% 2.1% 2.4% 0.4% 1.4% 2.0% 0.4% 1.3% 1987-2010 LOW 0.8% 1.1% 1.4% 0.0% 0.6% 1.2% 0.0% 0.3% MIDOLE 1.1% 1.5% 1.6% 0.4% 0.7% 1.4% 0.4% 1.3% HIGH 1.6% 1.9% sik 0.7% 1.2% 1.9% 0.8% 2.4% KENAL 1987-1995 LOW -3.1% -1.0% 0.6% 1.3% 0.6% 1.2% 1.4% -10.3% MIDDLE -0.7% -0.8% 0.8% 1.6% 0.6% 1.1% 1.8% -1.3% HIGH 0.0% -0.7% 0.9% 1.8% 0.9% 1.5% 2.0% 0.4% 1995-2010 LOW 0.6% 0.7% 1.2% -0.0% 0.7% 1.1% 0.0% 0.4% MIDOLE 1.0% 1.4% 1.7% 0.1% 1.0% 1.7% 0.1% 0.4% HIGH 1.3% 1.6% 2.0% 0.3% 131% 1.8% 0.3% 1.1% 1987-2010 LOW -0.7% 0.1% 1.0% 0.4% 0.6% 1.2% 0.5% +3.4% MIDDLE 0.4% 0.6% 1.4% 0.6% 0.9% 1.5% 0.7% -0.2% HIGH 0.8% 0.8% 1.6% 0.8% 1.0% 1.7% 0.9% 0.8% MATSU 1987-1995 LOW 72.1% -3.1% 0.5% 1.9% 0.2% 0.7% 1.9% 9.6% MIDDLE -1.8% -2.8% 0.8% 2.3% 0.2% 0.7% 2.3% 15.3% HIGH =1.3% -2.3% 1.3% 2.6% 0.7% 1.5% 2.6% 15.3% 1995-2010 LOW 0.8% 0.7% 1.6% 0.1% 0.8% 1.4% 0.1% 3.5% MIDDLE 2.0% 2.1% 2.6% 0.2% 1.6% 2.6% 0.2% 3.9% HIGH 2.3% 2.4% 2.9% 0.4% 1.8% 2.7% 0.4% 7.0% 1987-2010 LOW -0.2% -0.6% 1.2% 0.7% 0.6% 1.2% 0.7% 5.5% MIDDLE 0.7% 0.4% 2.0% 1.0% 1.1% 1.9% 1.0% 7.7% HIGH 1.0% 0.7% 2.3% 1.2% 1.4% 2.3% 1.2% 9.8% Table 4.12: Electric Demand Growth Rate Summary 4 - 34 = ma 4.7 Sensitivity to Gas Penetration Several critical assumptions happen to be the same in all three representative cases.* These include the technical change assumption and the gas penetration assumption. Table 4.13 reports the changes in residential and commercial sales attributable to the High natural gas penetration assumption with all other assumptions held constant. The MIDDLE projection case is our reference point and produces the lower load because it assumes High gas penetration. The sensitivity result is obtained by changing the gas penetration assumption to Base instead of High. Specifically, this means no gas to Homer or to the remote MatSu Valley. The effect of the gas penetration assumption is modest but significant: In order to show it clearly in Figure 4.16 we confine the plot to residential and commercial sales. If the effect appears small, it should be kept in mind that we made conservative assumptions, based largely on the end use survey data, about the actual amMOUuNt Of mame Table 4.13: Effect of Gas Penetration Assumption on expected to be backed out by Sales, Relative to MIDDLE Case electricity which could be extension of the gas system. Effect of Gas Penetration Assumption Region : Gas : Case Kenai : High : Base % Diff. MatSu : High : Base % Diff. 1987 270.2 270.2 0 282.8 282.8 0 1990 271.2 274.11 1.1 259.2 259.2 0 1995 261.5 278.0 5.9 242.4 254.7 4.8 2000 270.1 288.7 6.4 243.6 267.2 8.8 Residential + Commercial Sales, GWh 2005 285.4 311.3 8.3 272.9 296.5 8.0 2010 It is important to remember that the lower gas penetration assumption was used in producing the overall distribution of outcomes from which the three representative cases were chosen. Therefore, although none of the three cases is explicitly based on low gas penetration, the range of outcomes described by the cases does encompass the outcomes produced from an assumption of low gas penetration. 4-35 Effect of Gas in HOMER BASE vs HIGH Penetration, all else MID Res + Comm. Sales, GWh 350 300 250 200 150 100 50 1987 1990 1995 2000 2005 —>- BASE ——HIGH Effect of Gas in MATSU BASE vs HIGH Penetration, all else MID Res + Comm. Sales, GWh 400 350 300 250 200 150 100 50 0 | _L 4 1 L i 1 1 1 1 4 1 1 1987 1990 1995 2000 2005 —- BASE —— HIGH Figure 4.16: Sensitivity to Gas Penetration Assumption: Kenai, MatSu 4 - 36 SSS ee 4.8 Peak Load We projected peak load under the assumption of a constant load factor in each region. Load factor can and will change as the end use pattern of demand changes. However, we know too little about end-use specific load factors to feel confident in making any projections of such changes, given their small magnitude. We did perform some scoping calculations of peak load using disaggregated load factors by end use and found that the overall factor changed very little. The reduction in heat load, which the forecasts generally show, clearly improves the average factor, but the key uncertainty is the load factor for the Miscellaneous end use, since that use grows in relative importance. To derive a suitable average load factor, we developed average historical load factors for the 1982-1987 period in consultation with Railbelt utility staff. This analysis is summarized in Appendix E. Load factors throughout the Railbelt have been notably increased during the past three years. Part of this phenomenon is directly attributable to warm weather. We therefore use the 1982-87 average here, rather than the 1984-87 figure. Little hard data exists on coincidence factors between utilities and between regions. The one analysis of which we are aware showed a 95-97% coincidence factor between the CEA and MLP. Utility staff agreed that 97% would be a reasonable figure for all aggregations. This figure is applied twice in some cases: once to aggregate across utilities and a second time to aggregate across regions to derive total Railbelt peak load. Since the load factor is assumed to remain constant throughout the forecast period within a region, the peak load is a constant multiple of the energy requirements. Both quantities grow at exactly the same rate. For this reason we are able to present both peak load and energy requirements on the same graph at the regional level, as in Figure 4.18. Since the Railbelt total peak load is a weighted average of the regional loads, it is shown separately along with requirements in Figure 4.17. Peak load growth rates are identical to those for energy, and may therefore be read off the "requirements" column of the growth rate summary. Railbelt Total Energy Requirements Low, Middle, High Cases Gwh 5000 4000 3000 2000F 1000 tt 1987 1990 1995 2000 2005 2010 — Low —~Middle — High Raibelt Peak Load (MW) Low, Middle, High Cases 1000 800 600¢ 400F 200F gut A fe et fe ee | ee eer ee 1987 1990 1995 2000 2005 2010 | —Low —~Middle —High Figure 4.17: Railbelt Energy Requirements and Peak Load, 3 Cases 4 - 38 OSiGLy: SOseD ¢ ‘peo yvog pur sjuowsmboy AdIoU_ [eUuOIZOY ‘ETP ONS Total Energy and Peak Load, Anchorage Low, Middle, High Cases Energy, GWh 3000 bat Fay pe eg 1987 1990 1995 2000 2005 —— Low ——Middle High Total Energy and.Peak Load, Kenai Low, Middle, High Cases Energy, GWh ee ee ee 1987 1990 1995 2000 2005 —— Low —~Middle — =~ High Total Energy and Peak Load, Fairbanks Low, Middle, High Cases Energy, GWh Peak, MW 1000 800 600: 400 200 ‘a eee 1987 1990 1995 2000 ——Low — Middle — High Total Energy and Peak Load, MatSu Low, Middle, High Cases Energy, GWh 0 been 1987 1990 1995 2000 2005 ——Low — Middle — High 4.9 Comparison With Econometric Results In this section we briefly compare our forecast results to several recent econometric results prepared by the Anchorage utilities. We confine our attention to the Anchorage region and the MID case forecast. 4.9.1 ISER Growth Rates vs. Econometric Forecast Growth Rates end use model results with recent Seneca] Conpeuueons econometric results obtained by CEA #1: 1987 CEA PRS econometric mid case and MLP. Part 1 compares the 1987 forecast:ve) [SER forecest CEA Electric Power Requirements Residential GWh Commercial GWh Study’ middle case econometric load CEA —ISER CEA —ISER forecast of residential and commercial 1987-1995 0.2% -0.3% -0.4% 0.1% sales to the current study MIDDLE 1995-2010 0.4% 1.6% 3.0% 1.4% case projections of residential and 1987-2010 0.2% = 1.0% 1.8% = 1.0% commercial sales. In both sectors, the #2: 1988 MLP PRS econometric equation two models produce the same general with ISER exogenous variables pattern. The difference between them veLISER forecast lies in the sectoral differences between Residential GWh Commercial Gwh residential and commercial sales. CEA MLP = ISER MLP = ISER shows higher commercial growth, while 1967-1995 -1.0% -0.3%. 0.7% 0.1% we show higher residential growth. 1995-2010 1.5% 1.6% 1.7K 1.4% 1987-2010 0.6% = 1.0% 1.4% 1.0% In creating part 2 of the table we refined the comparison by re-creating the MLP 1988 PRS econometric Table 4.14: Econometric Results vs ISER Results forecasts of residential and commercial sales using the same values for exogenous variables as employed in this study. That is, we employ the MLP forecasting equation with MLP’s coefficients which have been estimated on Alaska historical data but enter the same independent variables as are used to drive the end use models. The MLP model forecasts customers and kWh/customer separately. The equation for sales per customer is: log(kWh/CUST) = C + bi*log(PRICE(-1)) + b2*log(INCOME) + b3*log(WEATHER) (42) where WEATHER = Heating Degree Days We forecast residential sales per customer using this equation populated with values for PRICE and INCOME taken from the middle case price and economic scenarios discussed in section 4.1. (We set WEATHER =1 and adjusted the constant term accordingly so the fitted equation still passed through 1987 actuals). We then multiplied kWh/CUST by our exogenous household forecast to derive total forecast residential sales figures for the Anchorage region. The sales growth rates obtained are reported in Table 4.14 part 2. They agree quite well with each other, especially over the 1995-2010 period when the complicating factor of excess vacancy is not present. Sables IV-3 and IV-4 4-40 The same procedure was repeated for the commercial sector. The equation used is: log(GWh) = C + b1*log(/EMPLOYMENT) + b2*log(GWh(-1)) (43) where GWh = total commercial sales Once again, the growth rates and patterns of growth are in substantial agreement. Both of these results shown in part 2 of Table 4.14 are not surprising, given that the same economic and price scenarios are driving sales. The congruence between results is reassuring, however. It tells us that the end use models, which are loaded with hundreds of individual elasticities and engineering relationships estimated from national data, seem to faithfully reproduce the aggregate response to price and economic growth which these equations embody. 4.9.2 Synthetic Regression Models of End Use Forecast Results check on the reasonableness of our Synthetic Regressions on Forecast Results results, we fitted a "synthetic" regression Residential Sector model to our forecast results. We use . Model: log(SALES)= C+b1*log(PRICE(-1))+b2*log(#CUST ) the term synthetic because the "data" to Pe which the equation is fitted are not real Regression Output: historical data but a set of forecast Soe aia eee —o values for price, floorstock, and sales. __ R Squared 0.989481 The coefficients of the synthetic No. of Observations 3 regressions reveal the aggregate price STi “ elasticity of demand consistent with the Variable Name: PRICE(-1) #CUST end use model results ee T-statistic: -6.28 26.33 Table 4.15 reports the results of this CEA coefficient: -0.65 1.034 exercise. We started the process with the Chugach residential and commercial Commercial Sector equation forms, but had to drop Model: log(SALES) = c+b1*log(FSTOK)+b2*log(PRICE) employment from the commercial Regression Output: equation because it is almost perfectly Constant 4.623881 colinear with floorstock. The key tel eee variable to compare is the price No. of Observations Gihizs elasticity. In the residential sector the Degrees of Freedom 20 implied AKREM elasticity is -.23, vs Variable Name: FLOORSTOCK PRICE CEAs estimate of -.65 obtained from X Coefficient(s) 0.6% -0.250 historical data. Both values are well Sec) Eliot cost pro dec O-O7n ek T-statistic 31.78 -3.50 within the range of _ reported CEA Coeff: NA - 296 econometric estimates, although tC AKREM model is at the low end of the Table 4.15: Synthetic Regressions on Forecast Results spectrum. One reason for this may be the influence of federal efficiency standards. The standards drive consumption down independently of the movements in price. In the commercial sector, the elasticity estimate implicit in the COMMEND MID case forecast, -.25, agrees almost perfectly with the CEA result of -.296. This is a particularly reassuring result since the current study represents the first attempt to apply an end use model to the commercial sector. 4-41 REFERENCES Guide to References EPRI Electric Power Research Institute reports are referenced by EPRI report number. All are published by EPRI Research Reports Center, Box 50490, Palo Alto, CA 94303. (415) 965-4081. ACEEE American Council for an Energy-Efficient Economy reports are published by ACEEE, 1001 CT Ave NW, Suite 535, Washington DC 20036. (202) 429-8873. Alaska Electric Power Statistics, 1960-1985. 1986. Juneau: Alaska Power Authority. Alaska Electric Power Statistics, 1960-1986. 1987. Juneau: Alaska Power Authority. Alliance to Save Energy 1987, Industrial Decision-Making Interviews: Findings and Recommendations, for Michigan Energy Options Study. Lansing MI: Michigan Public Service Commission. Barakat, Howard & Chamberlin, 1988. (415-893-7800). Demand Side Management Program Analysis. Volume III: Commercial/Industrial Sector Report. Prepared for Long Island Lighting Co. Berman, Matthew, 1988. Hedonic Price Equations for the Alaska Housing Market. Anchorage: ISER internal memorandum. Brookhaven National Labs 1987. Analysis and Technology Transfer Annual Report - 1986. Washington DC: US. Dept. Energy, August. California Public Utilities Commission, 1985. 1984 Energy Conservation Program Summary. San Francisco. Chiu, S.A., and F.R. Zaloudek, 1987. R & D Opportunities for Commercial HVAC Equipment Richland, WA: Pacific Northwest Laboratories PNL-6079. Chugach Electric Association, 1987. Power Requirements Study. Anchorage AK. Cleary, Colleen, and Barbara Crimmin. 1986. Commercial Hourly End-Use Study. Status Report, 1982-1985. Seattle: Seattle City Light. Colt, Steve. 1986. AKWARM: Life-Cycle Cost Worksheet for Residential Buildings, User’s Guide. Anchorage, AK: Institute of Social and Economic Research prepared for Department of Community and Regional Affairs. Decision Focus Incorporated, 1989. Railbelt Intertie Feasibility Study, Interim Report. Los Altos, CA: Decision Focus, Inc. Electric Power Research Institute, 1986. Manual on Indoor Air Quality. Palo Alto: EPRI EM-3469. Easton Consultants, 1987. Commercial Market Fuel Selection Study. Prepared for the American Gas Association. Geller, Howard S. 1986a (revised 1987). Energy and Economic Savings from National Appliance Efficiency Standards. Washington, D.C.: ACEEE draft report prepared for OTA. Geller, Howard S. 1987. Residential Equipment Efficiency: A State-of-the-Art Review. Washington, D.C.: American Council for an Energy-Efficient Economy (review draft). Golden Valley Electric Association 1987. Power Requirements Study Prepared by CH2M Hill Inc. Goldsmith, O. Scott 1988. Economic and Demographic Projections for the Alaska Railbelt: 1988-2010. Prepared for the Alaska Power Authority. Anchorage: Institute of Social and Economic Research. Goldsmith, Scott and Lee Huskey, 1980. Electric Power in the Railbelt: A Projection of Requirements Anchorage: Institute of Social and Economic Research. Hirst, Eric, and Janet Carney. 1978. The ORNL Engineering-Economic Model of Residential Energy Use. Oak Ridge, TN: Oak Ridge National Laboratory for U.S. Department of Energy. ICF-Lewin Energy Group, 1988. Fuel Price Outlook for the Alaska Railbelt Region: Oil and Natural Gas. Prepared for the Alaska Power Authority. Knight, Frank 1921. Risk, Uncertainty, and Profit. New York: Harper & Row. Lann, R.B., et al. 1985. An Implementation Guide for the EPRI Commercial Sector End-Use Energy Demand Forecasting Model: COMMEND, Vols. 1 and 2, rep. EPRI EA-4049-CCM. Atlanta, GA: Georgia Institute of Technology. Lin, William, et al. Fuel Choices in the Household Sector. 1976. Oak Ridge, TN: Oak Ridge National Laboratory for Energy Research and Development Administration. McDonald, Craig et. al. 1987. Commercial Sector Demand Side Management Options Analysis, final report. Lansing: Michigan Public Service Commission. McMenamin, J.S. 1988. Commercial End-Use Data Development Handbook: COMMEND Market Profiles and Parameters, Vols. 2 and 4, interim rep. EPRI P-4463-SR. San Diego, CA: Regional Economic Research, Inc. McMenamin, J.S. 1988b. Commercial Sector Payback Criteria file: a collection of excerpts from published studies. McMenamin, J.S. 1987. Models of Commercial Sector Equipment and Fuel Choice Decisions for the COMMEND Code: Framework Design and Data Development Plan. EPRI EM-5356. August. City of Seward 1985. Review and Analysis of Retail Electric Rate Designs.Prepared by R.W. Beck and Associates, Inc. Regional Economic Research, Inc. 1986. (ph619-481-0081) A Review of Commercial Energy Use Indexes. Draft report to EPRI, September. Ruderman, Henry et al. 1987. "Energy-Efficiency Choice in the Purchase of Residential Appliances" in Eneryg Efficiency: Perspectives on Individual Behavior, ed. Willett Kempton and Max Neiman. Washington D.C.: American Council for an Energy-Efficient Economy. Scott, M.J., et al. 1983. RED Model (1983 Version) Technical Documentation Report. Richland, WA: Battelle Pacific Northwest Laboratories prepared for Harza-Ebasco Susitna Joint Venture. Statt, T.G. 1988. The Use of Chloroflourocarbons in Refrigeration, Insulation, and Mobile Air Conditioning in the United States. Washington DC: US Dept. Energy , prepared for the 1988 EPA Conference on CFCs and Halon. Synergic Resources Corporation, 1986. Commercial End Use Metering Workshop Proceddings. EPRI EM-4393. Verzhbinsky, Gleb, et al. 1985. The Residential Hourly and Peak Demand Model: Description and Validation. Berkeley, CA: Lawrence Berkeley Laboratory prepared for U.S. Department of Energy. Appendix A: Regression Analysis of Residential Electric Consumption This page intentionally left blank A. REGRESSION ANALYSIS OF RESIDENTIAL CONSUMPTION This appendix presents results of multiple regression analysis of residential end use survey data from the ISER 1987 end use survey. The primary goal of the analysis was to develop electric heat EUI estimates. This task is complicated by the presence of so many houses using multiple fuels for heating (see Table 2.3). Dummy variable regression provides a powerful tool for assigning EUI values to "all-electric", "primarily electric", and "some electric" usage intensities. The residential end-use survey lists in great detail which appliances are used in a sample of about nine hundred residences in the Railbelt. For about half of the households in this survey, we were able to obtain solid consumption variables from actual (raw) read data.’ Using simple regression techniques we can account for the various factors affecting appliance use and can find the most likely decomposition of total electric use into use by individual appliances. Section 1 proposes a way of thinking about appliance use which is amenable to statistical analysis. The regression techniques and data needed to estimate the use by each appliance are discussed in Section 2. The results of these regressions and the estimated use by each appliance are presented in Section 3. A.1 Decomposing the Use of Electricity The variety of appliances in residences can be limited to those that are the dominant users of electricity: electric space heating, water heater, dryer, stove, roof heat tape, waterbeds, refrigerators, freezers, and other major appliances. Each of these appliances contributes to total electricity consumption in different ways. The predominant use of electricity in many residences is space heating. Heating consumption is a function of alternative fuel use, housing unit size and surface exposure, weather, and occupant behavior. Large appliance consumption - hot water heaters, stoves, and clothes dryers- depends on largely on the number of people living in the household. Other appliances use electricity in relatively stable amounts independent of the configuration of the house, the number of people in the house, or other characteristics of the household. These include waterbeds, refrigerators, roof heat tape, block heater, separate freezers, electric heating fan, dishwasher, water well pumps, and jacuzzis. The remaining appliances fall into two categories: those appliances that almost everyone has and those appliances that very few have. Over ninety percent of all of the residential households have a clothes washer, a refrigerator, a T.V., lighting, a radio or two, and many small appliances like toasters and clocks. This common electric usage can be treated as a single block of common or base electric consumption which is the same in all households. In addition, each household has unique appliances or unique uses of common appliances. Records were dropped for several reasons, including: multiple consumption records linked to one identifier (phone number), missing data, bad account number supplied by customer, and several outlier screening criteria. We have no reason to suspect that the dropped records are correlated with the regressors; hence we believe the truncated data set is still representative. A-1 One residence may have special electrical machinery while another has a collection of electronic equipment. Some residences may maintain the temperature at night while others do not. Some may have elaborate stereo equipment which plays continuously, while others have a broad collection of electrical kitchen appliances. All of these unique uses of electricity are treated as random variations in the use of electricity relative to the average household. All of these appliances, from electrical furnaces to stereo equipment are used simultaneously in the household. Some appliances are strong complements, such as electric clothes washer and an electric dryer. The use of an electric hot water heater moves with the use of appliances that rely on hot water such as the clothes washer and dishwasher. Other appliances are present only when other larger appliances are used. An electric heat fan distributes hot air only when a gas or oil furnace is used. A water well pump is usually needed only outside urban areas where many other types of special electrical appliances, such as additional lighting, are also more common. As a matter of statistics, this complementarity of appliances makes the separation of electrical usage into individual appliances difficult. One solution is to explicitly restrict the use of these appliances to be proportional to or connected with each other according to technological relationships. For example the amount of heat used to heat the water for use in a clotheswasher or dishwasher could be specified. We prefer to adopt an alternative approach and to limit the analysis only to those appliances which have a distinct, independent level of electrical use which is not critically dependent on other appliances. For this reason, appliances such as the clothes washer and well pump are omitted from the analysis. The drawback of this approach is that the appliances which complement the omitted appliances must now represent not only themselves but also the use of the complementary appliances. Other factors which may contribute to the consumption of electricity in the household include location, family income, the utility providing the electricity, the availability of gas, oil, coal and wood as alternative heat, and other economic, demographic, and geographic characteristics. All of these factors are assumed not to alter the contribution by an individual appliance to total electric consumption in the household. This assumption certainly simplifies actual conditions. A.2 Methodology and Data This section presents an estimable model of appliance use based on the structure outlined above. We can think of each household in the survey as having a collection of electric appliances. Each appliance is assumed to use the same amount of electricity in every household where it is present with some adjustments for the number of people and type of house. The total consumption of electricity in each house is the sum of the individual uses by every appliance in the house. Since each house has a slightly different mixture of appliances its total consumption of electricity is unique. We could estimate the contribution to total electricity consumption by one appliance by finding two households which have the identical appliances except that one house has one additional electrical appliance, say a clothes dryer. The difference between the total electricity consumptions in the two houses could serve as an estimate of A-2 the consumption of electricity by the clothes dryer which is present in only one of the houses. This estimate relies on the fact that not all households have the same appliances. An extension of this simple method would construct a list of all the appliances found in each household in the survey. Comparing these lists of appliances for all of the houses at once would enable us to find out what additional contribution each appliance makes to total electricity consumption when it is present in a household. A statistical regression can do this large comparison by finding the contribution from each appliance, when it is present in a household, to the total electricity consumption of the household. For each household, we can observe whether an appliance is present. For each appliance, a dummy or dichotomous variable takes the value one if the appliance is present and zero if the appliance is not present. A regression of total consumption for each household on the collection of appliance dummy variables for each household and appliance will estimate the contribution of each appliance (as represented by a dummy variable) to total electricity consumption. For those appliances which use more electricity when the area of the house or the number of people is different, the dummy variables for those appliances are interacted with those factors. The space heating variables are multiplied by the area of the house to account for the fact that more electricity is used in large houses than in small houses for heating. In addition, we allow the space heating dummy variable to be different when different types of other heating are available in the house. When some other dominant heat source is available, such as oil or gas, a dummy variable for "some electric heat" takes the value one. When electric heat is the primary heat source, but it is supplemented by other sources, the dummy variable "primarily electric heat" takes the value one. And when only electric heat is used, the variable "only electric heat" takes the value one. The water heater, dryer, and stove dummy variables are multiplied by the number of people to account for the variation in use of these appliances with the number of people. The heating dummy variable is split into different effects to allow for different contributions of electric heating depending on what other forms of heating are available. The descriptive statistics of the variables included in the model for the sample of 430 households for which end-use survey data and total consumption data is available are summarized in Table A.1. In mathematical terms, the regression model is Total Electricity Consumption = Constant + (Some Electric Heat)*(Area)*(Multiple) + (Primarily Electric Heat)*(Area)*(Multiple) + (Only Electric Heat)*(Area)*(Multiple) + (Some Electric Heat)*(Area)*(Single) + (Primarily Electric Heat)*(Area)*(Single) A-3 + (Only Electric Heat)*(Area)* (Single) + (Water Heater)*(Number of People) + (Dryer)*(Number of People) + (Stove)*(Number of People) + (Waterbed) + (Roof Heat Tape) + (Additional Refrigerator) + (Block Heater) + (Separate Freezer) All of the independent variables are dichotomous variables or the product of dichotomous variables and one of two continuous variables, number of people or area of the house. The constant in this model incorporates the use of electricity in appliances which are common to nearly all of the households in the survey. The constant includes lighting, refrigerator, T.V., clotheswasher, and numerous miscellaneous small appliances that are common to all households. This common or base electricity consumption can not be decomposed with statistical methods since no households (or very few) lack the components of the base use and no statistically meaningful comparison can be made among households. This model of appliance use emphasizes the sum of all uses of electricity rather than explaining the reasons for use or effect of complementary use among appliances. As discussed in the previous section, not all appliances are included in the regression. Inclusion of complementary appliances would result in multicolinearity and estimates biased toward zero. But limiting the number of appliances included in the model forces those appliances that remain to account for the variations in use of complementary appliances, resulting in upwardly biased estimates of appliance use. By carefully selecting the appliances included in the regression, we have attempted to minimize the effect of complementary use. Notably absent from this model are the list of "Outside conditions," as listed in table 1 which also affect the use of appliances. As discussed in the previous section, this model emphasizes the decomposition of electricity into individual uses rather than the reasons or motivations for using electricity. By assuming that if an appliance is present in a household, it is used the same regardless of the economic, demographic, and locational characteristics of the household. Technically, these outside characteristics are assumed to be orthogonal to the use of appliances in the households. If this assumption is not true, then the estimates of the coefficients in the specified model will not necessarily be biased but they will not be the statistically "best" estimates: the standard errors of the estimates will be larger. Table A.1: DESCRIPTIVE STATISTICS OF DICHOTOMOUS VARIABLES Variable Description Mean Standard Deviation ANN Annual Electricity Consumption 12725.056 8111.676 WIN Winter Electricity Consumption 3857.584 2733.781 SUM Summer Electricity Consumption 2517.9 1643.035 SINGLE _ Single Family house or mobile home 0.853 0.354 MULTIPLE Attached, multiple family house 0.147 0.354 A Area of house in square feet 1730.551 802.397 ES some of heat is electricity 0.181 0.386 EP most of heat is electricity 0.04 0.195 EO all of heat is electricity 0.056 0.23 SAES Single*area*es 297.744 732.024 SAEP Single*area*ep 63.837 363.525 SAEO Single*area*eo 60.628 357.909 MAES Multiple*area*es 15.116 131.086 MAEP Multiple*area*ep 4.419 91.626 MAEO Multiple*area*eo 28.077 185.964 Fi people 3.021 1333/7) W electric waterheater 0.367 0.483 WE electric waterheater*people 1.095 1.631 D electric dryer 0.795 0.404 DP electric dryer*people 2.405 1.698 S electric stove 0.672 0.47 SP electric stove*people 1.984 1.739 R second refrigerator 0.035 0.184 5 separate freezer 0.716 0.451 WB waterbed 0.328 0.608 FN electric heating fan 0.33 0.471 H hottub or jacuzzi 0.044 0.206 BH block heater 0.502 0.501 HT roof heat tape 0.021 0.143 DW dishwasher 0.772 0.42 CW clothes washer 0.947 0.225 CWP clothes washer * people Z93) 1.463 WW water well pump 0.442 0.497 ANCH in Anchorage 0.505 0.501 KENAI in Kenai Penninsula 0.126 0.332 MATSU _ in Matanuska Valley 0.191 0.393 FBNKS __ in Fairbanks 0.179 0.384 Unless explicitly noted (for area and consumption), all variables are dichotomous variables which take the value one if the characteristic is true for an individual household and zero otherwise. Not all of these variables are included in the regression, as noted in the text. A3 Results The model identified in the previous section was estimated using ordinary least squares regression techniques. The results of this regression using annual electricity consumption as the dependent variable are presented in Table A.2. The estimated coefficients can be interpreted as the additional electricity consumption contributed by an individual appliances relative to the base or constant level of consumption present in each household. In general, the coefficients on the heating variables and other appliances which do not depend on the number of people are robust: these coefficients retain their magnitude and significance regardless of alternative specifications. Nearly all of the coefficients are significant and the R Square statistic is very high for this type of cross sectional regression -- suggesting a generally good fit of the model. Most of the variation in consumption levels is explained by the collection of heating variables. As expected, residences which use only electric heat use more electricity per square foot than houses which supplement electric heat with another source. Comparing the use of electiricty in single family houses using only electric heat to single family houses using primarily electric heat, we find that the houses which supplement their electric heat save about four kilowatt hours per square foot of forty percent of the total electric use of a household using exclusively electric heat. Of those households that supplement their electricity use by an alternative heat source, sixty percent use wood and twenty percent use gas. Comparing the use of electricity in single family houses with some electric heat to the houses with primarily electric heat may involve a shift from the use of small space heaters to the installation of a central electric heater. The additional electric consumption in electricity is four and a half kilowatt hours per square foot, or a four fold increase the consumption of electricity for heating. The only interesting multi-unit coefficient is for those units which use electricity only. The dummy variable for multi-units which supplement their electric heat is non-zero for only one observation and therefore would measure only the residual of that observation if the variable were included in the regression. The coefficient for multi-units which use some electricity is statistically insignificant and cannot be meaningfully compared to the other coefficients. Multi-family units in Anchorage are dominated by rental units which are equipped to only use electricity for heating. These units are also dominated by renters, many of whom may have heat included in the rent. This incorporation of the heating bill removes the incentive to make substitutions of alternative heating sources for electric heat. These results suggest that among homes that use electricity exclusively, single family homes use almost twice as much electricity per square foot than do multi-family homes. This difference in electricity consumption may be due to differences in construction, shared heat from adjacent residences in a multi-unit building, smaller rooms in multi-unit residences, or higher ceilings in single family houses. A-6 Table A.2: Regression of Annual Consumption of Electricity on dichotomous appliance variables Estimated Coefficient | Standard Error of coefficient Use of electricity per square foot per year for electric heat in Single Family Houses using some electric heat 51) || 5 0.37 using primarily electric heat Gi2T 7) > 0.75 using only electric heat 1007) 2 0.74 Multi-Family Homes using some electric heat 0.7 1.98 using primarily electric heat iq using only electric heat 3:4))|||/* 1.43 Use per person per year for electric dryers 623:2' | * 171.8 electric hot water heaters 1395.3 * 176.4 electric stoves 160.6 165.5 Use per year for waterbeds 1172:9"* 427.8 roof heat tape 5830.6 * 1833.8 additional refrigerator 975.8 1428.5 block heater 521.6 3259) separate freezer 2367.4 * 589.7 base use per year for all appliances 522133 ¢ 667.9 common to all households (constant) R Squared = 0.58 Degrees of Freedom = 416 * Coefficients which are statistically significantly different from zero at the ninety percent confidence level are designated by a "*" The coefficients on the three variables which include the effect from the number of people are not statistically significantly different than our a priori expectations of the use of these appliances. However, the expected use of electricity in each of these appliances is not robust: the significance and magnitude of the estimated coefficients change substantially if the effect of the number of people is not included or if additional fixed effects from the appliances independent of the number of people are included in the regression. Attempts to resolve this sensitivity by altering the measurement of the number of people from the average number of people to the number of people above or below the average number of people in a household were not successful. One further difficulty in estimating the effect of the number of people is that the variation in the number of people in households is not large -- ranging from 1 to 7 with a mean of 3 and a standard deviation of 1.3. The remaining appliances which contribute a constant amount of electricity, independent of the number of people and the configuration of the house all have coefficients which are not statistically significantly different from expected use of these appliances. For the sake of comparing all appliances, the estimated use of electricity per square foot for heating can be scaled by the mean area of all houses and the estimated use of electricity per person in dryers, water heaters, and stoves can be scaled by the mean number of people in all houses. Based on the coefficients reported in Table A.2 and the mean of area and number of people in the houses (as reported in Table 2), the estimated electricity use per year of each appliance is as follows: Table A.3: Use per year of Appliances For electric heating in Single Family Houses using some electric heat 2595 using primarily electric heat 10726 using only electric heat 18511 in Multi-Family Homes using some electric heat 1211 using primarily electric heat y using only electric heat 9342 electric dryers 1870 electric hot water heaters 4186 electric stoves 482 waterbeds 1173 roof heat tape 5831 additional refrigerator 976 block heater 522 separate freezer 2367 constant 5222 Electric heat is the largest total user of electricity. Roof heat tape, hot water heaters, and separate freezers use less than one third of the electricity used in electric furnaces. The remainder of the appliances make much smaller contributions. A-8 A.4 Extensions Comparisons of annual, winter, and summer consumption levels can serve as an extra verification of our results. The regression results for winter and summer consumption are approximately one fourth the magnitude of the coefficients in the annual consumption regressions because winter and summer consumption covers only three months each while annual consumption occurs over twelve months. After adjusting by this factor of four, the electricity consumption closely associated with heating is much larger in the winter. All electric heating levels are significantly higher in the winter. In addition the use of heat tape, and block heaters are much higher. Appliances which are not significantly different are freezers, waterbeds, stoves, and dryers. The notable exception to these winter/summer comparisons is water heaters. When both a variable for the constant use of water heaters and the per capita use of water heaters are included in the regression, the constant or base use of water heaters is greater in the winter, but the per capita use of water heaters is the same in both seasons. The energy required to keep the water heater at a stable base temperature is greater in the winter than in the summer. Including some of the appliances which were excluded from the regression due to complementarity with other appliances gives expected results. For example, a heating fan is negatively correlated with electric heat use. When the heating fan dummy variable in the regression, its coefficient significantly negative, as expected. The existence of a water well pump indicates the presence of many other additional uses of electricity such as additional lighting. Including the well pump dummy variable in the regression leads to an estimated coefficient which is technologically far too large for use only in a well pump: the coefficient is most likely incorporating the effect from other rural appliances. Appliances which are present in nearly all households, such as clothes washers and refrigerators, have little variation and their coefficients were not significant when entered into the regression. Removing the clothes washer variable leads to a substantial increase in the coefficient on dryers, verifying our hypothesis that the use of these appliances are highly correlated. As a result, the coefficient on dryers includes use of electricity both in dryers and in clothes washers. An alternative specification of the model accounts for the effect of both the area of the house and the area of the house squared is intended to estimate any nonlinear effects of area on electricity consumption. The estimated effect from the area of the house squared is very small or insignificant. When the effect is significant, it is negative, suggesting that electricity use per square foot decreases as the total area increases. These results suggest that small economies of scale exist for heating larger homes. These economies of scale to heating are most appropriate for residences with areas more than fifty percent larger than the mean. At this level the estimated electricity use per square foot in linear estimation becomes more than the estimated electricity use per square foot in the quadratic specification. A-9 A.5 Conclusions We modeled and estimated the use of electricity in individual appliances by using statistical methods to distribute total electricity consumption in a household among the appliances which are present in the household. Estimated electricity use in appliances, as presented in Table 3 reliably reflect the expected use of these appliances. The advantage of these estimates is that they are based on actual consumption data and information about the actual appliances in use. Our conclusions regarding space heating are particularly noteworthy since previous estimates of electricity consumption for space heating could not account for the mixture of several energy sources for heating. The most useful result of this model is the estimates it produces of partial fuel use EUIs and the impact of alternative energy sources on the use of electricity for heat. Those who supplement their electricity use with wood or other sources, reduce their electricity consumption for heating by forty percent, or about 8000 kilowatt hours per year in single family houses. Those residences who convert from electric to gas heating will reduce their electricity consumption by as much as 18000 kilowatt hours per year if they used only electric heat initially. This number is substantially less than simple engineering calculations (or the Gas company) might suggest. If the hot water heater and dryer are also converted to gas, the savings in electricity consumption may total as much as 6000 kilowatt hours per year. A - 10 Appendix B: Residential Technology Curves This Page Intentionally Left Blank RELATIONSHIP OF EQUIPMENT PRICE TO UNIT ENERGY CONSUMPTION FOR RESIDENTIAL APPLIANCES, INCLUDING HEATING AND COOLING EQUIPMENT James E. {cMahon Author is a staff scientist at Lawrence Berkeley Laboratory, University of California, Berkeley, California, USA. (Institutional affiliation given for identification purposes only.) All opinions of those of the author. ACKNOWLEDGEMENTS: This report was prepared under funding from the University of Alaska, In- stitute of Social and Economic Research, Anchorage, Alaska. October, 1988 This Page Intentionally Left Blank B-2 Section APPENDIX A TABLE OF CONTENTS Name FOREWORD DATA SOURCES METHOD RESULTS FEDERAL APPLIANCE STANDARDS PAST EFFICIENCY CHANGES BEST AVAILABLE MODELS IN 1988 RETAIL PRICE AND UNIT ENERGY CONSUMPTION RESULTS B-3 Page This Page Intentionally Left Blank B-4 RELATIONSHIP OF EQUIPMENT PRICE TO UNIT ENERGY CONSUMPTION FOR RESIDENTIAL APPLIANCES, INCLUDING HEATING AND COOLING EQUIPMENT James E. McMahon FOREWORD For any residential appliance, many alternative designs are possible. Projections of future residential energy consumption can be based upon engineering estimates of the designs likely to be available. Since different designs have different costs, projections of the market penetration of new designs also require esti- mates of the prices of alternative designs. The U.S. Department of Energy (DOE) has commissioned studies of aternative designs of residential appliances, inclding heating and cooling equipment, during analyses of possible federal minimum efficiency standards for these products. This report draws upon the publicly available DOE reports to characterize the range of alternative designs, specifying for each appliance the annual energy consumption and retail price associated with alternatives differing in energy efficiency. By constructing smooth curve fits to the data, the results are transformed into parameters directly usable in the LBL Residential Energy Model (REM) for projecting future equipment efficiencies and prices. The reasonableness of the range of alternative designs is checked by comparing the highest efficiency included to the best efficiency currently available for sale in the U.S. in 1988. The National Appliance Energy Conservation Act of 1987 (NAECA) set federal minimum efficiency requirements for appliances. These standards apply to all appliances manufactured after a specified date (specific to each product). This report summarizes the federal requirements, and expresses them in units comparable to those used in characterizing the range of alternative designs. In addition, a comparison of the federal requirements with recent average efficiencies of appliances is made. B-5 DATA SOURCES The alternative designs of appliances are obtained from DOE reports: (1) U.S. Department of Energy, Assistant Secretary, Conservation and Renewable Energy, Test and Evaluation Branch, Consumer Products Efficiency Standards Engineering Analysis Document, March, 1982. (DOE/CE-0030) (2) U.S. Department of Energy, Assistant Secretary, Conservation and Renewable Energy, Test and Evaluation Branch, Supplement To: March 1982 Consumer Products Efficiency Standards Engineering Anaysis and Economic Analysis Documents, July, 1983. (DOE/CE-0045) Federal minimum efficiency requirements are from Public Law 100-12, 100th Congress [S. 83], March 17, 1987. Average efficiencies currently available are from press releases and reports of industry associations (some reported in DOE publications), including the Association of Home Appliance Manufacturers (AHAM), Air Conditioning and Refrigeration Institute (ARI), and Gas Appliance Manufacturers Associa- tion (GAMA). Best available efficiencies are from the American Council for an Energy Efficient Economy, "The Most Energy-Efficient Appliances," 1988 Edition (Washington, D.C.). METHOD The major sources of data for prices of residential appliances representing a range of alternative designs (differing in energy consumption) are from the U.S. Department of Energy. This report presents the results of collecting the publicly available reports, in a form compatible with the LBL Residential Energy Model for the PC. The products included are furnaces, air conditioners (room and central), water heaters (electric and gas), refrigerators (including refrigerator-freezers), and freezers. No review of the price and efficiency of clothesdryers, lighting, or miscellaneous products was conducted. (For these, the default values in the B-6 LBL-REM/PC data base are assumed correct.) For each product, such as refrigerators, DOE provides data for several classes, such as top mount auto-defrost refrigerator-freezers. For each class, a set of designs is specified, characterized by equipment price and unit energy consumption (or efficiency). The data from several classes are mapped together, appropriately weighted by relative sales, to construct a single representative relationship of equipment price to unit energy consumption for each product. RESULTS Parameters of the relationship between retail price and unit energy consumption (UEC). For each product, a page appears in Appendix A, containing two tables of data, a graph showing the data points and a curve fit, and the parameters of the curve fit. The products included are furnaces, air conditioners (room and central), water heaters (electric and gas), refrigerators (including refrigerator-freezers), and freezers. In all cases, the formulation is the exponential function used in the LBL-REM/PC. This formulation requires 2 parameters: A (called EALF in the LBL-REM program code) and EINF. These parameters are displayed at the top left of the page for each product. Below the parameters are 2 tables showing values of X and Y. X is energy consumption, and Y is equipment price. The first table contains values expressed relative to the first point, which has X and Y set to 1. The second table contains the absolute values. Units for energy consumption are million Btu/year for gas and oil appliances, and kWh/year for electricity. Prices are in 1985 dollars. The intent of this format is to estimate a curve fit that captures the dependence of relative price on equipment efficiency. In that way, local adjustments of the price or actual energy consumption can be made without the need to recalculate the parameters. B-7 FEDERAL EFFICIENCY STANDARDS Table 1 presents the minimum efficiency levels required by the National Appliance Energy Conserva- tion Act of 1987. The table shows, for each product and fuel, the first year in which the efficiency level must be met by manufacturers, and the efficiency level. For comparison, the average efficiency (shipment- weighted) for new units sold in 1985 is shown. Table 1 also shows the relative energy consumption of a new unit at the minimum efficiency level, relative to the first point (baseline unit) on the price/energy curves. This value can be input into LBL- REM/PC as the variable EUN, to restrict future appliances to comply with the federal regulation. Table 1. -- NAECA-87 Proposed Standards Efficiency Levels NAECA Minimum 1985 Baseline Appliance Year Efficiency SWEF_ Efficiency EUN Central Space Heater Gas 1992 78% AFUE 73.8 63.3 .812 Oil 1992 78% AFUE 78.6 75.8 971 Room Air Conditioner 1990 8.60 EER 720 6.72 .781 Central Air Conditioner 8.82 7.06 -706 Split System 1992 10.0 SEER Single Package 1993 9.7 SEER Heat Pump 8.56 7.06 -706 Split System 1992 10.0 SEER Single Package 1993 9.7 SEER Water Heater Gas 1990 54% AFUE 49.4" 47.9 .887 Electric 1990 88% AFUE 83.6" 78.3 .890 Refrigerator /Freezer 1990 7.52 EF 6.78 4.85 645 Freezer 1990 13.82 EF 11.55 9.73 704 Source: Senate Bill S.83 ~* LBL estimates for 1984 from partial data B-8 PAST SHIPMENT-WEIGHTED ENERGY FACTORS Table 2 shows the past average energy factors for each product and fuel, for each year from 1972 to -1987. For some products, several data sources are indicated, including DOE surveys, trade association sur- veys, and individual manufacturer’s product line. BEST AVAILABLE MODELS IN 1988 Table 3 shows the highest efficiency model available in 1988, the highest efficiency included in the equipment price/UEC curves, and the ratio (as a percent). For most of the products, the highest efficiency model currently available is 90% or more as efficient as the most efficient model considered in the equip- ment price/UEC relationship. This is a good indication that those relationships are composed of practical designs, not simply theoretical constructs with little likelihood of actual implementation. The exceptions are heat pumps, refrigerators, and freezers. For heat pumps, the DOE data are extended to approximately the same efficiency as for central air conditioners, but actual models available for heat pumps are lagging the efficiency of central air conditioners. For refrigerators and freezers, current models do not approach the efficiencies achievable by such measures as putting foam insulation in the doors, increasing the thickness of foam insulation (in walls and doors), more efficient compressor, and hybrid evaporator. Some experimental designs, notably in Denmark, have achieved efficiencies more com- parable with those proposed in the DOE studies than have current U.S. or Japanese models. B-9 d ‘ em Oo Table -- Shipment Weighted Energy Factors (SWEF) Appliance Source 1972 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 (787 Gas Central CS-179 62.7 - = -- 63.6 - 65.9 - = - -- - -- Space Heater Lennox -- 65.0 65.0 65.1 65.5 66.3 66.6 67.0 - -- -- - -- (AFUE %) Carrier = = «63.7 = - 651 663 66.7 66.5 = = = = = GAMA = = = = = = = - 69.6 730 738 743 75.1 Oil Central CS-179 73.6 - - ss 75.0 = 76.0 mt = = a = = Space Heater GAMA -- - -- -- -- - -- - -- 78.3 78.6 78.6 19-6 79.8 (AFUE 9%) Room Air CS-179 6.22 - - = 6.75 = 7.03 = = = = =< oa Conditioner AHLAM - -- -- 6.72 - 7.02 7.06 7.14 7.29 7.48 7.70 7.80 (EPR) Heat Pump ARI - _ 6.87 = 7.24 7.34 7.51 7.70 7.97 8.23 8.45 8.56 8.70 G.13 (SEER) Central Air CS-179 6.66 - -- -- 6.99 - 7.76 -- - - -- - - Conditioner Lennox -- 6.19 6.94 7.02 7.00 7.05 714 7.73 8.18 - - - - (SER) ARI 6.66 - 7.03 7.13 7.34 7.47 7.55 7.78 8.31 8.43 8.66 8.82 8.87 $.97 Electric Water CS-179 79.8 - - -- 80.7 - 81.3 -- - - 83.6 - - Heater (Percent) Gas Water CS-179 47-4 - -- -- 48.2 -- - - = = 49.4 = = Heater (Percent) Refrigerator CS-179 4.22 -- -- -- 5.09 -- 5.72 -- - -- -- -- - (cuft./kwh/day) AHAM 384 9 — = = 496 - 5.59 609 612 639 657 678 6.88 7.93 Freezer CS-179 8.08 - - -- 10.07 -- 10.83 -- - -- -- - == (cufu/kwh/day) ANAM 7.29 — = = 992 1085 13128366055 12.07, 22.93 Data Sources: ALAM - Association of Home Appliance Manufacturers ARI - Air-Conditioning and Refrigeration Institute Carrier - Carrier Corporation CS-179 - Department of Energy Survey of Manufacturers Lennox - Lennox Corporation Sn Table 3. COMPARISON OF BEST AVAILABLE U.S. MODELS IN 1988 TO HIGHEST EFFICIENCY DESIGN INCLUDED IN DOE STUDY Appliance Highest efficiency Highest efficiency Ratio Type available, 1988 in DOE study (Percent) Gas furnace 96 94.3 102% Oil furnace 90.8 91.9 99% Room air conditioner 12.0 12.5 96% Central air conditioner 15.0 16.7 90% Heat pump (SEER) 12.0 17.1 70% Electric water heater 97 97.9 99% Gas water heater 83 83.6 99% Refrigerator 8.7 23.2 38% Freezer 19.7 34.7 57% B- 11 This Page Intentionally Left Blank B- 12 RESIDENTIAL APPLIANCES: EQUIPMENT PRICE AND EFFIENCY James E. McMahon, Ph.D. Introduction. The major sources of data for prices of residential appliances representing a range of alternative designs (differing in energy consumption) are from the U.S. Department of Energy. This report presents the results of collecting the publicly available reports, in a form compatible with the LBL Residential Energy Model for the PC. The products included are furnaces, air conditioners (room and central), water heaters (elec- tric and gas), refrigerators (including refrigerator-freezers), and freezers. No review of the price and efficiency of clothesdryers, lighting, or miscellaneous products was conducted. (For these, the default values in the LBL-REM/PC data base are assumed correct.) Parameters of the relationship between unit energy consumption (UEC) and retail price. For each product, a page appears in this report, containing both two tables of data, a graph showing the data points and a curve fit, and the parameters of the curve fit. In all cases, the formulation is the exponential function used in the LBL-REM/PC. This formulation requires 2 parameters: A (called EALF in the LBL-REM program code) and EINF. These parameters are displayed at the top left of the page for each product. Below the parameters are 2 tables showing values of X and Y. X is energy consumption, and Y is equipment price. The first table contains values expressed relative to the first point, which has X and Y set to 1. The second table contains the absolute values. Units for energy consump- tion are million Btu/year for gas and oil appliances, and kWh/year for electricity. Prices are in 1980 dollars. The intent of this structure is to estimate a curve fit that captures the dependence of relative price on equipment efficiency. In that way, local adjustments of the price or actual energy con- sumption can be made without the need to recalculate the parameters, in most cases. B- 14 Shing FURNACES/BOILERS! GAS a EINF vALUE 4.3738 @.5243 SIGMA = 91.8527 6.0386 DELTA +8000 @.000e FLAMDA- 1,00000@1E-13 NO. x - VFIT 1 1.00038 1.0000 2 @.9478 1.0367 3° 0.8161 1.1033 1.1118 4 @.6715 1.2712 1.2682 NO x Y YFIT 1-91.38 2420.26 2420.26 2@ 86.S3 2509.02 2484.65 3 24.51 2670.18 2690.75 4 61.31 3076.58 3069.43 CHISOR- 2.2918839E-04 MORE INCREMENTAL PRICE MULTIPLIERS ? (Y/N) 1,e¢e@ 1.0266 (4 WEIGHT 1.0000 1.0008 1.0000 1.0000 WEIGHT 1.00 1.00 1.00 1.00 Price 14 1.0 FURNACES/BOILERS: GAS o7 oe O68 o4 Relstive Energy Use 02 01 AVo- Aili ti? Sled Furnaces/porters: O’L- a EINF VALLE 16.6728 @.8165 SIGMA 428.0023 1.4193 DELTA @.0220 @.0000 FLAMDA- 1.0@002@1E-09 NO. x Y YFIT 1 1.000@ 1.0000 1.0000 2@ @.9378 1.0248 1.0248 3 0.8245 1.1879 1.1879 NO. x Y YFIT 1 = 76.37 3508.46 3508.46 2 71.62 3595.59 3595.59 3. 62.97 4167.86 4167.86 CHISQGR- @.0000000E+00 EIGHT 1.0000 1.0820 1.0800 WEIGHT 1.00 1.00 1.00 MORE INCREMENTAL PRICE MULTIPLIERS ? FURNACES/BOILERS: 90/2 14 (Y/N) 12 Price 10 e o (012 (00) 08 (077 68 2 C6 (64. 091202 o4 Relstive Energy Use SMS Ll-4@ BI easy FOCM SIR CONDITIONERS A EINF VALUE $.e813 @.5544 SIGna 16.42e5 @.Se2? DELTA @.e0000 @.e020 FLAMDA- 1.0000e01E-09 _HO. x Y YFIT 1.0000 :.020@ 1.0020 @.9284 1.0417 1.0345 @.9084 1.0472 1.0453 @.8276 1.077e 1.0963 @.761@ 1.138@ 1.1513 @.7355 1.1863 1.1772 2.7266 1.172@ 1.1871 @.6986 1.2332 1.2221 @.6717 1.2856 1.2628 0.6533 1.3162 1.2964 @.6263 1.3932 1.3575 @.6239 1.5473 1.4326 8.5913 1.6265 1.4905 @.5385 1.7949 99.0200 ated SWUM OODVONU SUN Zz ° AWN OODVHNHAUMH- x ¥ YFIT 1167.85 $79.31 $79.31 2 603.48 599.29 1062.82 606.63 625.56 $66.52 623.92 635.10 888.76 659.26 €66.94 853.21 669.86 €81.95 848.56 678.95 €87.72 815.84 714.40 727.97 784.42 744.74 731.55 2762.50 762.46 751.02 732.11 807.08 786.41 7@5.2? 896.37 829.92 €3@.55 942.23 863.45 €28.92 1039.8157351.69 CKISGR- 4.8477523E-04 1 1 1 1 1 MORE INCREMENTAL PRICE MULTIPLIERS ? (Y/N) UEIGHT 3.0088 2.0808 1.0000 1.0002 1.0082 1.0000 2.00200 2.0eee 1.0ee@ 1.0002 2.0002 2.0002 1.0002 3.0000 UEIGHT 3.20 2.02 1.02 1.02 1.02 1.00 2.28 2.20 1.22 1.20 2.02 2.e2 1.00 3.20 10 ROOM AIR CONDITIONERS o7 o@ 08 O48 Reistive Energy Use Qt 81-4 CENTRAL AIR CONDITICNERS A EINF VALUE 1.4218 @.2414 SIGHA @.936? @.2S61 beta see CENTRAL AIR CONDITIONERS FLANDA- NO. x Y YFIT UEIGHT 1 +0000 1.0200 2.0000 @ 1282 1.0512 1.0000 e +1586 1.1020 1.0000 e +2217 1.1835 1.0008 @ +2527 1.2491 1.0008 @. +2839 1.2899 1.0000 @.7095 1.3184 1.3397 1.e000 @.6880 1.3525 1.3729 2.0000 @.6499 1.4172 1.4356 1.0000 @.6318 1.4486 1.4674 2.0000 @.5924 1.5228 1.5424 2.0000 @.5743 1.5677 1.5798 1.0080 @.5483 1.6425 1.6368 2.0000 @.5099 1.7191 1.7309 3.0000 0.4678 1.8582 1.8511 2.0e0ee 0.4434 1.9451 1.9314 2.0000 @.4225 2.0397 2.0081 2.0000 VON SLYV$COHOVOHNAwWNM— oe oe oe oe oe oe oe oe Zz o x NOUSWVKSOOVHUsAWWH-+ Y YFIT UEIGHT 4439.99 1334.53 1334.53 4159.S@ 1505.68 1416.19 3985.64 1546.22 1470.62 3666.80 1630.34 1579.46 3435.89 1671.83 1666.99 3302.79 1713.35 1721.41 3150.24 1759.41 1787.94 3854.62 1805.01 1832.17 2885.63 1891.30 1915.84 2805.4@ 1933.18 1958.33 2630.12 2032.19 2e58.44 2549.67 2092.10 2108.23 2434.63 2191.97 2184.34 2264.12 2294.13 2309.88 2076.92 2479.89 2470.30 1968.SS 2595.81 2577.46 1875.97 2721.98 2679.81 CHISQR- $.9418491E-04 WVU UMH WM Vee meet rere et tt td MORE INCREMENTAL PRICE MULTIPLIERS ? (Y/N) 10 O08 O08 OF Of O08 O04 O3 oO2 a oo Reletive Energy Use 61-4 HEAT PUMPS a VALUE 1.7388 @.2538 SIGMA 0.0044 eo eo eo = = DELTA 0.0208 —-@. 0800 FLAMDA- Si . NO. 672 WR SOHVETNsAWH eaten ee 4447.76 3958.36 3737.38 3621.28 3411.64 3261.22 3128.00 2990.14 2818.27 2703.41 2569.96 2487.47 2325.82 2 24 2098.65 1963.82 1832.37 Ferner ee pepe meme me VON SWUM OOH VHVAWNM--- 8 MORE ANCRERENTAL BRACE PULTIPLIERS 7? (Y/N) 4 (pio 4 ] pw J 1.2483 1.25 1.2756 1.29 1.3705 1.3326 1.4010 1.38: 1.4204 1.42 1.4459 1.47: 1.4867 1.51 1.5610 1.58 1.6103 1.63! 1.7224 1.70 1.8014 1.79 1.8968 1.89, y YFI 1515.64 1515. 1654.63 1654. 1722.37 1725. 1769.79 1765. 1843.62 1841. 1891.93 1901. 1933.30 1957. 2077.17 2019. 2123.34 2104. 2152.80 2165. 2191.51 e242. 2253.33 2294. 2365.86 2404. 2440.63 2478. 2618.59 2587. 2730.27 2718. 2874.80 2867. - 2@.S$355074E-04 ? ( jee 1.00 maa tate rhe ees SSSSSSSSSSSSES 10 os os o7 HEAT PUMPS oe 0&6 04 Relative Energy Use 02 o1 UATER HEATERS! ELECTRIC VALUE 10.5477 SIGMA 2.2273 DELTA @.0000 FLAMDA- 10.00000 i 2 3 4 5 NO. x 1 5093.86 2 4449.81 3 4349.81 4 4162.72 S 4075.09 wo N o 223.49 239.62 248.83 258.05 269.57 CHISQR- 5.4253698E-05 EIN 0.7768 @.0670 6.0000 VFIT 1.0000 1.0788 1.1001 1.16@5 1.2417 VFIT 223.49 241.16 245.86 es9.35 270.79 MORE INCREMENTAL PRICE MULTIPLIERS ? (Y/N) & § ee & eeieal Lo WATER HEATERS: ELECTRIC oe 08 4 Relstive Energy Use MC Lhe “UP? LG ick : . es WATER HEATERS: GAS (DELETE PTS 6,7,8) NG MP o.sre BE 0.0076 le000 WATER HEATERS: GAS (DELETE PTS 6,7,8) ¥ YFIT 1.6000 1.0000 1.0583 1.0943 1.1000 1.1085 1.1687 1.2280 1.2014 3.9167 99.0000 ~ $3823e" ONaAwWRe> v YFIT 24.03 241.92 241.92 20.91 256.03 264.73 20.53 266.11 268.18 19.14 286.27 282.73 18.S1 296.35 290.64 13.77 947.5223950.08 CHISQR- 6.185234S5E-04 MORE INCREMENTAL PRICE MULTIPLIERS ? (Y/N) 14 1@ 18 20 22 24 26 28 3.0 32 34 30 38 40 42 12 1o o8 o8 oF Of O8 O48 OF o2 @ @ Rolstive Energy Use REFRIGERATCR-FREEZERS A EINE VALLE 10.1276 @.2385 SIGMA @.0330 DELTA 2000 FLAMDA- NO. x YFIT 1 ee0e 1.8000 2 0.9026 1.0135 3 6.8861 1.0160 4 0.8468 1.0222 S 0.7987 1.0303 6 0.7176 1.0459 2? 0.6946 1.0506 2 0.6050 1.0722 9 0.5498 1.0886 10 0.5246 1.0967 11 0.4954 1.1073 12 4350 1.1338 13 @.4039 1.1508 14 @.3656 1.1768 15 @.3306 1.2086 a 16 @.3032 1.2435 ' 17 @.2093 99.0000 N N NO. x Y YFIT 1 1427.85 763.81 763.81 2 1288.71 771.12 774.14 3 1265.22 772:¢7 776.03 4 1209.07 774.28 780.75 $ 1140.38 781.7 786.97 6 1023.77 786.91 798.85 7° (991.85 807.37 9802.46 @ 863.91 825.03 818.96 9 783.87 833.40 831.47 1@ 749.06 837.12 837.64 11 707.39 848.11 845.76 12 621.06 89 865.99 13. $76.75 880.78 878.96 14 $22.08 @ 898.81 1s 472. 4 923.16 16 432.92 8 949.77 1? 298.88 -6275617.19 CHISQR- 2.3425773E-83 MORE INCREMENTAL PRICE MULTIPLIERS ? (Y/N) VEIGHT 2.00 1.00 1.00 1.00 2.08 2.00 2.00 2.00 1.00 1.00 1.00 1.20 4.00 1.00 2 10 REFRIGERATOR-FREEZERS 07 oe 08 04 Reletive Energy Use 02 ot FREEZERS sana na = DELTA E ZERS FLANDA- FRE NO. x % YFIT WEIGHT 1 1.0000 1.000@ 1.6¢2@ 2.0000 2 @.8618 1.0281 1.@287 1.0000 3 @.7981 1.0466 1.0443 1.0008 4 @.7481 1.0526 1.0577 2.0000 S @.6751 1.0630 1.801 1.0000 6 @.6046 1.0915 1.1856 3.0000 ? @.5834 1.1392 1.1143 2.0000 8 @.5393 1.1670 1.1343 2.0000 9 0.4464 1.1876 1.1887 2.0000 18 4357 1.1926 1.1966 2.0000 11 3936 1.2174 1.2326 2.0000 12 @.2805 1.3856 1.4462 2.0000 nO. x Y YFIT 1 974.93 470.46 470.46" e. 2@ 840.22 483.70 483.98 1. 3. 778.07 492.37 491.29 1. 4 729.38 495.22 497.63 2. 5 658.28 500.09 Se8.14 1. 6 $89.49 $13.52 $20.15 3. ? $68.74 $35.97 $24.25 2. 8 525.76 $49.03 $33.65 2. 9 435.18 558.73 $59.26 e. 10 424.75 $61.07 562.96 2. 11 383.69 $72.72 579.91 e. 12 273.43 651.85 680.38 2. CHISOR- 2.2205820E-03 MORE INCREMENTAL PRICE MULTIPLIERS ? (Y/N) T = gee 7 10 08 6& oF Of 68 O44 O38 o2 Of @ Reletive Energy Use Appendix C: End Use Model Summary Output This appendix contains end use model summary output at the regional level. The tables are organized at the highest level by region, then by sector, then by case. This page intentionally left blank AKREM Output Summary Region: ANCHORAGE Case: LOW PART I 1987 1988 1990 1995 2000 2005 2010 Occupied Housing Stock (000) Single 36.3 36.5 36.9 36.3 34.3 41.3 45.7 Multi S515) 34.1 5334 3730 39.9 42.5 45.5 Mobile Shel wat 5.8 6.2 63) 6.6 6.9 TOTAL Ties 76.3 76.1 73.8 $3.7 90.4. . 98.1. New Housing Units (000) Single 0.0 0.0 0.0 0.0 0.7 0.9 do Multi 0.0 0.0 0.0 0.0 O..5 0.7 0.8 Mobile 0.0 0.0 0.0 0.0 0.1 0.1 Ont. TOTAL 0.0 0.0 0.0 0.0 dae p By / 2.0 Sales by End Use (Gwh/yr) Heat 132 128 122 113 106 100 96 Water 85 81 76 78) 73 70 69 Frig 93 91 87 82 82 84 89 Freez 45 45 44 44 43 42 43 Cook 38 37 37 39 40 42 44 Dry 59 58 58 58 58 59 63 Lite 97 96 96 99 104 13 124 Misc 162 160 160 165 175) 193 214 TOTAL SALES 711 697 680 673 680 704 742 Usage per Customer 9,191 9,131 8,930 8,465 8,132 7,789 7,568 New Equipment Electric Shares HEAT.) New Buildings (7) ) |) coe Giese cto losers ioe =| ecco) ee Single O05) (005) | 0.05 + 10.05) | | (0.105 210,05 Multi O.10) || O10) | 0510) <10.-10) O10" 70.10 Mobile 0510) |) O10) |) 0510 | 0.10! | | 0.210 > ~10- 10 OTHER Water 0.22, 0827 | 05207, OAl9| 4.0517 || 0.17 Frig 12.03) 2.103) || 203" | AF-03" ~ 1502) 1. 1202 Freez 0257 | O58) | 0557) On57) 30.574 O57 Cook On 72 | 0.722; (0.72) | 0569) | *0,.677.' O66 Dry O62 (7) 02615) 1/0459) 10859) | 0.5772 0255 Lite 1500 | 1.00 * 0.99 %570.98 0.98 0.97 Misc 1.00 1.00 0.99 0.98 0.96 0.94 Residential (AKREM) and Commercial (COMMEND) End Use Model Output Call AKREM Output Summary Region: ANCHORAGE Case LOW PART II 1988 1990 1995 2000 2005 2010 Average Electric Shares of total housing market Heat Shares by building type Single 0.06 0.06 0.05 0.04 0.04 0.03 Multi 0:26 0723) Of27 10.19 Oc. LIS Mobile 0.05 0.04 0.02 0.01 0.00 0.00 Water 0.21 0.20 0.19 0.18 0.16 0.15 Frig 1.03 1.03 2.03 21.03 1.03 2.03 Freez O557 10557 . 0557 10:57 O.57 0.57 Cook 0.73 0.73 0.74 #O.72 O.70 0.69 Dry 0.67, 0.67 0.65 0.62 0.59 0.58 Lite 1.00 1.00 1.00 0.99 0.98 0.98 Misc 1.00 1.00 1.00 0.99 0.98 0.97 NEW EQUIPMENT EUIs (Kwh/House/Yr) Heat by building type Single 18447 18345 17989 17918 17810 17637 Multi 8418 8371 8209 8178 8145 8087 Mobile 10851 10791 10581 10536 10415 10232 Other End Uses Water 4777 4512) 4461 44424438 = 4423 Frig 945 853 854 858 854 853 Freez 897 753 Jou 752 753) 754 Cook 650 649 645 644 643 642 Dry 1100 1099 1093 1091 1089 1086 Lite 1222 1221 1214 1220 1230 = 1233 Misc 1951 1948 1936 1931 1924 1915 AVERAGE EQUIPMENT EUIs Heat by building type Single 18447 18345 17989 17983 18101 18236 Multi 10010 9954 9761 9712 9648 9567 Mobile 10851 10791 10582 10668 11280 11667 Other End Uses Water 4866 4824 4625 4510 4443 4413 Frig 1123 1076 973 926 883 858 Freez 999 994 955 883 799 757 Cook 650 649 645 644 643 642 Dry 1100 1099 1093 1091 1089 1086 Lite 1226 .~1230. 1215 1218 1236 £1257 Misc 1952, 1956 1974 2021 2086 2156 INDEX VALUES (1987=1) 1987 1988 1990 1995 .2000 2005 2010 Housing Stock 1.000 0.987 0.984 1.028 1.082 1.169 1.269 Use per House 1.000 0.993 0.972 0.921 0.885 0.847 0.823 Total Sales 1.000 0.980 0.957 0.947 0.957 0.990 1.045 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-2 AKREM Output Summary Region: ANCHORAGE Case: MIDDLE PART I 1987 1988 1990 1995 2000 2005 2010 Occupied Housing Stock (000) Single 36.3 36.5 36.9 36.3 40.9 47.1 55.0 Multi 351.3) 34.0 33 25 39.5 42.3 46.6 BZ at Mobile auf Sid. 5.8 6.2 6.6 rik 7.7 TOTAL 771.3) 76.2 76.1 82.0 89.8 100.7 114.9 New Housing Units (000) Single 0.0 0.0 0.0 0.0 1.3 pI 1.8 Multi 0.0 0.0 0.0 0.0 0.9 La Lae Mobile 0.0 0.0 0.0 0.0 0.1 0.1 0.2 TOTAL 0.0 0.0 0.0 0.0 25) 2:8 350 Adjusted Sales by End Use Heat 132 128 122 119 112 110 aS Water 85 81 76 76 75 73 76 Frig 93 91 87 84 88 94 104 Freez 45 45 44 46 46 48 52 Cook 38 37 a7 40 42 46 51 Dry 59 58 58 60 63 68 78 Lite 97 96 96 102 113 130 152 Misc 162 160 160 169 191 222 263 TOTAL SALES 711 696 680 696 731 792 889 Kwh per Customer 9,191 9,130 8,936 8,492 8,144 7,863 7,741 New Equipment Electric Shares HEAT: New Buildings = ------ rene re werner renee e eee e eee eee Single 0.05 0.05 0.05 0.05 0.05 0.05 Multi 0.10 0.10 0.10 0.10 0.10 0.10 Mobile 0.10 0.10 0.10 0.10 0.10 0.10 OTHER Water 0.22 0.21 0.20 0.17) 0.16 0.16 Frig 1.03 1,03. 103) 1.03 -1.03,_ 1.08 Freez 0.57 0.58 0.58°7'0.58 0.59 0559 Cook 0.71 O.72 O.72 0.68 0.66 0.62 Dry 0.61 O.61 0.60 0.61 0.60 0.60 Lite 1.00 1.00 1.00 0.99 0.99 0.99 Misc 1.00 1.00 1.00 1.00 1.00 1.00 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-3 AKREM Output Summary Region: ANCHORAGE Case: MIDDLE PART II 1988 1990 1995 2000 2005 2010 Average Electric Shares of total housing market Heat Shares by building type Single 0.06 0.06 0.05 0.04 0.04 0.04 Multi 012477) 10-23) 0.2177) OnkS | 1016) |) One Mobile 0.105) | (0,04 | | 0702)" (0401 | 1110.01"! || O01 Water 0.20) (0.20) 0,19 057 10.15) On14 Frig LOSI) OS 70311111103) F103 05) Freez 0.157) 057) 1110.57) || 0557 | 10-58) (0758 Cook 0.73) 03730. 74-0571 70.69) 0.167 Dry 0.67 0.67 0.65 0.62 0.60 0.60 Lite 1.00) 1.00 | 4..00;7 (71.100 | {1007 0.199 Misc 1-005-100-2100) 1200) 2-00 1.500 NEW EQUIPMENT EUIs (Kwh/House/Yr) Heat by building type Single 18447 18355 18049 17998 17929 17797 Multi 8418 8376 8236 8223 8213 8180 Mobile 10851 10797 10617 10577 10499 10370 Other End Uses Water 4777 =4514 4468 4450 4457 4451 Frig 945 853 854 859 856 857 Freez 897 DS 752 753 192) 756 Cook 650 649 646 645 645 644 Dry 1100 1100 1096 1096 1097 1098 Lite 1222, 1222, | «1217 2361245 1.249 Misc 1951 1949 1942 1940 1938 1934 AVERAGE EQUIPMENT EUIs E Heat by building type Single 18447 18355 18049 18166 18382 18649 Multi 10010 9959 9793 9736 9689 9633 Mobile 10851 10797 10617 10817 11228 11665 Other End Uses Water 4866 4825 4631 4521 4459 4440 Frig 1123 1076 972 925 884 861 Freez 999 994 953 877 797 759 Cook 650 649 646 645 645 644 Dry 1100 .1100 1096 1096 1097 1098 Lite 1226) 1231 1213 77123471263, 1299) Misc 1952 1957. 1980 2042 2119 2207 INDEX VALUES (1987=1) 1987 1988 1990 1995 2000 2005 2010 Housing Stock 1.000 0.985 0.984 1.061 1.161 1.302 1.486 Use per House 1.000 0.993 0.972 0.924 0.886 0.856 0.842 Total Sales 1.000 0.978 0.955 0.980 1.029 1.114 1.250 Residential (AKREM) and Commercial (COMMEND) End Use Model Output Cc-4 AKREM Output Summary Region: ANCHORAGE Case: HIGH PART I 1987. 1988 1990 1995 2000 2005 2010 Occupied Housing Stock (000) Single 3643:-°36.5 36.9 ° 37.8 44.3 52.2” <60).3 Multi 3543. 34.1. ..34.3 -40.2. 44.7 ° 50.12" ". 55.9 Mobile a. a 5.8 6.3 6.8 Tow 8.2 TOTAL 77.3.. 76.3 77.0 84.3: 95:8 109.9 124.4 New Housing Units (000) Single 0.0 0.0 0.0 1.6 1.7 1.6 1.9 Multi 0.0 0.0 0.0 1.1 ia a 1.4 Mobile 0.0 0.0 0.0 0.1 0.2 0.2 0.2 TOTAL 0.0 0.0 0.0 2.8 31.2 310) Sn4 Sales by End Use (Gwh/yr) Heat 131.5 128.1 124.2 120.0 116.7 117.5 119.9 Water 85.1 *81.2 76.6 77.5% 79.4** 8052 ““85).1 Frig 93.2 90.7 87.6 86.9 93.5 102.2 112.3 Freez 45.3 44.7 44.8 47.1 49.1 51.4 55.3 Cook 37.8 37.4 37.7 40.9 44.6 49.3 54.2 Dry 58.9 58.0 58.4 61.2 66.7 73.5 81.7 Lite 96.9 96.1 97.1 104.7 120.6 140.5 160.2 Misc 161.9 160.4 161.2 173.9 202.7 238.8 274.1 TOTAL SALES 710.5 696.5 687.6 712.2 773.3 853.4 942.7 Usage per Customer 9191 9130 8928 8446 8074 7764 7581 New Equipment Electric Shares HEAT: New Buildings = = ------- wenn ne ween ee wee eee eee eee ------ Single 0.05 0.05 0.05 0.05 0.05 0.05 Multi 0.10 0.10 0.10 0.10 0.10 0.10 Mobile 0.10 0.10 0.10 0.10 0.10 0.10 OTHER Water 0.22 0.21 0.17 0.17 0.17 0.17 Frig 1.03 1.03 1.03 1.03 1.02 1.02 Freez 0.57 0.58 - 0.58, .0.58 0.57°.°'0:57 Cook 0.71 O.72 0.65 0.68 0.66 0.62 Dry 0.61 O.61 0.60 0.60 0.58 0.57 Lite 1.00 1.00 0.99 0.99 0.98 0.97 Misc 1.00 1.00 0.99 0.99 0.97 0.93 Residential (AKREM) and Commercial (COMMEND) End Use Model Output c-5 AKREM Output Summary Region: ANCHORAGE Case: HIGH ~ PART II 1988 1990 1995 2000 2005 2010 Average Electric Shares of total housing market Heat Shares by building type Single 0.06 0.06 0.05 0.04 0.04 0.04 Multi 0224-0323 (0-21-0218 0-16 0.14 Mobile (02055 0504-008 = 0.020.201 10-01 Water O%215 == 0520 == 0719-0 18) = OF 1610.15 Frig E.03- 1.03 1.03. 1-03-1103 __-1.02 Freez OS ea Or a1) OE =O O17, Cook 0:73 0.75 0.73 0.71; 0.68 0-66 Dry 0367- 0-67 — (0/65 — 0-62'- 0.60 10.59 Lite 1.00 1-00 1.00 0.99 0-98 0:97, Misc 17005 1500; 1-00 -0:99'---0599-—-0.97, NEW EQUIPMENT EUIs (Kwh/House/Yr) Heat by building type Single 18447 18321 17890 17785 17623 17393 Multi 8418 8360 8164 8132 8084 8004 Mobile 10851 10777 10523 10448 10325 10160 Other End Uses Water 4777 4507 4427 4410 4409 4388 Frig 945 853 857 857 854 854 Freez 897 753 751 752 753) D3 Cook 650 649 644 643 642 639 Dry 1100 1099 1093 1092 1089 1084 Lite 1222 122) 1225 1236) 1239-1237, Misc 1951, 1948 1936 193119231913: AVERAGE EQUIPMENT EUIs Heat by building type Single 18447 18321 17925 18027 18163 18263 Multi 10010 9941 9692 9601 9504 9394 Mobile 10851 10777 10563 10744 10996 11227 Water 4866 4816 4599 4477 4406 4374 Frig 1123. 1076 975 924 881 858 Freez 999 994 951 873 793 756 Cook 650 649 644 643 642 639 Dry 1100 1099 1093 1092 1089 1084 Lite 1226 1228 1213 1237 1265 1292 Misc 1952 1956 1979 2041 2112 2184 INDEX VALUES (1987=1) 1987 1988 1990 1995 2000 2005 2010 Housing Stock 1500-0599" 2-00 1.09 1-26 1.42, 161 Use per House 1.00 0.99 0.97 0.92 0.88 0.84 0.82 Total Sales 005 =0598=-=.0,9/= ==) 0027109 == 1-20 1-95) Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-6 COMMEND Output Summary Region: ANCHORAGE Case: LOW Floorstok (Million Ft2) 55.0 54.8 54.7 57.2 58.9 63.0 67.3 New Equip. Elec. Shares (%) HEAT (1) (1) 3.4 3.6 4.1 4.2 4.1 WATR (1) (1) 20.8 21.4 23.8 25.7 25.8 Average Elec. Shares (%) HEAT 6.8 6.8 6.7 6.3 5.8 5.1 5.0 WATR 23.5 23.4 23.4 23.4 23.5 23.9 24.3 New Equipment EUI (Kwh/Ft2/Yr) HEAT (1) (1) 7.1 6.9 6.6 6.5 6.5 coOoL (1) (1) 0.9 0.7 0.6 0.6 0.6 VENT (1) (1) Lad Ld iu? 13 a WATR (aly) Gd) 1.9 LD dnd) iis ome COOK (1) (1) 1.3 0.7 0.6 0.5 0.5 REFR (1) G@) 1.9 1.6 ait 1.4 1.4 LITE (1) (1) 6.7 6.3 6.2 6.0 5.8 MISC (@) (1) 2.8 a2 2.0 1.9 1.8 Average EUI (Kwh/Ft2/Yr) HEAT 10.0 8.1 8.0 7.2 To 6.7 6.6 COOL 0.7 0.7 0.7 0.7 On7 0.6 0.6 VENT 2.4 2.4 2.4 2.3 2, 2.0 1.9 WATR 1.3 1.3 1.8 ia hud 1.3 ia COOK 0.4 0.4 0.4 0.4 0.5 0.5 0.5 REFR 1.7 1.7 1.7 1.6 ig 1.3 1.4 LITE 7.8 Pub 7.6 7% 6.3 6.3 6.1 MISC 1.8 1.8 1.8 1.8 1.9 1.9 1.9 Electric Intensity (Kwh/Ft2/Yr) ; HEAT 0.5 0.5 0.4 0.4 0.4 0.3 (RK COOL 0.7 0.7 0.7 0.7 0.7 0.6 0.6 VENT 2.4 2.4 2.4 2.2 mak 2.0 1.9 WATR 0.2 0.2 On 0.2 0.2 0.2 0.2 COOK 0.4 0.4 0.4 0.4 0.5 0.5 055) REFR Lad hat 1.7 1.6 4.3 p 1.4 LITE 7.8 7.8 736 Vik 6.5 6.3 6.2 MISC 1.8 1.8 i aut 2.3 2.6 i TOTAL 15.4 15.4 15.3 14.8 14.2 13.9 13.8 Adjusted Electric Sales (Gwh/yr) HEAT 31 a 30 28 27 23 23 COOL 49 49 49 49 48 49 52 VENT 163 161 159 158 152 152 159 WATR 16 16 16 16 17 18 20 COOK 25 25 26 30 33 37 41 REFR LS 112 111 113 112 113 120 LITE 526 525 514 499 472 489 510 MISC 120 120 124 147 169 198 223 TOTAL ELECTRIC SALES 1043 1039 1029 1040 1030 1081 # 1147 Note (1): Due to lack of new construction during these years, weighted averages across building types are not meaningful. Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-7 COMMEND Output Summary Region: ANCHORAGE Case: MIDDLE 1987 1988 1990 §=1995 2000 2005 2010 Floorstok (Million Ft2) 55.0 54.9 54.7 5716) 621.2 68.5 77.3 New Equip. Elec. Shares (%) HEAT (1) (1) 4.2 4. WATR (1) (1) 24.5 23% Average Elec. Shares (%) HEAT 6.8 WATR 23.5 a3. fo i] Ww fn nN w ™~ np + own nN ru er NR fu wo New Equipment EUI (Kwh/Ft2/Yr) HEAT (1) (1) cooL (1) (1) VENT (1) (1) WATR (1) @) COOK (1) (1) REFR () Gt) LITE (1) @) MISC Gis) (@5) Average EUI (Kwh/Ft2/Yr) HEAT COOL VENT WATR COOK REFR LITE MISC Electric Intensity (Kwh/Ft HEAT COOL VENT WATR COOK REFR LITE MISC TOTAL NVAFORRFOD FAWONA UG NARFPORPROD CWUUNWBOAN PArRPORROD OWE ENOUYW FPArFPORPFON OOP EFNOUY PUPORPROD CorrFNoUs r UrRPNRPOONOCONRFNRFORFRNOO rPNPRPOrRNOO rPNPRPOrRNOO CONN FE WENE ONDDEFWEHENO EPs F OF Nn O~* COrRrDFWNHND rar OrNON OoNDNFWRrNND rPaArRPOrRNOD wowWNFNOANY RFPAOrRrOrRrF OO worFrEnNODA VL UrRNRPOONOO FoOomNFneuu UrRNRPOONOO NONDDFNFENSES FNUNRPOONOCO CORRrFAFNNNS FNArFOONOO NPFDDENRPASHLE WNHNHArFOOrFODOO CUWUNFENYNODW WNHArFOCOrRPOCO NNRPPFEFNODW » is » es = = Adjusted Electric Sales (Gwh/yr) HEAT 31 31 30 30 28 25 27 COOL 49 49 49 48 49 Sw 5D VENT 163 163 159 158 158 164 181 WATR 16 16 16 7. 18 20 22) COOK 25 29 26 28 33 37 42 REFR 113 112 niall LU5 118 123 136 LITE 526 525 514 505 503 534 585 MISC 120 120 124 148 180 213 253 TOTAL ELECTRIC SALES 1043 1040 1029 1049 1088 1166 1301 Note (1): Due to lack of new construction during these years, weighted averages across building types are not meaningful. Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-8 COMMEND Output Summary Region: ANCHORAGE Case: HIGH 1987)! 1988))))))'990)))))//'1995)))))||2000)))))))'2005 |)))(2010 Blloorstok)) (Million Ft2))| ||| S24 Gil) i) See) eS StL) 0,25) 68120) 17710 3-19) |) | SAernO New Equip. Elec. Shares (%) HEAT (1) (1) 5.0 4.0 4.0 a 4 WATR (1) (1) 26.8 6 3 Average Elec. Shares (%) HEAT 6.8 6.8 6.6 6.12 5.9) 4.9 4.7 WA TRY |/)/-23) 51H ll (23 105)) I) k2-Sta9, I /23!20 7A N23) 5H) 2 Ste Oe |S New Equipment EUI (Kwh/Ft2/Yr) HEAT (1) (1) 615) 6.6 6.5 6.4 6}s3) COOL (1) (1) O75) 0.5 0.6 0.5 0.6 WENT) )/(G1)) (1) eT, 9 Lag) 1.8 8 WATR (1) (1) a aa Bz, Li 2 1.2 COOK (1) (1) 0.5 0.4 0.4 0.4 0.4 REFR (1) (a) 1.6 1.4 io) 1.4 1.4 tate | RD (1) 6.8 6.3 62) 6.0 Se) MISC. § (1) (1) aan 2.0 r9) 18 LAS Average EUI (Kwh/Ft2/Yr) HEAT 10.0 a 8.0 726 io 6.6 6.5 COOL 0.7 (47) 0.7 ..7 0.6 0.6 0.6 VENT 2.4 2.4 2.4 2.2 2.0 me +0 WATR oot eS ets) nhs} Le — 4,2 COOK 0.4 0.4 0.4 0.4 0.4 0.4 0.4 REFR aie aud Le? 1.6 kes Lad 1.4 LITE a8 7.8 Boe 7.8 675) 6.3) 6.1 MISC 8 1.8 aS 9 Ls9 ad 1.9 Electric Intensity (Kwh/Ft2/Yr) HEAT 0.5 0.5 0.4 0.4 (ys) 0.3) 0.3) COOL 0.7 On On7 0.7 0.6 0.6 0.6 VENT 2.4 2.4 233 22 2.0 eg) 1.8 WATR O;:2) 0.2 0.2 On) Os2 On2) 0.2 COOK 0.4 0.4 0.4 0.4 0.4 0.4 0.4 REFR eT TAZ Loy, a6 Leo) 1.4 1.4 LITE To 738 7/26 Tk .2 6.2 6.0 MISC 1.8 Le Li9 2k 2.4 25) Dial, TOTAL. | )/e:5)x45)) 11): 5)505 lS oto (N/a 7 NL LS AG esa Adjusted Electric Sales (Gwh/yr) HEAT oi) 30 30 30 28 26 a COOL 49 49 49 Si 52 55 59 VENT 161 160 159 161 168 177 191 WATR 16 16 16 7 20 21 23 COOK 25 25 26 31 36 41 44 REFR 113) 112 112 121 129 137 147 LITE 525 524 519 527 542 588 627 MISC 120 120 127 158 Uo; 239 276 TOTAL ELECTRIC SALES 1040 1035 1038 1096 1172 1284 1395 Note (1): Due to lack of new construction during these years, weighted averages across building types are not meaningful. Residential (AKREM) and Commercial (COMMEND) End Use Model Output c-9 AKREM Output Summary Region: FAIRBANKS Case: LOW PART I 1987 ©1988 1990 1995 2000 2005 ‘2010 Occupied Housing Stock (000) Single 2ac0* 158 13,7 SL UE 19.7 Multi 6.7 6.7 Tek BiD 9.1 9.9 10.9 Mobile z.3 a2 2.4 22 2.6 ey, 2.8 TOTAL 24.5 26.7% §=Sb alt 26.6 25.3 8057, 33.4 New Housing Units (000) Single 0.0 0.0 0.0 (ane 0.3 0.4 0.4 Multi 0.0 0.0 0.0 OT 0.2 0.2 o.2 Mobile 0.0 0.0 0.0 0.0 0.0 0.0 0.0 TOTAL 0.0 0.0 0.0 OR2) 0.5 0.6 0.7 Sales by End Use (Gwh/yr) Heat 18 18 19 20 2 25) 25) Water 40 43 45 46 43 43 46 Frig 28 28 29 29 28 27 29 Freez 15 1S 15 15) 15 15 16 Cook 11 lal: 12 13 13 14 LS) Dry 19 19 19 20 2: 23 25 Lite 30 31 31 33 35 39 43 Misc 49 49 50 53 58 64 712} TOTAL SALES a0 215 220 228 234 248 271 Usage per Customer 8,595 8,718 8,732 8,562 8,261 8,093 8,114 New Equipment Electric Shares HEAT: New Buildings = -wwn ne ween renee weer cee eee ee eeee Single 0.04 0.04 0.04 0.04 0.04 0.04 Multi 0.09 0.09 0.09 = 0.09 0509 0.09 Mobile 03025205022 0202-42 0.022 050252 10.02 OTHER Water 0 333i= OSS) 05926 210.501 0.30 0.30 Frig 1703103) 03 1.03108 1303 Freez 0.63 0.64 0.64 0.64 0.64 0.64 Cook O- 74 0572 ~ 0777 0.78 0.61, 0578 Dry 0.74. 05735 = 0574:- 0-73 40:74. 207 Lite 1-00" 1-00 1:00 = 1.00 1.00 0:99 Misc 1.005100 100 0,99. 20:99) 0798 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-11 AKREM Output Summary Region: FAIRBANKS Case: LOW PART II 1988 1990 1995 2000 2005 2010 Average Electric Shares of total housing market Heat Shares by building type Single 0.04 0.04 0.04 0.04 0.04 0.04 Multi 0.09 0.09 0.09 0.09 0.09 0.09 Mobile 0.02 0.02 0.02 0.02 0.02 0.02 Water 0.36 0.37 0.37 0.33 0.31 0.30 Frig 1.03 1.03 1.03 1.03 1.03 1.03 Freez 0.63 O.63 0.63 0.63 0.64 0.64 Cook 0.77. 0.78 O.80 0.78 0.76 0.76 Dry 0.74 6,73 6.72 |} 0.71 0.71 £@.71 Lite 1.00 1.00 1.00 1.00 1.00 0.99 Misc 1.00 1.00 1.00 1.00 1.00 0.99 NEW EQUIPMENT EUIs (Kwh/House/Yr) Heat by building type Single 20000 19997 20013 19976 19990 19951 Multi 9000 8999 9006 8996 9014 9007 Mobile 10000 9999 10006 9990 10002 9988 Other End Uses Water 4992 4732 4722 4739 4761 4766 Frig 963 864 865 865 865 867 Freez 897 754 755 756 758 759 Cook 650 650 650 650 650 650 Dry 1100 1100 1101 1100 1101 = 1101 Lite 1300 1299 1300 1297 1295 1293 Misc 2000 2000 2001 1997 1995 1990 AVERAGE EQUIPMENT EUIs Heat by building type Single 20000 19997 20019 20085 20274 20466 Multi 10000 9998 10002 9967 9988 10015 Mobile 10000 9999 10009 10020 10082 10140 Other End Uses Water 5041 5025 4907 4792 4760 4761 Frig 1155 1148 1097 998 903 869 Freez 999 991 933 855 793 762 Cook 650 650 650 650 650 650 Dry 1100 1100 1101 #1100 1101 = # 1101 Lite 1299 1296 1287 1301 1321 1346 Misc 2001 2008 2042 2098 2172 2252 INDEX VALUES (1987=1) 1987 1988 1990 1995 2000 2005 2010 Housing Stock 1.000 1.007 1.027 1.087 1.156 1.252 1.363 Use per House 1.000 1.015 1.016 0.997 0.960 0.941 0.944 Total Sales - 1.000 1.023 1.045 1.082 1.109 1.177 1.286 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-12 AKREM Output Summary Region: FAIRBANKS Case: MIDDLE PART I 1987. 1988 1990 1995 2000 2005 2010 Occupied Housing Stock (000) Single 25,5 -25,6 US a7: Mu 17.1 18.8 + ee | Multi 6.7 6.6 6.8 8.6 9.4 10.3 11.6 Mobile 203 2.3 2.4 4:5 2.6 aut 269) TOTAL 24.5 24.6 24.9 26.8 29.1 S159 35:6 New Housing Units (000) Single 0.0 0.0 0.0 0.2 0.4 0.5 6.5 Multi 0.0 0.0 0.0 0.1 O42) 0%3) 07.3) Mobile 0.0 0.0 0.0 0.0 0.0 0.0 0.0 TOTAL 0.0 0.0 0.0 0.4 0.7 0.8 0.8 Adjusted Sales by End Use Heat 18 18 18 20 22 24 27 Water 40 43 45 46 44 45 50 Frig ~ 28 28 28 29 29 29 31 Freez 15 pA is 15 15 16 17 Cook id 11 12 13 14 15 a7, Dry 19 19 19 20 22 24 27. Lite 30 31 31 33 a 41 47 Misc 49 49 50 54 60 68 80 TOTAL SALES 2a: 214 218 23 242 262 296 Kwh per Customer 8,595 8,730 8,754 8,604 8,318 8,222 8,323 New Equipment Electric Shares HEAT: New Buildings © 9 ------ ------ wre ene weer ee wee eee eee eee Single 0.04 0.04 0.04 0.04 0.04 0.04 Multi 0.09 0.09 0.09 0.09 0.09 0.09 - Mobile 0.02 0.02 0.02 0.02 0.02 0.02 OTHER Water 0533/0733 053105301030) #05311 Frig 1.03 1.03 1.04 1.04 1.04 1.04 Freez 0.63 0.64 0.65 0.65 0.65 0.66 Cook 0.74 0.73 0.78 0.79 0.82 0.80 Dry 0.74 O.73 0.75 0.75 0.74 0.75 Lite 1.00 1.00 1.01 1.01 1.01 1.01 Misc 1.00 1.00 1.01 1.01 1.01 1.02 Residential (AKREM) and Commercial (COMMEND) End Use Model Output c-13 AKREM Output Summary Region: FAIRBANKS Case: MIDDLE PART II 1988 1990 1995 2000 2005 2010 Average Electric Shares of total housing market Heat Shares by building type Single 0.04 0.04 0.04 0.04 0.04 0.04 Multi 0.09 0.09 0.09 0.09 0.09 0.09 Mobile 0.02 0.02 0.02 0.02 0.02 0.02 Water 0.36 0.37 0.37 0.33 O.31 0.31 Frig 1203"— 1203123038103) -1-03 1.104. Freez 0.63 0.63 0.63 0.64 0.64 0.65 Cook 0.77. 0.78 O.80 0.78 0.77 0.77 Dry 0.74 0.74 0.72 0.72 0.72 0.73 Lite 1.00 1.00 1.00 1.00 1.01 1.01 Misc 1.00 1.00 1.00 1.00 1.01 1.01 NEW EQUIPMENT EUIs (Kwh/House/Yr) Heat by building type Single 20000 20007 20077 20072 20132 20146 Multi 9000 9003 9035 9042 9082 9101 Mobile 10000 10003 10039 10039 10074 10088 Other End Uses ; Water 4992 4733 4731 4750 4778 4791 Frig 963 864 866 866 867 869 Freez 897 754 755 757 Uey) 761 Cook 650 650 651 651 652 653 Dry 1100 1101 1105 1106 1109 1112 Lite 1300 1300 1303 1302 1304 1305 Misc 2000 2001 2007 2007 2009 2010 AVERAGE EQUIPMENT EUIs Heat by building type Single 20000 20007 20091 20215 20473 20773 Multi 10000 10003 10031 10008 10055 10120 Mobile 10000 10004 10043 10078 10172 10279 Other End Uses Water 5042 5028 4914 4802 4777 4785 Frig 1155 1149 = 1096 995 904 871 Freez 999 991 932 852 793 764 Cook 650 650 651 651 652 653 Dry 1100 1101 1105 1106 1109 = 1112 Lite 1300 1300 1292 1310 1337 #1370 Misc 2001 2009 2050 2113 2195 2289 INDEX VALUES (1987=1) 1987 1988 1990 1995 2000 2005 2010 Housing Stock 1.00 1.00 1.02 1.09 | 1.19 1.30 1.45 Use per House 1.00 1.02 1.02 1.00 0.97 0.96 0.97 Total Sales 1.00 1.02 1.04 1.10 1.15 1.25 1.40 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-14 AKREM Output Summary Region: FAIRBANKS Case: HIGH PART I 1987. 1988 1990 1995 2000 2005 2010 Occupied Housing Stock (000) Single 15.55) veo sen) Lose 7) oe) heme ua ete eos Multi 6.7 6.7 D3 S810. OT 729 Mobile 2.3 2:3 2.4 2.6 2o7 229 3o1 TOTAL 245.2406.) 25.14) | 2785!) 3019... 35.9.1 39/5 New Housing Units (000) Single 0.0 0.0 0.0 0.3 0.6 Tet 0.6 Multi 0.0 0.0 0.0 0.2 0.3 0.7 0.4 Mobile 0.0 0.0 0.0 0.0 0.1 0.1 0.1 TOTAL 0.0 0.0 0.0 0.6 1.0 139) 1.0 Sales by End Use (Gwh/yr) Heat 28207) 1832) 1858) 20n4): 23500) 27.0) 30.1 Water 99.8 43.4 45.5 (47.0) 45.9" 349.4+"*5570 Frig DT ee ee ON SO .2 miso Saco Freez 1458 7 14587 153207 15.57 (16:07 1702) 28.4 Cook De 2 yf eA be Ory este ey yes © ii) LOOn 1) owe Dry 79:1 ° 19.2, 19.6 21.0 23.4 2770. 929.8 Lite SOF OTT OC SOIT SLs iT See ites eee IIe sei owas Misc 49,2 49.3. 50.5. 54.9 63.5 . 76.4 * 8628 TOTAL SALES 210.63 215.0: 221.7 - 235.6 - 255.1 291 6° *3237.8 Usage per Customer 8595 8725 8735 8560 8270 8132 8197 New Equipment Electric Shares HEAT: New Buildings = ------ enn n ee weer ee ween e eee eee ee eee Single 0.04 0.04 0.04 0.04 0.04 0.04 Multi 0.09 0.09 0.09 0.09 0.09 0.09 Mobile 0.02. 0:.02--'-- 0.02) -O02.--0.. 02-002 OTHER Water 0.33. 0.34 0.31 0.30 0.28 0.32 Frig 1.03 1.03 1.03 1.03. 1.03- 1.03 Freez 0.63 0.64 0.64 0.64 0.64 0.64 Cook 0.74. 0:73 0.78 + 0.79: 0.82 °° 0:79 Dry 0.74 O.74 0.74 0.74 0.72 0.72 Lite 1.00 © 1.00 1.00 (1.00 1.00 - 0.99 Misc 1.00 1.00 1.00 1.00 1.00 0.98 Residential (AKREM) and Commercial (COMMEND) End Use Model Output Cc-15 AKREM Output Summary Region: FAIRBANKS Case: HIGH PART II 1988 1990 1995 2000 2005 2010 Average Electric Shares of total housing market Heat Shares by building type Single 0.04 0.04 0.04 0.04 0.04 0.04 Multi 0.09 0.09 0.09 0.09 0.09 0.09 Mobile 0.02, 10502.) 002)10,.02..0..02).10).102 Water 0..36-10.370537.-410,334 10.30) -0531 Frig L303 617036517035 .6.103 51.03) 6.1,.03 Freez 0.63 0.63 0.63 0.64 0.64 0.64 Cook ON Oe 780 1802105792 :0177- Ol 78 Dry On 74710773) 1 O72) 10.72 (10.7.2) Ol72 Lite L005 21,0031 ,00--4100--=-1:.00)=--099 Misc 120077100 1.00) 100 7 1.00) 100 NEW EQUIPMENT EUIs (Kwh/House/Yr) Heat by building type Single 20000 19987 19926 19861 19836 19739 Multi 9000 8994 8968 8953 8958 8930 Mobile 10000 9993 9963 9935 9929 9887 Other End Uses Water 4992 4728 4701 4717 4743 4733 Frig 963 864 865 865 866 868 Freez 897 754 754 756 757 758 Cook 650 650 649 649 649 648 Dry 1100 1101 1102 1101 #1101 ~=# 1099 Lite 1300 1299 1298 1295 1293 1290 Misc 2000 2001 2002 1998 1994 1987 AVERAGE EQUIPMENT EUIs Heat by building type Single 20000 19987 19964 20068 20333 20469 Multi 10000 9993 9940 9885 9901 9907 Mobile 10000 9994 9974 9993 10079 10119 Water 5041 5019 4883 4763 4733 4728 Frig 1155 1148 1093 989 898 869 Freez 999 991 929 848 788 760 Cook 650 650 649 649 649 648 Dry 1100 1101 1102 1101 1101 + # 1099 Lite 1299 1295 1290 1310 1339 1366 Misc 2001 2009 2047 2112 2195 2271 INDEX VALUES (1987=1) 1987 1988 1990 1995 2000 2005 2010 Housing Stock 1 OOF ST OL Ls04 ssl 12 or 2646 1 6 Use per House 2 OO} 1.02->--102) 1.00) 05:96) 0950.95 Total Sales OOS O2 EEE OS Let 2 ol 27 3954: Residential (AKREM) and Commercial (COMMEND) End Use Model Output C- 16 COMMEND Output Summary Region: FAIRBANKS Case: LOW FLoorstok) (Mi 1lition, t2)))) | st2) 03.3) eS) 35) ee a 2 Gore ey New Equip. Elec. Shares (%) HEAT (1) (1) 1.4 253 2p oon 2.0 WATR':/ 7G) @) a 6 5 Average Elec. Shares (%) HEAT 3k D Sh WAIR 641.7. 41, $ — wu S oO N £ Oo © 7 w w £ re New Equipment EUI (Kwh/Ft2/Yr) HEAT GL) (Gh) 6.8 Tig) 659 6.8 6.6 CoOL (1) (1) 0.5 0.6 0.5 05 O55) VENT (1) (1) ileal TAS) 1. ee a4 WATR (1) @) 5 aC a2 D2 nd COOK (1) (@) 1.0) 0.8 0.7 0.7 0.7 REFR (1) @) 25 2.0 Zach 2h Qik. Lite, -t1) (@) 5.0 Sie baal 9) | =|!) (416 MISC) (1) (a) 2D Dae Zi 2a 2.0 Average EUI (Kwh/Ft2/Yr) HEAT 10.9 Si.-D 8.4 Pr Ta) 730 6.9 COOL 0.6 0.6 0.6 0.6 O25) On) 0.5 VENT ed 1.8 1.8 Ly 1.6 Aaa Aut WATR eS Li 13 is a Le nA COOK 0.5) 0.5) 0.6 0.6 0.6 Oni 0.7 REFR 256 2.6 g.3 2.4 2.3) 22 22 LITE 6.4 6.4 6.2 538 Si) aan 4.9 MISC 220) 2.0 2.0 2.0 elt 2.2 2a Electric Intensity (Kwh/Ft2/Yr) HEAT 0.4 O73) O73) 0.3 On) 0.2 0.2 COOL 0.6 0.6 0.6 075 0.5) 0.5 0.5 VENT Las 28 1.8 16 6 5) 1.4 WATR Ou5) 2.3 0.5 0.5 0.5 0.5 0.5 COOK 0x5) 0-5) 0.6 0.6 0.6 0.7 On? REFR 2.6 2.6 nD 2.4 253) ane aut LITE 6.4 6.4 6.2 5.8 5.3 ak 4.9 MISC 2.0 2.0 eek 203 a 2.6 2.7 TODAY) L.4"18))) 127077) ek 4G |) eae DG) I) desea a iso Adjusted Electric Sales (Gwh/yr) HEAT 6 5 5 5 5 4 4 COOL 10 10 10 10 10 10 10 VENT 31 31 31 31 Syl 31 32) WATR 9 9 9 10 10 10 LL COOK 9 9 9 2 13 14 14 REFR 43 43 43 45 45 46 48 LITE 108 108 107 107 103 104 108 MISC 33 33 36 42 48 3D) 60 TOTAL ELECTRIC SALES 250 248 250 261 265 274 288 Note (1): Due to lack of new construction during these years, weighted averages across building types are not meaningful. Residential (AKREM) and Commercial (COMMEND) End Use Model Output CeaT COMMEND Output Summary Region: FAIRBANKS Case: MIDDLE 1987 1988 1990 1995 2000 2005 2010 Floorstok (Million Ft2) 13.2 132 13.3 14.2 152 16.3 18.0 New Equip. Elec. Shares (%) HEAT (1) (1) @) aly a8 1.8 2.1 WATR (1) (1) (1) 38.1 39.8 41.6 42.9 Average Elec. Shares (%) HEAT aao 3.4 3.4 ash 2.9 219 2.4 WATR 41.7 41.7 41.6 40.5 40.5 40.6 41.2 New Equipment EUI (Kwh/Ft2/Yr) HEAT (1) (1) (1) 6.9 6.9 Gi, 6.7 coOoL (1) G.) (1) 0.5 0.5 0.4 0.35 VENT (1) (1) (1) 1.4 dwt Lid 1.4 WATR (1) (1) (1) pe cam fae cook (1) (1) (1) 0.8 0.7 0.6 0.8 REFR- (1) (1) (1) z.2 Zio asa ek LITE (1) (1) (1) Ba Buk 4.8 4.6 MISC «) (Gb) @) 21 Zid 27.0. 2.0) Average EUI (Kwh/Ft2/Yr) HEAT 10.9 8.5 8.4 Uae Tul 7.0 6.9 COOL 0.6 0.6 0.6 0.5 0.5 (0) 5) 0.5 VENT 1.8 1.8 1.8 ket 4.6 1.8 1.4 WATR ia daa aS ia Lo 22 Lea COOK 0.5 OF) 0.6 0.6 0.6 0.6 0.6 REFR 2.6 2.6 2.8 2.4 Z.9 2.2 252. LITE 6.4 6.4 G.2 5.8 30 Suk 4.9 MISC 2.0 2.0 2.0 2.0 2.0 2.0 Qiks Electric Intensity (Kwh/Ft2/Yr) HEAT 0.4 0.3 0.3 0.3 0.3 0.2 0.2 COOL 0.6 0.6 0.6 0.5 0.5 0.5 je) VENT 1.8 1.8 1.8 iezs 1.6 5 OR. | 1.4 WATR G.5 (55) 0.5 0.5 0.5 0.5 0.5 COOK 0.5 0-5 0.6 0.6 0.6 0.6 0.6 REFR 2.6 2.6 ZS 2.4 2.2 22) yy 4 LITE 6.4 6.4 6.2 518 SS ae 4.9 MISC 2.0 210) 2.0) ae 2.4 2.6 iT, TOTAL «=§©14,8 #$44.7 ##44.6 146.0 13.5 13.1 12.9 Adjusted Electric Sales (Gwh/yr) HEAT 6 5 5 5 5 4 4 COOL 10 10 10 10 10 10 10 VENT 31 31 31 29 31 31 33 WATR 9 9 9 9 10 10 11 COOK 9 9 9 10 11 13 14 REFR 43 43 43 43 45 46 50 LITE 108 107 104 104 102 104 112 MISC 33 33 34 41 47 54 61 TOTAL ELECTRIC SALES 250 247 246 252 261 271 296 Note (1): Due to lack of new construction during these years, weighted averages across building types are not meaningful. Residential (AKREM) and Commercial (COMMEND) End Use Model Output C- 18 COMMEND Output Summary Region: FAIRBANKS Case: HIGH Floorstok (Million Ft2) 13.2 13.2 13.8 15.0 16.7 18.9 20.3 New Equip. Elec. Shares (%) HEAT (1) (1) Lea ee 9 . WATR (1) (1) 36.6 36.0 40.0 39.9 42.2 Average Elec. Shares (%) HEAT au WATR 41. mR co DS row np S B NS + ° ro $ °o yD £ ° £ ° nw E) 7 New Equipment EUI (Kwh/Ft2/Yr) HEAT (1) (1) 6.7 7.0 6.7 6.5 6.4 cooL (1) (1) 0.4 OBE) O55) 0.4 0.4 VENT (1) @) Lz 1.4 1.4 ‘lie3 iS WATR (1) (1) eo) 1.4 wi Ue 2.10 COOK (1) (1) 2 0.9 0.5 0.4 055) REFR (1) @) 2.6 uk mk Amal 240) LITE (1) (1) Sek 5.0 eyauk 4.7 4.5 MISC (1) (1) 22 Zak mul 20) 250 Average EUI (Kwh/Ft2/Yr) HEAT 10.9 855 Se at 7.4 659) 6.7 COOL 0.6 0.6 0.6 0.5 0.5 055 O25) VENT 1.8 es 1.8 1.6 1.6 eae 1.4 WATR ee] lie Li eS Laid ind LZ COOK Oh5) 0.5 0.6 0.6 0.6 0.6 0.6 REFR 2.6 2.6 ae) 2.4 Cr Zoe ak LITE 6.4 6.4 6.2 5.8 Das) 5.0 4.9 MISC 2.0 2730) 2.0 2.0 ek el Zk Electric Intensity (Kwh/Ft2/Yr) HEAT 0.4 0.3 O73 O53 O73 O52 0.2 COOL 0.6 0.6 0.6 0.5 OF) 055) 0.4 VENT 1.8 138 1.8 1.6 5 ad 4: WATR O75 Woe) 0.5 0.5 0.5 0.5 ORS) COOK OR) (055) 0.6 0.6 0.6 0.6 0.6 REFR 2.6 2.6 ino 2.4 Zoo) 262 2.2 LITE 6.4 6.4 6.2 56 Daa) 530) 4.8 MISC 2.0 230 aie 253) 2.4 2.6 Zio TOTAL, 14°18 _ 14-7 14-6 13.9" —1353' 1239 12.57 Adjusted Electric Sales (Gwh/yr) HEAT 6 5 . 3 S 5 5 CooL 10 10 10 10 10 ae 11 VENT ol a1 SL 31 32 34 36 WATR 9 9 9 10 10 11 13) COOK 9 9 10 a u3 14 15 REFR 43 43 45 46 48 54 56 LITE 108 107 110 110 111 120 125 MISC So. 33 36 43 52. 62 69 TOTAL ELECTRIC SALES 250 247 255 266 282 312 330 Note (1): Due to lack of new construction during these years, weighted averages across building types are not meaningful. Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-19 AKREM Output Summary Region: KENAI Case: LOW PART I 9 87 988 e990 11995177 200071 2005/7 72020 Occupied Housing Stock (000) Single oS) 9.4 9.4 9.8 10.4 11.0 i138 Multi auk ae Baa 249 Syed: SsL3 355 Mobile > 2.3 2.4 Zit 2:8 23 Bex TOTAL 14.7 14.8 14.9 15.4 16.3 Li.a ak its} New Housing Units (000) Single 0.0 0.0 0.0 Onn OFs2) On: OR2) Multi 0.0 0.0 0.0 0.0 O.F 0.1 0.1 Mobile 0.0 0.0 0.0 0.0 0.0 0.0 0.0 TOTAL 0.0 0.0 0.0 0.2 0.3 0.3 0.3 Sales by End Use (Gwh/yr) Heat 26 26 24 16 15 16 17 Water 37 39 37 32 31 33 35 Frig 167) tT 17 17 16 1S 16 Freez 10 I: 10 10 10 10 10 Cook 6 6 6 6 6 6 Ts Dry 10 10 10 10 10 9 10 Lite 18 18 19 19 20 22 23) Misc 29 29 29 31 33 36 39 TOTAL SALES 153 156 152 141 141 147 157 Usage per Customer NO 409 11051410 2U BO L458 .667 2/8 552116, 50n) New Equipment Electric Shares HEAT: New Buildings = ------- -----2 eee eee ee eee eee -e- ------ Single OSS MOET SOP ORO STOUM OO Ln Ono Multi OMS Osean Osan OeLonit Ocelot Ors, Mobile O S20 UMNO SZO MMO TS NOLS HO. SiianO no) OTHER Water 025005397) 0843) O45 1 Ocaon OL44 Frig LOS ie OStatic OS aanHasOS aul OS imnuel Os) Freez 7 4s Ore 7.41 O's O27. 4a 10.2)7-44 0 O74 Cook OS 43 Ona 7 Ono On O50 nO .45tORnGS, Dry OF SIMO... S LO sol malLO4 5. leO. 5:1 1Or5 0 Lite Me, OO el sO Hi OS Si On 99 tO 209 OK & Misc LOOMS OOUMORS STO 99 MMOs .9 SuOn 917, Residential (AKREM) and Commercial (COMMEND) End Use Model Output C= 21 AKREM Output Summary Region: KENAI Case: LOW PART II 1988 1990 1995 2000 2005 2010 Average Electric Shares of total housing market Heat Shares by building type Single 0.21 0.19 0-13 0612 0-1 Ont Multi 0.28 0.26 0.20 0.18 0.17 0.17 Mobile 0.05 0.03 0.00 0.00 0.01 0.01 Water 0.53 0.50 0.41 0.38 0.39 0.40 Frig 1.031031 0351 .03—— 1-03-1508 Freez 0.74 O.74 0.74 0.74 0.74 0.74 Cook 0.52 0.52 0.51 0.47 0.48 0.48 Dry 0.65 0.65 0.62 0.56 O.51 0.51 Lite 1.00 1.00 1.00 0.99 0.99 0.99 Misc 1.00 1.00 1.00 1.00 0.99 0.98 NEW EQUIPMENT EUIs (Kwh/House/Yr) Heat by building type Single 10032 10007 9873 9861 9848 9813 Multi 6316 6301 6220 6219 6220 6207 Mobile 8025 8005 7886 7826 7786 7742 Other End Uses Water 5021 4726 4675 4690 4712 4701 Frig 953) 860 859 860 861 864 Freez 897 754 TES) 754 756 757 Cook 801 800 797 797 796 796 Dry 1101 1101 1097 1096 1096 1094 Lite 1287 1287 1305 1302 1298 1295 Misc 1982 1982 1973 1969 1965 1959 AVERAGE EQUIPMENT EUIs Heat by building type Single 10032 10007 9886 9911 9956 10002 Multi 7519 7500 7367 7329 7306 7282 Mobile 8025 8006 8099 8406 8554 8674 Other End Uses Water 5067. 5042 4875 4751 4701 4694 Frig 1148 1139 1077 981 892 861 Freez 999 991 940 861 790 TOT Cook 801 800 797 797 796 796 Dry 1101 1101 1097 1096 1096 1094 Lite 1288 1289 1296 1309 1326 1346 Misc 1984 1990 2019 2073 2139 2212 INDEX VALUES (1987=1) 1987 1988 1990 1995 2000 2005 2010 Housing Stock 1.000 1.008 1.012 1.048 1.105 1.171 1.244 Use per House 1.000 1.011 0.983 0.879 0.833 0.820 0.823 Total Sales 1.000 1.019 0.994 0.924 0.924 0.962 1.025 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C=22 AKREM Output Summary Region: KENAI Case: MIDDLE PART I 1987-1988 1990 1995 2000 2005 2010 Occupied Housing Stock (000) Single 9.3 9.4 9.4 9.9 10.8 13.7 12.9 Multi Sok Be 3.0 3.2 aoe Ped 4.1 Mobile 235 a 2.4 2.7 2.8 3.0 aoa TOTAL 14.7 14.8 14.9 13.7 We) 14 F2 New Housing Units (000) Single 0.0 0.0 0.0 0.2 0.2 01.3) 0.3 Multi 0.0 0.0 0.0 0.1 0.1 0.1 Ck Mobile 0.0 0.0 0.0 0.0 0.1 0.1 0.1 TOTAL 0.0 0.0 0.0 0.3 0.4 0.4 0.4 Adjusted Sales by End Use Heat 26 26 24 LZ 16 a7 19 Water 37 39 37 32 32 34 38 Frig’ 17 17 17 7 17 16 17 Freez 10 11 10 11 10 10 11 Cook 6 6 6 6 6 7 8 Dry 10 10 10 10 10 10 12 Lite 18 18 as 20 22 24 27 Misc 29 29 29 2 35 39 45 TOTAL SALES 33 156 152 144 148 159 177 Kwh per Customer 10,409 10,521 10,232 9,162 8,707 8,643 8,745 New Equipment Electric Shares HEAT: New Buildings = ------ rere cree ne were ee cere ee rere ee Single OS) O25) ))) 0-110) 3) (O20) 0.2105 07110 Multi OES | )))}02 15) | i) Od5))))) (0015) |) 0x25 ie O.en5 Mobile 0:20) 105:20)) O15) OTS) 0215 17) Ors OTHER Water ORSON Oe SO Ona Oraa Ones e044 Frig AOS OS) Os) LOS) yy sO31))) PL Od Freez ORE ORG) | OL75)11)) O79) (110). (75) ri O.a76) Cook 0.43) ||) (0547 | 10.50.°) 0550 | 0.46 || 0249 Dry OF:53))) (01253 110253) 11210 194) 11) 0545/10. .55 Lite 200 |||) 00) ||) 100))|)) E500) |) //1500) 7 02 Misc LOO LOO LOL Le Oly ils) | h01! Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-23 AKREM Output Summary Region: KENAT Case: MIDDLE PART II 1988 1990 1995 2000 2005 2010 Average Electric Shares of total housing market Heat Shares by building type Single C.22. O.49 O13 G.1k G.Ah. 6G.dd Multi O28) |) 0.26; |) (Ol L9) |) Ody.) Oz) 10117: Mobile 0.05 0.03 0.00 0.00 0.01 0.02 Water 0.53 0.50 0.41 0.38 0.38° 0.39 Frig i002 4S 61S 61.09: 1.08 7.03 Freez 0.74 O.74 O.74 O.74 O.74 0.75 Cook 0.52 0.52 0.51 0.48 0.49 0.49 Dry 0.65 0.65. 0.62 G.36 0.53 06.53 Lite 200) |) 15500) |) E00) 5) 200) |) 1500) | | 101 Misc DAO0) |) 00 |) 2.00) |) N00)) |) 1.0.) | | aOL NEW EQUIPMENT EUIs (Kwh/House/Yr) Heat by building type Single 10032 10012 9908 9912 9926 9917 Multi 6316 6304 6242 6254 6274 6280 Mobile 8025 8010 7912 7866 7853 7833 Other End Uses Water 5021 4727 4685 4702 4732 4730 Frig 953 860 860 861 862 865 Freez 897 754 754 755 757 759 Cook 801 800 798 798 799 799 Dry 1101 1102 1101 £1102 += 1105 1107 Lite 1287. 1287) =1289 =: 1292 1295 1297 Misc 1982 1983 1980 1980 1981 1981 AVERAGE EQUIPMENT EUIs Heat by building type Single 10032 10012 9925 9982 10073 10180 Multi 7519 7504. 7392 7357 7351 7347 Mobile 8025 8010 8167 8455 8616 8764 Other End Uses Water 5067 5044 4880 4762 4721 4723 Frig 1148 1139 1076 979 894 866 Freez 999 992 939 859 791 761 Cook 801 800 798 798 799 799 Dry ATOR |) 2102) | TLOL) |) 1102) | | aos || |1:107 Lite 1288 1290 1296 1316 1341 1371 Misc 1984 1991 2027 2090 2166 2254 INDEX VALUES (1987=1) 1987 1988 1990 1995 2000 2005 2010 Housing Stock 100 |) TeO1 ||) L501) || 107) |i Le: | | 325) | | 1.38 Use per House 1.00 1.01 0.98 0.88 0.84 0.83 0.84 Total Sales L300 | | d02 || 10599) | 0194) (0.97) | 1 [04) | | 16 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-24 AKREM Output Summary Region: KENAI Case: HIGH PART I 1987 1988 1990 1995 2000 2005 2010 Occupied Housing Stock (000) Single 953 9.4 954100 22135 Multi Sek 5.2 Sey Bae S55 326 4.2 Mobile ZS 250) 2.4 2a), 29) S22 3.4 TOTAL 14,7 14.8) 15.00 1538) 7.5 92 20 2 New Housing Units (000) Single 0.0 0.0 0.0 0.2 0.3 0.2 053) Multi 0.0 0.0 0.0 0.1 Onn Onn 0.1 Mobile 0.0 0.0 0.0 0.0 0.1 0.0 0.1 TOTAL 0.0 0.0 0.0 0.3 0.5 0.2 os Sales by End Use (Gwh/yr) Heat 26.1 26.0 23.8 17.0 16.7 18.0 19.8 Water 36.7 38.7 O7.0 S216 (3304 36.9 4157 Frig 1618 1'6),9\ 7,0 16-9. 17502 1750. 283: Freez 10.4 10.5 10.5 10.5 10.7 10.8 11.5 Cook 5.6 D9 6.0 G52 6.5 Vad. 8.0 Dry 10.2 10.3 10.3 10.4 10.4 10.6 11.9 Lite 18.3 18.5 18.6 19.8 22.2 24.6 27.6 Misc 28.9 29.0" 29.4 31.5 3538 40.5 - 46.1 TOTAL SALES 153.0 155.8 153.1 145.0 152.8 165.7 184.9 Usage per Customer 10409 10521 10207 9165 8742 8646 8717 New Equipment Electric Shares HEAT: New Buildings = --we ne ween ne weer ee cee ee eee eee ee eeee Single 0.15 0.15 0.10 0.10 0.10 0.10 Multi O15) —-0715)—_0.15 0215 0-15 _ 0.15 Mobile 0.20 0.20 0.15 O.15 O.15 0.15 OTHER Water 0.50 0.40 0.44 0.45 0.48 0.46 Frig 1.03/21, 08= 1.03) 108 1-03-1039) Freez 0.74 0.74 0.74 O.75 0.74 0.74 Cook 0.43 0.46 0.50 0.51 0.46 0.50 Dry 0553)" 0553 0.53, 2 0.53 10.53. * 70.52 Lite 1.00 1.00 1.00 1.00 0.99 0.99 Misc 1.00 1.00 1.00 1.00 0.99 0.98 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C=25 AKREM Output Summary Region: KENAI Case: HIGH PART II 1988 1990 1995 2000 2005 2010 Average Electric Shares of total housing market Heat Shares by building type Single 0.22. O19 ©.13 Odt . O12 O.T1 Multi 0.28 0.26 0.20 0.18 0.17 0.17 Mobile 0.05 0.03 0.00 0.01 0.02 0.03 Water 0.53 0.50 0.41 0.39 0.40 0.41 Frig 1.03 12.03 £03 ° 1.03 2.03 1:03 Freez 0.74 O.74 0.74 0.74 0.74 0.74 Cook 0.52 0.52 0.51 0.48 0.49 0.49 Dry 0.65 0.65 0.62 0.56 0.53 0.53 Lite 1.00 1.00 1.00 1.00 1.00 0.99 Misc 1.00 1.00 1.00 1.00 1.00 0.99 NEW EQUIPMENT EUIs (Kwh/House/Yr) Heat by building type Single 10032 10007 9862 9860 9830 9770 Multi 6316 6301 6214 6223 6218 6193 Mobile 8025 8006 7875 7824 7780 7722 Other End Uses Water 5021 4729 4667 4685 4702 4684 Frig 953 862 859 860 861 864 Freez 897 754 753 755 756 757 Cook 801 800 797 797 796 795 Dry 1101 1102 1099 1099 1098 1096 Lite 1287) 1288 1292 1293 1291 1287 Misc 1982 1983 1976 1975 1969 1961 AVERAGE EQUIPMENT EUIs Heat by building type Single 10032 10007 9880 9941 9995 10055 Multi 7519 = 7501 7352 7305 7267 7226 Mobile 8025 8006 8115 8413 8515 8621 Water 5067 5040 4861 4742 4692 4679 Frig 1148 1138 1074 976 891 864 Freez 999 992 938 856 788 759 Cook 801 800 797 797 796 795 Dry 1101 1102 1099 1099 1098 1096 Lite 1288 1288 1296 1316 1338 1363 Misc 1984 1991 2024 2087 2157 2235 INDEX VALUES (1987=1) 1987 1988 1990 1995 2000 2005 2010 Housing Stock 1.00 1.01 1.02 1.08 1.19 1.30 1.44 Use per House 1.00 1.01 0.98 0.88 0.84 0.83 0.84 Total Sales 1.00 1.02 1.01 0.95 1.00 1.09 1.21 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-26 COMMEND Output Summary Region: KENAT Case: LOW 1987 1988 1990 1995 2000 2005 2010 Floorstok (Million Ft2) 7.9 8.3 8.2 8.7 9.1 e.F 10.3 New Equip. Elec. Shares (%) HEAT (1) (1) 26.1 9.6 ; ae WATR (1) @) 14.3: 04.6 59.5 60.7 58.9 Average Elec. Shares (%) HEAT 12.8 12.6 ie] ih. SATE = 61.4% = =62.5 62. “Sl, un rR oO Oo ray oO wo New Equipment EUI (Kwh/Ft2/Yr) HEAT (1) (1) 7.8 720) (a7 655) 6.3 COOL (1) (1) Sra! 0.9 0.7 0.6 0.6 VENT (1) (1) Ze 1.8 1.6 1.6 15) WATR (1) (1) 3.9 1.8 ao nD) 1.4 COOK (1) (1) (sail kon 0.9 0.9 0.8 REFR (1) (1) 4.0 Zak 1.9 1.8 del LITE (1) (1) me) Sf 55) 5,2 5.0 MISC § (1) (1) 3.2 aun 1.9 hae a6 Average EUI (Kwh/Ft2/Yr) HEAT 8.0 8.1 8.0 7.6 7.8 6.7 6.5 COOL 0.5 0.5 0.5 0.5 OFS) 0.5 075) VENT 1.9 1.9 1.9 1.8 xan 1.6 WATR 5 Wt 5 Lo 1.3 1.4 COOK (ORS) O55 0-5) 0.6 OF (57) 0.7 REFR eh 2.0 2-0 2.0 9) 18 abay/ LITE 6.5 6.5 6.4 6.0 a. oa §.1 MISC Mee ss, le. 1 1.8 1.8 1.8 Electric Intensity (Kwh/Ft2/Yr) HEAT 110) 1.0 1.0) 0.9 0.8 0.7 0.6 COOL O25) 0.5) 0.5 (AE) 0:5) O75) 07,5) VENT 139 1.9 19 La? 1.6 lao p WATR 0.9 1.0 0.9 0.9 0.9 0.9 0.9 COOK 035) 0.5 0.5) 0.6 On 057 0-7 REFR ain 2.0 2.0 2.0 1.9 1.8 1.8 LITE 6.5 625 6.4 6.0 5.4 by574 Sel MISC Lit pe 1.8 2j.0) 22 2.4 265) aan: 281. 13.8 150 6 hl a Adjusted Electric Sales (Gwh/yr) HEAT 8 8 8 7 7 7 7 COOL 4 4 4 4 5 5 5 VENT 15 16 15 135 15 15 15 WATR 7 8 8 8 9 9 9 COOK 4 4 4 5 6 7 7 REFR 16 17 17 17 17 18 18 LITE 51 53 51 a 49 49 51. MISC 13 14 pE. 17 20 23 24 TOTAL ELECTRIC SALES D7, 122 121 E22 126 Sl 135 Note (1): Due to lack of new construction during these years, weighted averages across building types are not meaningful. Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-27 COMMEND Output Summary Region: KENAI Case: MIDDLE Floorstok (Million Ft2) 7.8 8.3 8.1 8.6 9.4 o.oo 25.2 New Equip. Elec. Shares (%) HEAT (1) (1) (1) 7.8 7.8 8.0 a9 WATR (1) @) @) 55.0 5529) 56.9 Average Elec. Shares (%) HEAT i276 26 liars) LL 6: 1 106) 9.7 9.3 WATR || GbaGs) 6225) || 162/50) | 61712) | | (60/3 | | 59.8 59.2 New Equipment EUI (Kwh/Ft2/Yr) HEAT (1) (1) (1) 6.8 6.6 6.4 6.3 cooL (1) (1) @) 0.6 o.3 0.5 0:5) VENT (1) (@)) @) 1.6 1.6 15 Lae WATR (1) (1) @) pe 1.4 dea La COOK (1) (1) (1) 0.8 0.7 Oj?) 0.7 REFR (1) (1) (1) 1.9 1.8 1.7 136 LITE (1) i) (1) 5.5 DS we Bae 4.9 MISC (1) (1) @) a) 138 ee) oa, Average EUI (Kwh/Ft2/Yr) HEAT 8.0 8.1 8.0 Tat Ive On 7 625 COOL 0.35 0.5 Orm5) O15) 0:/5) 0.5 0.4 VENT 1.9 9 1.9 1.8 hal 1.6 Te WATR 1.5 15 bout p sat 1.4 1.4 COOK 0.5 0.5 0.5 0.6 0.6 0.7 0.7 REFR ae 2.0 2.0 2.0 Lo) 158 aoF LITE 6.5 6.5 6.4 6.0 S19 5.2 ak MISC ue ei eed 1.8 1.8 1.8 1.8 Electric Intensity (Kwh/Ft2/Yr) HEAT 1.0 1.0 1.0 0.9 0.8 On 0.6 COOL (0%) 0).5 0.5 6.5 0.5 0.4 0.4 VENT 69) 1.9 Li az: 1.6 1.5 oe WATR 0.9 70) 0.9 0.9 0.9 0:19) 0.8 COOK On) ORS O75 0.6 0.6 Oo7) OLi7: REFR Suu 2.0) 2.0 230 19 1.8 Let LITE 6.5 6.5 6.4 6.0 5.4 De) a MISC Le? ee 1.8 240) Cars 10) 2.4 SOTA, «615.2 03.0 25.0 14.5 35.6 UA 13.1 Adjusted Electric Sales (Gwh/yr) HEAT 8 8 8 7 7 6 7 COOL 4 4 4 4 4 4 5 VENT 15 16 is 5) a5) 15 16 WATR 7 8 8 8 8 9 3 COOK 4 4 4 a 6 7 7 REFR 16 7, 16 17 18 18 13 LITE 51 53 51 51 50 51 55 MISC 13 14 14 17 20 23 25 TOTAL ELECTRIC SALES 117 122 aay) 122 126 ache 142 Note (1): Due to lack of new construction during these years, weighted averages across building types are not meaningful. Residential (AKREM) and Commercial (COMMEND) End Use Model Output C- 28 COMMEND Output Summary Region: Floorstok (Million Ft2) Wad (yap New Equip. Elec. Shares (%) HEAT (1) (1) WATR (1) a) Average Elec. Shares (%) HEAT 12.8 12.5 WATR 61.4 62.7 New Equipment EUI (Kwh/Ft2/Yr) HEAT (1) (1) COOL (1) (1) VENT (1) (1) WATR (@) (1) COOK (1) (1) REFR- (1) (1) LITE (1) (1) MISC (1) (1) Average EUI (Kwh/Ft2/Yr) HEAT CooL VENT WATR COOK REFR LITE MISC Electric Intensity (Kwh/Ft HEAT COOL VENT WATR COOK REFR LITE MISC TOTAL PANOFRFOE NNDOUUOUPF r) UFRADNOOCOrFPORPNEFANOFPF Of je arena Uncool en rama Ciao eno co UrRPADANHNOrRrF OF NNDOUCDWOWNO re er Adjusted Electric Sales (Gwh/yr) HEAT 8 8 COOL 4 4 VENT 15 15 WATR 7 8 COOK 4 4 REFR 16 16 LITE 51 52 MISC 13 14 TOTAL ELECTRIC SALES 117 120 ar PP ow HVPE rrFr ON RrPAaANnNOrRF COO NUP UU OWS UrRFANOOrROF rPouUrFUDOAUOo rR 8 4 15 8 4 17 52 15 122 oOoUuUrFrUOewWO rR Be ofr PFPuUuUrRPOFRF OC OwWwwouUnnawo rPaANOrRrF ON COOKnDUNUAWNDA FnNoNuUuUnOCOrFOCSO FOoWONnNUON Ue re i 4 15 8 >) 18 52 18 125 ro o uw PUPORROD NFNDWUUY rPURPORPRON CUWOAUAUN WNUrFOOrRGSO NRPFODWOAUN ray 7 4 16 9 6 19 52 21 132 uw wo wa PUPORRFOD CNN OLUD SEL rPUuUrROrRPrROD CNONFUUD WNUrFOOCOrFOCO NOwWRrPONoOULD e 7 5 16 9 7 Lg 54 - 24 139 w eo © RP PrOrRPr OO NODDWHEUF FPUFPOPROD COrnnN FUL LS NNHOUrRPOOrROCSO CWONNOHKED ry 7 5 Ly 9 8 20 Ey) 26 148 Note (1): Due to lack of new construction during these years, weighted averages across building types are not meaningful. Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-29 AKREM Output Summary Region: MATSU Case: LOW PART I O87) 9.8.8) ))))) 19.9.0) 995) 2000 101) 2005) ||/s2010 Occupied Housing Stock (000) Single war Pee) AOue |) Raee | Dhow Mah ~ Tat Multi aa Lie 0.7 0.3 1.0 Lew =o Mobile lad 1.2 hal 1.4 1.4 1.4 | TOTAL Beek |||) eee ||| date 267 P35) 14.7 16.1 New Housing Units (000) Single 0.0 0.0 0.0 0.0 0.0 0.2 0.3 Multi 0.0 0.0 0.0 0.0 0.0 0.0 G.1 Mobile 0.0 0.0 0.0 0.0 0.0 0.0 0.0 TOTAL 0.0 0.0 0.0 0.0 0.0 0.3 0.4 Sales by End Use (Gwh/yr) Heat ae 68 58 38 28 as 29 Water 34 32 28 21 Be 25 30 Frig 16 16 16 16 16 a3 16 Freez ty 11 11 PL 11 11 11 Cook a U 7 a 8 9 10 Dry 11 a 11 LL Ty 10 i Lite 19 19 19 20 21 23) 25 Misc 35 | bb 36 38 42 48 TOTAL SALES 207 198 184 161 aos 160 180 Usage per Customer 175,052/16,,586 15),414 12.4716) 115368 10), 92011 21/50 New Equipment Electric Shares HEAT: New Buildings = --- 2+ weer ee ween en eee eee eee eee - eee Single OLS) Ole 5) seein il Ol 3) | L:3 ane O alts) Multi OS 20 OSZON OILS OES NOLS gOS Mobile O10) (0.20) 1) 07X08) 0)08) 0110877008 OTHER , Water 04211) 0 36.1)):0. 40) Ose?) Oat O40 Frig 108) 1) 03.)) 1 OS | OSi iOS er Ms, 03 Freez OSH O84) 065) 0.183) One sea One8 Cook 01560) (0.68 1)))'0..65' 11) 07651) OnG2) 8 Oe fal: Dry OF 8) O59) Os 61111 O59) iO 99) MIMO ou Lite BSOO) |) 24200) (099) )10:2,997 11/099) 47 202.98 Misc L300) @ 2100) 0799) 705519. 8311)// 01.98." (05.96 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C31 AKREM Output Summary Region: MATSU Case: LOW PART II 1988 1990 1995 2000 2005 2010 Average Electric Shares of total housing market Heat Shares by building type Single 0.23 0.20 0.13 0.09 0.08 0.09 Multi 0.45 0.38 0.05 .0.00 0.01 0.02 Mobile 0.19 0.16 0.09 0.05 0.04 0.04 Water 0.46 0.39 0.28 0.26 0.28 0.31 Erig LOSI 03 TOS 103) 03) 103 Freez 0.83 0.83 0.83 0.83 0.83 0.83 Cook 0267) || O65 || (063 |) | 0.162) || 0.67) | |'0467 Dry 0273) || 0075 || 0.74:)) | 0.464) | 056) | |:0255 Lite 1.00 1.00 1.00 0.99 0.99 0.99 Misc 1-00) || 1500 || 200 | 0.99) (0.99) ‘0298 NEW EQUIPMENT EUIs (Kwh/House/Yr) Heat by building type Single 22930 22804 22538 22531 22509 22383 Multi 13399 13326 13173 13196 13286 13242 Mobile 11963 11898 11759 11755 11742 11666 Other End Uses Water 5013. 4684 4719 4742 4695 4688 Frig 971 875 874 879 881 875 Freez 896 753 752 754 755) 756 Cook 799 798 795 795 795 794 Dry 1099 1097. 1094 1094 1093 1092 Lite 1379 1379 1376 1376 1369 1361 Misc 2398 2394 + 2386 2383 2379 2371 AVERAGE EQUIPMENT EUIs Heat by building type Single 22930 22804 22539 22533 22657 22970 Multi 15950 15860 15632 15359 14509 14939 Mobile 11963 11898 11760 11760 11853 12083 Other End Uses Water 5092 5065 4893 4755 4688 4658 Frig ED70) |) 165) | | 1112) |) 1015 916 878 Freez 997 987 936 863 797 755 Cook 799 798 795 795 795 794 Dry 1099 1097 1094 1094 1093 1092 Lite 1382 1399 1416 1400 1387 1407 Misc 2399 2403 2432 2482 2552 2650 INDEX VALUES (1987=1) 1987 1988 1990 1995 2000 2005 2010 Housing Stock 1.000 0.985 0.983 1.043 1.110 1.209 1.329 Use per House 1.000 0.973 0.904 0.746 0.666 0.641 0.654 Total Sales 1.000 0.962 0.891 0.777 0.739 0.777 0.870 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C- 32 AKREM Output Summary Region: MATSU Case: MIDDLE PART I 1987. 1988 1990 1995 2000 2005 2010 Occupied Housing Stock (000) Single 9.7 96 ~ 100 D0 Ww ed 12.8 14.9 Multi 1.2 kik 0.6 0.6 2.0 2.4 2.4 Mobile 1.2 1.2 | 1.4 1.4 L316 1.8 TOTAL 12.1 12.0 18.9 13.0 14.5 16.4 19.1 New Housing Units (000) Single 0.0 0.0 0.0 0.0 0.0 0.4 0.5 Multi 0.0 _0.0 030) 0.0 0.0 0.1 015.0; Mobile 0.0 0.0 0.0 0.0 0.0 0.1 0.1 TOTAL 0.0 0.0 0.0 0.0 0.0 0.5 0.6 Adjusted Sales by End Use Heat 73 68 58 39 30 33 41 Water 34 32 28 22 22 29 35 Frig 16 16 16 17 17 V7, 20 Freez 1, ot a il 12 12 14 Cook aw 7 7 7 8 10 12 Dry aa) Vet a 12 11 12 14 Lite 19 19 19 21 22 26 31 Misc 35 35 35 37 41 49 59 TOTAL SALES 207 198 184 165 163 187 226 Kwh per Customer 17,052 16,581 15,430 12,687 11,281 11,425 11,796 New Equipment Electric Shares HEAT: New Buildings = ------ ------ -reee- eee eee ee eee -2--e- Single 0-15 0.15 0.13 (0.13 0.13) ..0.13 Multi 0.20 0.20 0.18 O.18 0.18 0.18 Mobile 0.10 0.10 0.08 0.08 0.08 0.08 OTHER Water 0.42 0.36 0.41 0.42 0.40 0.39 Frig 2.03 21.03 1.03: 1.03 1:03 1.03 Freez 0.83 0.84 0.84 0.84 0.84 0.84 Cook 0.60 0.69 0.66 0.65 0.65 0.74 Dry 0.58 0.59 0.57 0.61 0.60 0.54 Lite 1.00 1.00 1.00 1.00 1.00 1.00 Misc 1.00 1.00 1.00 1.00 1.01 1.01 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C - 33 AKREM Output Summary Region: MATSU Case: MIDDLE PART II 1988 1990 1995 2000 2005 2010 Average Electric Shares of total housing market Heat Shares by building type Single 0.23 0.20 0.13 0.09 0.09 0.09 Multi 0.45 0.38 0.12 0.07 0.08 0.09 Mobile 0.19 O.16 O.09 O.05 0.04 0.05 Water 0.46 0.39 0.27 0.24 0.29 0.31 Frig 1.03 1.03 1.03 1.03 1.03 1.03 Freez 0.83 O.83 0.83 0.83 0.84 0.84 Cook 0.67 0.65 0.64 0.65 0.68 0.68 Dry 0573; 0.75: OL73 0.61, 0.573, 10..58 Lite 1.00 1.00 1.00 1.00 1.00 1.00 Misc 1.00 1.00 1.00 1.00 1.00 1.00 NEW EQUIPMENT EUIs (Kwh/House/Yr) Heat by building type Single 22930 22815 22604 22642 22639 22564 Multi 13399 13332 13211 13236 13291 13312 Mobile 11963 11904 11793 11813 11806 11758 Other End Uses Water 5013. 4685 4721 4733 4700 = 4721 Frig 971 875 875 880 880 875 Freez 896 753 753 754 756 758 Cook 799 798 796 796 797 797 Dry 1099 1098 1097 #1099 1101 #1103 Lite 1379 =1380 1379 1380 1373 1371 Misc 2398 2395 2393 2394 2395 2394 AVERAGE EQUIPMENT EUIs Heat by building type Single 22930 22815 22605 22644 23089 23571 Multi 15950 15867 15701 15712 15438 15383 Mobile 11963 11904 11795 11818 12100 12410 Other End Uses Water 5092. 5067 4899 4761 4696 4682 Frig 1170 «#1165 1108 1005 916 881 Freez 997 987 935 859 797 758 Cook 799 798 796 796 797 797 Dry 1099 1098 1097 1099 1101 # 1103 Lite 1382. 1401 1409 1379 1412 1447 Misc 2399 2404 2439 2491 2602 2723 INDEX VALUES (1987=1) 1987 1988 1990 1995 2000 2005 2010 Housing Stock 13000598 0398) ha07 | 119) 13557, Use per House 1.00 0.97 0.90 0.74 0.66 0.67 0.69 Total Sales 1.00 0.96 0.89 0.80 0.79 0.91 1.09 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C - 34 AKREM Output Summary Region: MATSU Case: HIGH PART I 1987. 1988 1990 1995 2000 2005 2010 Occupied Housing Stock (000) Single OF, 9.6 10.0 11.0 12.0 3)59, 15.9 Multi Taz TL 0.9 a 2.0 253 2.6 Mobile 1.2 1.2 73 1.4 1.5 1.7 1.9 TOTAL 2 oL. 12.0 12 13.4 23.5 17.8 20.5 New Housing Units (000) Single 0.0 0.0 0.0 0.0 On 0.3 0.5 Multi 0.0 0.0 0.0 0.0 0.1 0.1 (0 Mobile 0.0 0.0 0.0 0.0 On 0.0 Ont TOTAL 0.0 0.0 0.0 0.0 0.6 0.4 0.7 Sales by End Use (Gwh/yr) Heat 73.4 68.0 59.6 42.1 38e1l--- 41.6 4922 Water 33.7 Ole 28.2 2257. 25:29 52.3 38.4 Frig 16.4 16.2 16.3 17.2 18.0 18.7 20.8 Freez Ale S else aie) 11.8 12.5 13.4 14.6 Cook Fad Tse 742 daw 9.0 10.8 12.5 Dry 11.0 10.8 11.1 11.7 aS 1235) 14-3 Lite 18.7 18.5 19.0 20.9 24.0 28.0 32.6 Misc BS. 34.7 35.0 STATE 43.9 52.2 61.9 TOTAL SALES 207.0 198.1 187.6 171.8 183.1 209.5 244.1 Usage per Customer 17052 16581 15452 12799 11836 11769 11923 New Equipment Electric Shares HEAT): = New Bul Ging 5 snes sees = aio wwii im = mine omnia ai oe inin wninn Single O55 0.150.137 0.13 0.13" (0.13 Multi 0.20 0.20 0.18 O.18 0.18 0.18 Mobile 0.10 0.10 0.08 0.08 0.08 0.08 OTHER Water 0.42 0.36 0.42 0.38 0.42 0.41 Frig 1.03. __1.03__1,03:_1.03 1.03. _ 1°03 Freez 0.83 0.84 0.84 0.84 0.84 0.83 Cook 0.60 0.69 0.66 0.67 0.66 0.73 Dry 0.58 0.60 0.57 0.58 0.58 0.52 Lite 1.00 1.00 0.99 0.99 0.99 0.98 Misc 1.00 1.00 1.00 0.99 0.98 0.97 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-35 AKREM Output Summary Region: MATSU Case: HIGH PART II 1988 1990 1995 2000 2005 2010 Average Electric Shares of total housing market Heat Shares by building type Single 0.23 0.20 0.13 0.10 0.09 0.09 Multi 0.45 0.40 0.26 0.21 0.20 0.19 Mobile 0.19 O.16 0.09 0.05 0.05 0.05 Water 0.46 0.39 O.28 O.28 O.31 0.32 Frig 1.03 1.03 1.03 1.03 1.03 1.03 Freez 0.83 O.83 0.83 0.83 0.83 0.83 Cook 0.67 0.66 0.65 0.65 0.68 0.68 Dry 0.75 0.76 O72, 0362 0.557 __ 0.157 Lite 1.00 1.00 1.00 1.00 0.99 0.98 Misc 1.00 1.00 1.00 1.00 0.99 0.98 NEW EQUIPMENT EUIs (Kwh/House/Yr) Heat by building type Single 22930 22791 22477 22473 22360 22190 Multi 13399 13318 13136 13141 13128 13075 Mobile 11963 11891 11727 11724 11660 11564 Other End Uses Water 5013. 4681 4689 4696 4678 4682 Frig O71 875 874 876 877 874 Freez 896 753 IE? 753 Ta5 735) Cook 799 798 794 794 794 792 Dry 1099 1098 1095 1096 1094 1092 Lite 1379 1380 1375 1369 1362 1356 Misc 2398 2395 2387 2387 2379 2369 AVERAGE EQUIPMENT EUIs Heat by building type Single 22930 22791 22478 22643 22925 23225 Multi 15950 15851 15627 15535 15401 15298 Mobile 11963 11891 11729 11823 12009 12211 Water 5092 5054 4858 4711 4656 4640 Frig 1170 «1163 1107 1004 914 880 Freez 997 987 937 859 798 760 Cook 799 798 794 794 794 792 Dry 1099 1098 1095 1096 1094 1092 Lite 1382, 1393 1393 1388 1411 1439 Misc 2399 2405 2432 2504 2598 2699 INDEX VALUES (1987=1) 1987 1988 1990 1995 2000 2005 2010 Housing Stock 1.00 0.98 1.00 1.11 1.27 1.47 1.69 Use per House 1.00 0.97 0.91 0.75 0.69 0.69 0.70 Total Sales 1.00 0.96 0.91 0.83 0.89 1.02 1.18 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C=36) COMMEND Output Summary Region: MATSU Case: LOW Floorstok (Million Ft2) 4.9 4.9 4.9 5.2 535 5.9 6.4 New Equip. Elec. Shares (%) HEAT (1) (1) -6.4 6.4 4.9 6.6 5.7 WATR (1) (1) 20 Oc OUn PG. vO lo.) Average Elec. Shares (%) HEAT 639 G5) 65 6.3 6572 6.0 6.0 WATRe SBS 48/13) eS 2) 71.9) Oir ely ode LORD New Equipment EUI (Kwh/Ft2/Yr) . HEAT (1) @) 5.8 6.7 6.2 6.3) 6.1 COOL (1) (1) (}5ik 0.4 0.4 0.4 0.4 VENT (1) Gb) eel 1.6 14 tas) 1.4 WATR (1) (1) 1507) eles lea) ino 1.1 cook (1) (1) 0.1 0.3 0.4 0.4 0.4 REFR (1) (1) 1.4 10 u beac! Das) is LITE, (1) (1) 4.8 B44 50 Sid: 4.8 MISC (1) (a) 1.6 1.6 4:3 15) 16 Average EUI (Kwh/Ft2/Yr) HEAT 9.4 8.2 Siok Wad, Tine 6.4 6.4 CooL (AS) 045 0.5 On) Ou5 0.4 0.4 VENT 1.9 a9 a9 1.8 16 1.4 14 WATR 1.2 3 3 8 13) bi! v2 COOK 073 O53: 073) 0.4 0.4 0.4 0.4 REFR Ses er ey, thee) 1.6 1.4 bar LITE 6.7 6.7 657 6.2 Si5) Seo! Bue MISC 5 16 Li6 1.6 ey abs) a Electric Intensity (Kwh/Ft2/Yr) HEAT 0.6 0.6 0.6 OR5) 0.5 0.4 0.4 COOL 0.5) 035 On) OS) 0.5 0.4 0.4 VENT 19 9) 1.9 Nd, 16 Ina aya: WATR 0.2 0.2 0.2 0.2 0.2 On, 0.2 COOK 0-3) 0.3 o.3 0.4 0.4 0.4 0.4 REFR Ted, aly) ey a liAy Louk 1.4 1.4 LITE 6.7 6.7 6.6 6.1 5.4 wae aoe MISC 1.5 T36 1516 Leo 2130) Ziad 2.12) TOTALS 47 lor kona) 1) beso) ete) | dO) LL ag Adjusted Electric Sales (Gwh/yr) HEAT 3 3 3 3 3 3 3 COOL 2 2 2 2 3 2 3 VENT 10 10 10 10 10 9 10 WATR - 1 1 1 af a ut COOK 2 2 2 2 2 2 3 REFR 9 9 9 10 10 9 10 LITE 38 38 37 37 35 36 38 MISC 9 9 9 10 13 14 16 TOTAL ELECTRIC SALES 76 76 75 77 79 77 87 Note (1): Due to lack of new construction during these years, weighted averages across building types are not meaningful. Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-37 COMMEND Output Summary Region: Floorstok (Million Ft2) 4.9 4.9 New Equip. Elec. Shares (%) HEAT (1) (1) WATR (1) (1) Average Elec. Shares (%) HEAT oo WATR: | 1853 New Equipment EUI (Kwh/Ft2/Yr) HEAT @) (1) COOL (1) (1) VENT (1) (sy WATR (1) Gy COOK (1) (1) REFR (1) i) LITE (1) (1) MISC (1) G1) Average EUI (Kwh/Ft2/Yr) HEAT COOL VENT WATR COOK REFR LITE MISC Electric Intensity (Kwh/Ft HEAT COOL VENT WATR COOK REFR LITE MISC TOTAL rPaArPOrFRrF OC DANN WWOUPD r) WRARrFPOCOORP CONF DOF OFF OW PUN Se ilo Uvowe ued <1 wes wll WRArOOrFOCO FUNINWNHOUD r hr Adjusted Electric Sales (Gwh/yr) HEAT 3 COOL 2 VENT 10 WATR 1 COOK 2 REFR 9 LITE 38 MISC 9 TOTAL ELECTRIC SALES 76 ray Ww HDowmonronw ~ Note (1): Due to lack of new construction during these years, weighted averages across building types (1) (1) (1) (1) qQ) QQ) (1) (1) KF Ar Orr Oo Co WrRArOCOOrOCSO WAANWNHWOU OD ray Rr w ~N are not meaningful. rPUuUrROrRrROD NWWENFHE HE ANnNwWwWwWOUrF PArFORRON NrFArOCOOrFOCSO COrNEFNMNUUYN rR WON ONPFONW wow DANOnNFWOUN 3 2 10 1 2 10 37 10 didi PURPORPROD PURPORPRON DUDE WAUNW NNUFOOrROCO rFPOF HE NAL UW rR PR AWEENERE 3 3 10 L 2 10 36 alk) 80 PURPORPROD DAOPENEWN FPUPRPORROD rPNHUrPOOrFOCO UPNFENHEHE roy Case: (3 DAWRENERE 3 2 10 al 2 10 39 15 84 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C - 38 MIDDLE woo PPrOrRPrROO MWOwWwWr Wwe rPUrROrRrF ODN Deen W rPHyHuUuUrPOOrOCO WNHOFFEFNHLHKE rR COMMEND Output Summary Region: Case: HIGH Floorstok (Million Ft2) 4.9 4.9 New Equip. Elec. Shares (%) HEAT (1) (1) WATR (1) (1) Average Elec. Shares (%) HEAT 6.5 WATR, 18:3, 18: New Equipment EUI (Kwh/Ft2/Yr) HEAT (1) (1) coOoL (1) (1) VENT (1) (1) WATR ~ (1) (1) COOK (1) (1) REFR- (1) (1) LITE (1) (ab) MISC (1) (1) Average EUI (Kwh/Ft2/Yr) HEAT COOL VENT WATR COOK REFR LITE MISC Electric Intensity (Kwh/Ft: HEAT COOL VENT WATR COOK REFR LITE MISC TOTAL KF nr OF FO ANN WWwWOUPD WRFARrFOORFOONKRFDFOFF OW Unfon Si sup G tol Girone ona sg) calor toy unl es 2 WRrRArOOrFCO FANNWNHUOUD Pr rw Adjusted Electric Sales (Gwh/yr) HEAT 3 CooL 2 VENT 10 WATR 1 COOK 2 REFR 9 LITE 38 MISC 9 TOTAL ELECTRIC SALES 76 re Ww HDowmwonronw ~ Note (1): Due to lack of new construction during these years, weighted averages across building types are not meaningful. ww NOUrFPOrRrF OD ON OWHRErWO KFPuUuUrF OFF ON UrFRNNNUWA PARPORFOD AnNwNwwour PArRPORPRON WRFArOOrFCO FADNWNHAOUD NVEFArOOrFCS NOrPNFYONUW r rw r w HDOowonron w ~ won DANN FwWNUD 3 3 10 1 2 10 38 12 81 Fur OFF ON FurFrorrd on NNUPRPOOrFOSO COFDFEFNUFU rR PR NWEENE LW NUN WAWUN 3 3 12 al 2 12 39) 14 87 PURPORRFOM WUONWNEWW FPUuUrFOrFrFON RPNoUrFOOrFOCSO FRNHFEFNHHKE re oo NOP RNR ERE 3 3 12 iz 3 12 43 | 95 PURPORPRFOD PRPRPORFRFOD PNUrFOOrFCO NNOKFEFNWHEE rh DAowkrHwwo wo NPR ENP RW 3 3 13 Z 3 13 47 21 105 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C-39 Residential (AKREM) and Commercial (COMMEND) End Use Model Output C- 40 Appendix D: Industrial Forecast Assumptions by Industry This appendix contains 3 tables describing the LOW, MIDDLE, and HIGH Case industrial forecast assumptions at the industry level. This page intentionally left blank BASE CASE PROJECTED UTILITY SUPPLIED INDUSTRIAL LOADS 1987 1990 1995 2000 2005 2010 BY REGION AND USE GWh MW GWh MW GWh MW GWh MW GWh MW GWh MW KENAI TOTAL (S70 33, 15925 31° 142.5 33° (145.5 34 148.5 3 151.5 36 Petroleum Refining 128.7 21 105.0 47 7105-0 17 105.0 iy | Aa. 47 905.0 a7 Chevron USA Ref 8.2 8.7 8.7 8.7 8.7 8.7 Tesoro Refinery 89.3 eh) 59.5 59.5 59.5 59-5 ARCO Alaska 12.6 12.6 12.6 12.6 12.6 12.6 Phillips 17.8 19.1 19.1 19.1 19.1 1921 Other Petroleum 0.9 51 5.1 5.1 5.1 5.1 General Man 20.2 Z| eos0 6) 27-5 9 30.0 10) 32-5 ot 65-0) 1 Fish Processing 8.8 6 925 6 10.0 OCs a tl eO i We 8 Gas Liquifaction 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 FRBKS TOTAL 64.6 13) 6827, 14 69.7 16) 72.8 15 74.8 15, +76-8) 16 Pet Refining 49.8 ur inD Bin52-0 85320 Be 5620 Sr 5520) 8 Mining 929 5) ) S 9:9 S)al2e0) 4 1550 4 14.0 4 Pet Transportation 4.1 1 7.0 i 7.0 ie. 7.0 2 7.0 2 7.0 2 Construction 0.8 1 0.8 ‘| 0.8 1 0.8 1 0.8 1 0.8 1 ANCH TOTAL See 14 34.0 1Syeson7, 15) Seo 16 39.4 ae 18 General Manufacturing 52.2 14 34.0 Ses orig, 1Si sta 16° 39.4 Wienned a>: 18 Homeporting 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 MATSU TOTAL 1.4 2 iS 2 4.5 4 7.0 5 1.5 5 8.0 6 Construction 1.4 2 4-5 fe 1.5 2 1.5 2 1.5 2 1.5 2 Ski Facilities 0.0 0 0.0 0 1.5 2 deo 2 1.5 Z 4.5 2 Mining 0.0 0 0.0 0 a5: 0 3.0 1 3.0 1 ae0) 1 Other 0.0 0 0.0 0 0.0 0 1.0 0 5 1 2.0 1 TOTAL 255.9 62 243.7 62 252.4 66 262.8 70 270.2 Ws) 206 DB NOTES: KENAI petroleum production based on HEA projections. KENAI general manufacturing increases with Homer dock, Seward industrial facilities and unspecified addition KENAI fish processing increases based on HEA projections extrapolated. FAIRBANKS petroleum refining includes GVEA forecast to 1990 and no significant subsequent expansion. FAIRBANKS mining includes new mining activity in the late 1990s. FAIRANKS construction and petroleum transportation based on GVEA projections. ANCHORAGE general use increases 1% annually after 1990. MATSU construction remains constant. MATSU ski developments occur in early 1990s. MATSU unspecified mining commences in mid-1990s. MATSU other unspecified manufacturing activities begin in 2000. Table D.1: MIDDLE Case Industrial Assumptions by Industry LOW CASE PROJECTED UTILITY SUPPLIED INDUSTRIAL LOADS 1987 1990 1995 2000 2005 2010 BY REGION AND USE GWh MW GWh MW GWh MW GWh MW GWh MW GWh MW KENAI TOTAL 157.7 351395 31 66.0 20 67.5 21 69.0 2170-5 22 Petroleum Refining 128.7 21 105.0 17 30.0 5 30.0 5) | 330.0 5 30.0 5 Chevron USA Ref 8.2 8.7 8.7 8.7 8.7 8.7 Tesoro Refinery 89.3 62.5 0.0 0.0 0.0 0.0 ARCO Alaska 12.6 12.6 0.0 0.0 0.0 0.0 Phillips 17.8 1954 19.1 1931 19.1 19.1 Other Petroleum 0.9 2.1 eee 2.2 2.2 eee General Man 20.2 @ 25.0 8 26.0 827.0 9 28.0 Oi 2920 9 Fish Processing 8.8 6 9:5 6 10.0 @ | 10-5 une AY f. 19E5 8 Gas Liquifaction 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 FRBKS TOTAL 64.6 13, 68.7 14 68.7 14 68.7 14 68.7 14 68.7 14 Pet Refining 49.8 (51-0 Si D120 8 51.0 8 51.0 8 51.0 8 Mining S29) 3 99, S 9.9 5 9.9 3) 9.9 3 9.9) 5 Pet Transportation 4.1 1 acO 2 7.0 2 7.0 2 7.0 2 7.0 Q Construction 0.8 1 0.8 1 0.8 1 0.8 1 0.8 q 0.8 1 ANCH TOTAL Sena 14 34.0 45) | (34.0 15 ||) 3420 15 34.0 15 34.0 15 General Manufacturing Base 14 34.0 15 34.0 15. 34,0 15)..34.0 15 34.0 15 Homeporting 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 MATSU TOTAL 1.4 2 1.5 2 3.0 4 4.0 4 4.5 4 5.0 5 Construction 1.4 2 Vor) 2 1.5 2 oe} 2 1.5 2 125 2 Ski Facilities 0.0 0 0.0 0 V5 2 1.5 a 155) ie 1.5 2 Mining 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 Other 0.0 0 0.0 0 0.0 0 1.0 0 1.9) 1 2.0 1 TOTAL 255.9 62 243.7 62 171.7 53. 174.2 54 176.2 55 178.2 55 NOTES: KENAI petroleum production based on HEA projections. KENAI general manufacturing increases with Homer dock, Seward industrial facilities and unspecified addition KENAI fish processing increases based on HEA projections extrapolated. FAIRBANKS petroleum refining includes GVEA forecast to 1990 and no significant subsequent expansion. FAIRBANKS mining includes new mining activity in the late 1990s. FAIRANKS construction and petroleum transportation based on GVEA projections. ANCHORAGE general use increases 1% annually after 1990. MATSU construction remains constant. MATSU ski developments occur in early 1990s. MATSU unspecified mining commences in mid-1990s. MATSU other unspecified manufacturing activities begin in 2000. MATSU no unspecified mining occurs. MATSU other unspecified manufacturing activities begin in 2000. Table D.2: LOW Case Industrial Load Assumptions by Industry HIGH CASE PROJECTED UTILITY SUPPLIED INDUSTRIAL LOADS 1987 1990 1995 2000 2005 2010 BY REGION AND USE GWh MW GWh MW GWh MW GWh MW GWh MW GWh MW KENAI TOTAL 1S far Sa1S9-5 SI 162e5) 40 165.5 41 188.5 AO) 19455 50 Petroleum Refining 128.7 21 105.0 17 105.0 17 105.0 17° 105.0 17 105.0 17 Chevron USA Ref 8.2 8.7 8.7 8.7 a0 8.7 Tesoro Refinery 89.3 62.5 62.5 62.5 62.5 0.0 ARCO Alaska 12.6 12.6 12.9 13-2 1359 15.9) Phillips 17.8 19.1 19.1 19.1 1954 Wat Other Petroleum 0.9 2.1 1.8 We) 122 63.3 General Man 20.2 72520 8 era 9 30:0 10 3225 afi 3520 1 Fish Processing 8.8 6 955 6 10.0 Olas Cate) Wh Veo 8 Gas Liquifaction 0.0 0 0.0 0 20.0 WA ede. Tt eoso 14 40.0 14 FRBKS TOTAL 64.6 IS OSaf, 14 92.7 18), 9928 19 105.8 20 111.8 21 Pet Refining 49.8 oitieO) 875.0) 11. 80.0 12 8550 13)/590.0 aS) Mining 9.9 3 9.9 5 9.9 3) es0) G 1350 1650 4 Pet Transportation al 1 7.0 2 720 2 7.0 2 7.0 2 7.0 2 Construction 0.8 1 0.8 1 0.8 1 0.8 1 0.8 1 0.8 1 ANCH TOTAL 32.2 14 34.0 15) 5186 24 5501 26) (5925 28 «63.8 30 General Manufacturing Seve, NG 3620) 15.5134 16; 41.1 18 45. 20 «49.8 22 Homeport ing 0.0 0 0.0 0 14.0 8 16. 8 14.0 8 1620 8 MATSU TOTAL 1.4 2 5: 2 4.5 4 7.0 > 1055 8) 225) 9 Construction 1.4 2 i-5) 2 1.5 2 1.5 2 1.5) 2 AS. 2 Ski Facilities 0.0 0 0.0 0 1.5 2 1.5) 2 3.0 4 3.0 4 Mining 0.0 0 0.0 0 a5) 0 3.0 1 4.5 1 6.0 2 Other 0.0 0 0.0 0 0.0 0 1.0 0 i) 1 2.0 1 TOTAL 255.9 62 243.7 62nrsiiet 86 327.4 91 364.1 104 379.6 109 NOTES: KENAI petroleum production based on HEA projections. KENAI general manufacturing increases with Homer dock, Seward industrial facilities and unspecified addition KENAI fish processing increases based on HEA projections extrapolated. FAIRBANKS petroleum refining includes GVEA forecast to 1990 and no significant subsequent expansion. FAIRBANKS mining includes new mining activity in the late 1990s. FAIRANKS construction and petroleum transportation based on GVEA projections. ANCHORAGE general use increases 1% annually after 1990. MATSU construction remains constant. MATSU ski developments occur in early 1990s. MATSU unspecified mining commences in mid-1990s. MATSU other unspecified manufacturing activities begin in 2000. LE a a i IS IB Ai te AAA TRE TR TETAS. i ih ETE ERA TT OE EE ARE COE, Table D.3: HIGH Case Industrial Load Assumptions by Industry This page intentionally left blank Appendix E: Development of Historical Load Factors This appendix contains four tables on which an average historical load factor for each of the four regions of the Railbelt is derived. This page intentionally left blank LOAD FACTOR CALCULATION ANCHORAGE BOWL CEA(1) ANLEP (2) TOTAL ENERGY PEAK LOAD ENERGY PEAK LOAD ENERGY PEAK LOAD ROMNTS LOAD FACTOR RQMNTS LOAD FACTOR RQMNTS LOAD(3) FACTOR YEAR (MWH) = (MW) (3) ( MWH) (MW) ($) (MWH) = (MW) (% 1970 238395 49.1 55.4 1971 282088 55.7 57.8 1972 312705 62.7 56.9 1973 350252 66.6 60.0 1974 488866 = 131.6 42.4 378711 74.4 58.1 867577 199.8 49.6 1975 692470 156 50.7 436769 89.5 55.7 1129239 238.1 54.1 1976 759792 = 149.6 58.0 500618 93.4 61.2 1260410 = 235.7 61.0 1977 840723 181 53.0 535430 = 101.5 60.2 1376153 274.0 57.3 1978 872801 = 181.1 55.0 550194 99 63.4 1422995 = 271.7 59.8 1979 938752 186 57.6 578870 109 60.6 1517622 286.2 60.5 1980 945215 214.6 50.3 585778 121 55.3 1530993 325.5 53.7 1981 959170 = 215.7 50.8 590908 14 59.2 1550078 319.8 55.3 1982 1062803 = 227.1 53.4 650627 121 61.4 1713430 = 337.7 57.9 1983 1079609 = 223.4 55.2 676727 127 60.8 1756336 339.9 59.0 1984 1103444 = 196.3 64.2 744170 150 56.6 1847614 335.9 62.8 1985 1119357 196.9 64.9 849192 = 153.4 63.2 1968549 339.8 66.1 1986 1125814 = 181.5 70.8 843443 143 67.3 1969257 314.8 71.4 1987 996526 = 175.1 65.0 823899 141 66.7 1820425 306.6 67.8 AVERAGE: 59.7 1984-1987 AVERAGE: 67.0 1982-1987 AVERAGE: 64.2 GENERAL: SYSTEM ENERGY REQUIREMENTS FOR EACH UTILITY=GENERATION+PURCHASES-SALES FOR RESALE PEAK LOAD FOR EACH UTILITY=PEAK(GENERATION+PURCHASES-SALES FOR RESALE) LOAD PACTOR=ENERGY REQUIREMENTS/(PEAK LOAD#8760 HOURS) (1) CBA ANCHORAGE BOWL REQUIREMENTS=TOTAL SYSTEM REQUIREMENTS-(MBA SALES+HEA SALES+SES SALES) CEA PEAK LOAD=TOTAL SYSTEMPEAK LOAD-WHOLESALE COINCIDENT PEAK LOAD WHERE COINCIDENT PEAK WHOLESALE LOAD IS ACTUAL FOR 1982-1987, ESTIMATED FOR 1974-1982 USING THE AVERAGE RATIO OF (NONCOINCIDENT PEAK/COINCIDENT PEAK) FOR ACTUAL 1982-1986 DATA SOURCE: 1983 AND 1987 CEA POWER REQUIREMENTS STUDIES (2) SOURCE: MIKE KIECH, AML&P. AUGUST, 1988 Table E.1: Load Factor Calculations, Anchorage Region E-1 LOAD FACTOR CALCULATION FAIRBANKS GVEA(1) FNUS (2) TOTAL ENERGY PEAK LOAD ~=©=»sENERGY © PEAK «= LOAD. «=SsENERGY © PEAK ©~=—sLOAD ROMNTS LOAD FACTOR =ROMNTS LOAD FACTOR ~=ROMNTS~=«LOAD(3) FACTOR YEAR (MAH) = (MW) 8) (KWH) = (MW) (8) (MWR) = (MW) (8 1970 1583468 N/A 1971 18725946. 1972 210619 45. 1973 22401552. 49, 1974 24884464, 44. 1 46, 9 2 4 1975 327766 81.6 45, 4 9 6 bz. 1976 339498 76. 50. 1977 353110 89. 44, 1978 = 341489 11. 54, 1979 325518 75.7 49, 1980 © 314589 70 51. 1981 = 316861 68.7 52. 1982 350159 67.9 58. 1983 361887 72.2 57. 1984 394071 77.5 58. 1985 422495 81.4 59. 1986 © 446239 85.2 59, 1987 459011 82.7 63. 128105 26.5 55.2 478264 91.6 59.6 124783 28.2 50.5 486670 97.4 57.0 134693 29.2 52.7 528764 = 103.5 5 141433 27.4 58.9 563928 105.5 61.0 145601 31.2 53.3 591840 112.9 5 136904 28.5 54.8 595915 107.9 BPOwWOM DO IWH ek &® ISH oO & AVERAGE: 1984-1987 AVERAGE: 1982-1987 AVERAGE: nn wow Co mH & GENERAL: SYSTEM ENERGY REQUIREMENT FOR EACH UTILITY=GENERATION+PURCHASES-SALES FOR RESALEE PEAK LOAD FOR EACH UTILITY=COINCIDENT PEAK(GENERATION+PURCHASES-SALES FOR RESALE) LOAD FACTOR=ENERGYREQUIREMMENTS / (PEAK LOAD¥8760 HOURS) (1) SOURCE: GVEA POWER REQUIREMENTS STUDY (2) SOURCE: ALASKA POWER AUTHORITY (3) ASSUMES 97% COINCIDENCE FACTOR - RECA eee AREAS ER iia i CO kl AAT OE i i 0 a aD TO Table E.2: Load Factor Calculations, Fairbanks E-2 LOAD FACTOR CALCULATION KENAI ENERGY PEAK LOAD ENERGY PEAK LOAD ENERGY PEAK LOAD ROMNTS LOAD FACTOR RQMNTS LOAD FACTOR RQMNTS LOAD(3) FACTOR YEAR (MWH) (MW) (3) (MWH) = (MW) (8) (MWH) = (MW) (8) 1970 1971 1972 197396485 18.5 59.5 1974 104600 20.4 58.5 16008 5.8 31.5 120608 25.4 54.2 1975 104338 25.8 46.2 15483 3.9 45.3 119821 28.8 47.5 1976 = 152034 30 57.9 19181 4.1 53.4 171215 33.1 59.1 1977 178116 38.7 52.5 18035 4.2 49.0 196151 41.6 53.8 1978 = 211219 45.2 53.3 23175 7 37.8 = 234394 50.6 52.8 1979 232250 50 53.0 24454 4.9 57.0 256704 53.3 55.0 1980 252628 51.8 55.7 25964 5 59.3 278592 55.1 57.7 1981 284836 59.3 54.8 27254 5.1 61.0 312090 62.5 57.0 1982 312708 60.4 59.1 29542 5.3 63.6 342250 63.7 61.3 1983 348593 67.5 59.0 30533 5.7 61.1 379126 71.0 61.0 1984 372628 64.7 65.7 32990 6.2 60.7 405618 68.8 67.3 1985 404775 11 65.1 31457 5.9 60.9 436232 74.6 66.8 1986 © 405133 72.6 63.7 34367 6.3 62.3 439500 76.5 65.6 1987 = 412774 74.6 63.2 36251 6.7 61.8 449025 78.9 65.0 AVERAGE: 58.9 1984-1987 AVERAGE: 66.2 1982-1987 AVERAGE: 64.5 GENERAL: SYSTEM ENERGY REQUIREMENTS FOR EACH UTILITY=GENERATION+PURCHASES-SALES FOR RESALE PEAR LOAD FOR EACH UTILITY=PEAK(GENERATION+PURCHASES-SALES FOR RESALE) LOAD PACTOR=ENERGY REQUIREMENTS/(PEAK LOAD* 8760 HOURS) (1),(2) SOURCE: 1987 CHUGACH ELECTRIC POWER REQUIREMENTS STUDY (3) ASSUMES 97% COINCIDENCE FACTOR LS a A SANE AE TILA he A I TT BE SY SAR AIR LE AS AA A NS ET RR ST A Table E.3: Load Factor Calculations, Kenai Region E-3 LOAD FACTOR CALCULATION MATSU YEAR ENERGY RQMTS(1) PEAK LOAD(2) LOAD FACTOR(3) MWH MW PERCENT 1970 55878 13.7 46.56 2973 65847 14.8 50.79 i972 73910 16.6 50.83 1973 84151 20.8 46.18 1974 98919 29 38.94 1975 133860 35.4 43.17 1976 159979 ao. 51.44 L977 199625 52. 43.74 1978 243811 59.3 46.93 1979 253646 69.7 41.54 1980 268014 68.7 44.53 1981 273860 aoa 41.63 1982 345886 83.4 47.34 1983 362083 89.2 46.34 1984 416417 91.3 52.07 1985 462205 88.7 59.48 1986 444484 90.5 56.07 1987 430616 93). 52.80 AVERAGE: 47.51 1984-1987 AVERAGE: 55.10 1982-1987 AVERAGE: 52.35 (1) SYSTEM ENERGY REQUIREMENTS=GENERATION+PURCHASES-SALES FOR RESALE (2) PEAK LOAD=PEAK(GENERATION*PURCHASES-SALES FOR RESALE) (3) LOAD FACTOR=ENERGY REQUIREMENTS/(PEAK LOAD*8760 HOURS) SOURCE: CHUGACH ELECTRIC ASSOC. NOVEMBER 1987 POWER REQUIREMENTS ST ean ————_—_—_—_—_= Table E.4: Load Factor Calculations, MatSu Region rau Appendix F; Summary Responses to the Residential End Use Survey This appendix contains a reprint of Appendix A from ISER’s Technical Memorandum of 18 February 1988 describing the Residential End Use Survey. The residential survey data set consists of 905 responses to over fifty major questions about fuel choice, appliance holdings, thermal integrity, and energy-related behavior. More than half of the records have actual electric consumption data attached to the characteristics of the household. The data is maintained as an SPSS system file on the University of Alaska’s VAX mainframe computer. Researchers interested in using the data for energy-related research are invited to contact the authors. This Page Intentionally Left Blank APPENDIX A: PRELIMINARY SUMMARY RESULTS The following tables report summary data in the form of raw frequencies for each of the four major Railbelt regions and a weighted total Railbelt frequency distribution. Because the numbers in the "Weighted Railbelt" column are weighted results, the row totals of responses across the four constituent regions does not add to the "Weighted Railbelt" total number of responses. Due to rounding operations during the weighting process, the total number of "weighted respondents" for the Railbelt sometimes adds to 904 instead of the exact correct total of 905. Guide to Tables +3 i io fa Oo Description Major Heating Fuel Source Fuels used for at least Some Heating Hot Water Heater Fuel Measures Taken by Electric Water Heater Users Type of Home Home Ownership Floor Space in Home Household Size Wall Construction Ceiling Insulation Thickness Panes of Glass in Windows Type of Stove Separate Freezer Used ? Clothes Dryer Ownership & Type Appliance Saturations: Clothes Washer, Microwave, Diswasher, TV, Computer, VCR 1 RRR RPP ODNADUFWHH UPWNrO ne Bee Dee te te ie ie i oe Weighted Natural Gas Electricity Cy YY Zs \ iS 8 Me a See Sed Table A-2 Fuel Source Used For at Least Some Heating Weighted Railbelt Anchorage Fairbanks Kenai Matsu Fuel Source: n % n % n % n % n % Natural Gas 513 57 394 82 0 0 319))||||29 20 19 oil 203°) (22 13 3 155: 86 54 40 30 28 Electricity 2:22) 25 94 20 33) |/28 S2.||||' 38 50 46 Wood 197 | 22) 50 10 51 28 65 48 59 55 Propane 18 2 4 1 4 2 4 3 10 9 City Steam 17 2 15 3 2 a 0 0 0 0 Coal 12 dL 0 0 4 2 5 4 0 0 Kerosene 9 rE 4 a 2 1 1. a 4 4 Solar Gain 7 a 1 0 1 ni 4 3 2 2 Other 4 |I\|0 II a Jal) || 2 oI) (0]||||| (0 Total 2 1202 133 578/120 253 141 223 164 175 162 % Using Fuel For Some Heat Note: Total percentages exceed 100 since some respondents use more than one fuel. Fuel Used for Some Heat N NN UN VA WA MHAMACHON aad a ANNTAAAON ddl ad onwvovronn Naw trRe+TONONMNM ra ONNADANr tS aon On OCOANXNAN St ad ad ~ADOOAMNDAO 0 N ANMWAANYTHAO N oa qo ~HOAONMNAKR eA oO AANNMNOADKDHE NOOO MANA oO j3 UV0BAIG=- DM jPHVOUIOOLE GAA ONVOAO oqo BZoOowMmeaocagym Hot Water Heater Fuel ocmUCTTCGSSCsCiaaSs—C BGS S§§§ C DFTs - GFT— ss DF ss DF sD o ® O& RF © OB + MW NN & Table A-4a Electric Water Heater Users Only Extra Insulation on Water Heater Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % Yes 86 30 26 | 27 Ly 24 30 41 2a, | 34 No 175) ) (62) 60 61 SO) | 7a. 40 54 38 61 Boiler Only (Inap.) a 0 1 a 0 0 0 0 0 0 Don't Know 20 7 pe Se & 3 3 4 4 5 2 5 Total) % 282 100 98 100 70 100 74 100 62 100 Table A-4b Electric Water Heater Users Only Hot Water Reduced to 120 Degrees or Lower Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % Yes 799) 35 26 27 316) |) 52: 22) 29 20 32 No 86 31 29 6 «639 2029 47 42 an 630 Don't Know 96 34 43 44 14 20 43 39 20) )) SZ Total : 281 100 98 100 70 100 111 100 62 100 F-6 Yes No Don't Know Total Table A-4c Electric Water Heater Users Only Showers Fitted With Flow Restrictors Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % 95 34 33 34 25) SS 20. 27 27 44 179 63 62 63 44 63 53 72 BS (53 8 3 3 3 3 4 1 i 2 3 282 100 98 100 70 100 74 100 62 100 F-7 Table A-5 Type of Home Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % Single Family 557 62 262 54 129 72 100 74 89 82 Mobile Home 84 9 39 8 LS 8 20 15 9) 8 Multi-family 262 29 178 37 36 20 16 12 10 9 Not Ascertained 2 0 2 0 0 0 0 0 0 0 Total : 905 100 481 100 180 100 136 100 108 100 Table A-6 Home Ownership Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % Own — 655 72 319 66 147 82 103 76 91 84 Rent 212 =—=23 133 28 oa sd 26 19 16 6415 Occupy Without Rent 33 4 25) 5 2 1 Zz 5 ay 1 Not Ascertained 5 1 4 ad: 0 0 0 0 0 0 Total : 905 100 481 100 180 100 136 100 108 100 F-8 Square Feet Under 1000 1000 to 1499 1500 to 1999 2000 and over Don't Know Total # in Household WODNIADAUARWNE Total : Table A-7 Floor Space in Home Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % 147 16 65 14 37) 22 29° 21 14. 13 215 24a 117 24 36 20 39°29 2S ta 188 21 99 721 36: -20 29° 2 28 26 244 27 13%, 28 45 25 27° 20 S7* aa a2 66 14 26 14 a2 9 6 6 905 100 481 100 180 100 136 100 108 100 Table A-8 Household Size Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % 118 13 63 13 29 16 16 #12 13; “12 276 3L 141 29 58 32 49 36 32 30 198 22 109 23 29 16 32 24 20 19 192 21 109 23 38) 2a 27 20 22 20 85 9 40 8 21°. 12 8 6 15 14 a3 2 ak 2 al 2 4 3 5 § 8 al 5 al 3 2 0 0 0 0 3 0 2 0 0 0 0 0 a 1 2 0 i 0 1 pI 0 0 0 0 903 100 481 100 180 100 136 100 108 100 F-9 i) X 2x 2x Logs Mobile Home Double Wall Othe Don't Know percent of respondents Ons ite Total ¢ 100 90 ¢ 80 70 60 50 40 30 20 10 Table A-9 Wall Construction Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % 454 50 283 59 61 34 62 46 39 36 275 310 128 27 61 34 42 31 51 47 12 al: aL: 0 8 4 a 1 2 2 42 5 La 2 20° Ad 10 7 5 5 55 6 29 6 4 2 13 10 7 6 9 1 0 0 8 4 1 dl: 0 0 od 2 10 2 5 3 3 2 a J 37 4 19 4 iS) 7 4 3 3 3 905 100 481 100 180 100 136 100 108 100 Wall Construction n he LLL p a9 RF SSNS KKK SPY 9 222 eae Table A-10 Ceiling Insulation Thickness Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % Less than 6 inches 97 7 LL 54 ti as) 7 a1, 38 7 6 6 to 10 inches 214 24 90 19 54 30 ne 38 35 11 to 30 inches 125) |) LA 45 9 46 26 24 18 19 18 Don't Know 467, 52 29262 66°37 43 32 44 41 Not Ascertained 2 0 0 0 al 1 1 1 0 0 Total >: 905: 100 481 100 180 100 1316 |-00 108 100 Table A-11 Panes of Glass in Windows Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % i 146 9 7916 29) 6 33 24 10 9 2 674 41 Sian ves 106 59 SO 7S 91 84 3 Td 9 22 5 44 24 4 3 i 6 4 2 0 3 al 0 0 0 0 0 0 Don't Know/NA 4 0 3 al al 1 0 0 0 0 Total. ; 904 59 481 100 180 100 136 100 108 100 AVDOONOO 6 N nnRONNOCOO 3 N NoONtdOd , a AANA Od CVA wOONAAO SS N ~9O ee a'° o anmdA OOO ~N NANMNNTAO ro NADTAOO won aOoOrUuoNnad am O==0 =O >= 0 -0-— o> — © 6-0-2 -o oOo @ @ - <¢@ 0) ¥ 1 An = 7 Ty @nuoOgagYv Ho@m@yvood a > WMVUANAOZ!SN Yes Total Table A-13 Separate Freezer Used Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % 545 60 260!) 154 1S") 63) O17) eval 90% 83 360 40 ay) 46 Gyan siz Sine) oT oer 905 100 481)) 2.00 180 100 1316) TOO 108 100 F - 13 percent of respondents None Electric Natural Gas Bottled Gas Don't Know Total) 3 100 90 80 70 60 50 40 30 20 10 Table A-14 Clothes Dryer Ownership and Type Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % 124 14 62)) 23 29) |) 126 228) aS) Bae | | ke 658/73 337 || 70 147 82 94 69 eZ | | 8a 106 12 Ti) Xe) 0 0 16) a2 5 5 10 1 0 0 4 2 5 4 4 4 az a 5 1 0 0 0 0 0 0 905 100 481 100 180 100 136 100 108 100 Dryer Type Table A-15a Number of Clothes Washers Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % 0 120 T2 sy ee ae 28 16 19 #14 9 8 1 792. 87 424 88 151 84 117 86 99 92 2 5 al 3 1 1 ab 0 0 0 0 Total : 906 100 481 100 180 100 136 100 108 100 Table A-15b Number of Microwaves Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % 0 159 18 86 18 3a «17 25 - 18 live “16 z 728 80 384 80 145 81 110 81 89 82 Z 18 a ulal 2 4 2 1 al 2 2 Total : 905 100 481 100 180 100 136 100 108 100 Table A-15c Number of Dishwashers Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % 0 241 27 88 18 85 47 55 40 30 28 1 661 73 392 81 94 52 81 60 78 72 2 2 0 1 0 1 1 nn) 0 0 Total : 904 100 481 100 180 100 136 100 108 100 F- 15 0 1 2 3 4 5 6 7 Total 0 1 2 3 4 5 6 Total Table A-15e Number of Televisions Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % 5) a 6 aL 2 a 5 4 3 3 316 16 149 31 69 38 66 49 43 40 335 377 194 40 58 32 34 25 40 37 145 16 79 416 30 A 20 @J5 16 15) 64 7 Si 8 9 5 8 6 6 6 18 2 8 2 8 4 2 1 0 0 9) 1 6 a 2 al AE a 0 0 8 0 2 0 all a: 0 0 0 0 905 81 481 100 180 100 136 100 108 100 Table A-15f Number of Computers Weighted Railbelt Anchorage Fairbanks Kenai Matsu n % n % n % n % n % 658 73 353 73 122 68 102 75 87 81 212, 23) 112° «23 44 24 29° 22 20 29 29 3 2) 7) ul} 7 5 4 Al 1 2 0 Zz 0 a 1 0 0 0 0 2 0 1 0 al 1 0 0 0 0 a 0 i: 0 0 0 0 0 0 0 al 0 1 0 0 0 0 0 0 0 904 100 481 100 180 100 136 100 108 100 Appendix G: Calculation of Projected Gas Conversions This appendix contains calculations of projected conversion activity from electricity to gas in residential households. The tables are organized by region and case. Two cases are developed for use in the probability tree: the BASE case (probability 50%) assumes continuing conversion of existing heating systems at historical rates, consistent with the total housing stock available for retrofit. The HIGH penetration case (probability 50%) reflects extension of gas service to Homer and to several major unserved subdivisions in the MatSu valley (Meadow Lakes, Big Lake, and Butte/Lazy mountain). These cases are discussed in greater detail in section 2.2.5. This page intentionally left blank Electric to Gas Conversions Calculations Region: ANCHORAGE Case: BASE Penetration Total Units Total Units Gaining Potent- Electric Retrofit Total # of Retrofits in Access ial Percent Rate Retrofits *ELECTRIC TO GAS* Initial to Gas of CALL CAUD eaae ema e eae Year Gas Retrofits Pool Fuels) Fuels) Single Multi Mobile 1987 : 88,750 21000 21000 0.6 0.005 105 47 8 8 1988 : 88,306 0 20895 0.6 0.005 104 47 8 8 1989 : 87,865 0 20791 0.6 0.005 104 47 8 8 1990 : 87,425 0 20687 0.6 0.005 103 47 8 8 1991 : 86,988 0 20583 0.6 0.005 103 46 8 8 1992 : 86,553 0 20480 0.6 0.005 102 46 8 8 1993 : 86,121 0 20378 0.6 0.005 102 46 8 8 1994 : 85,690 0 20276 0.6 0.005 101 46 8 8 1995 : 85,262 0 20175 0.6 0.005 101 45 8 8 1996 : 84,835 0 20074 0.6 0.005 100 45 8 8 1997 : 84,411 0 19973 0.6 0.005 100 45 7 a 1998 : 83,989 0 19873 0.6 0.005 99 45 7 7 1999 : 83,569 0 19774 0.6 0.005 99 44 7 7 2000 : 83,151 0 19675 0.6 0.005 98 44 7 rh 2001 : 82,735 0 19577 0.6 0.005 98 44 7 7 2002 : 82,322 0 19479 0.6 0.005 97 44 t EA 2003 : 81,910 0 19382 0.6 0.005 97 44 7 7 2004 : 81,501 0 19285 0.6 0.005 96 43 7 7 2005 : 81,093 0 19188 0.6 0.005 96 43 7 % 2006 : 80,688 0 19092 0.6 0.005 95 43 7 7 2007 : 80,284 0 18997 0.6 0.005 95 43 7 7 2008 : 79,883 0 18902 0.6 0.005 95 43 o 7 2009 : 79,483 0 18807 0.6 0.005 9% 42 7 iu 2010 : 79,086 0 18713 0.6 0.005 9% 42 i v TOTAL RETROFITS 21000 POTENTIAL; ACTUAL= 2380 =1071 +179 179 NOTES:* 21000 residential units ("full unit equivalents") do not currently use gas in Anchorage for space heating. * Rate of retrofits from personal communications with utilities and others. * Distribution of retrofits: 0.75 single family 0.125 multifamily 0.125 mobile home LE, STOTT LT EEO ID ETN EE TT TE, TIE OE TEDL LEE, ELLE, STEELE EELS, SOLACE Table G.1: Projected Gas Conversions, Anchorage, Base Penetration Electric to Gas Conversions Calculations Region: ANCHORAGE Case: HIGH Penetration Total Units Total Units Gaining Potent- Electric Retrofit Total # of Retrofits in Access ial Percent Rate Retrofits *ELECTRIC TO GAS* Initial to Gas of (ALL CAL terete ea nmaninnem Year Gas Retrofits Pool Fuels) Fuels) Single Multi Mobile 1987 88,750 21000 21000 0.6 0.005 105 16 41 6 1988 88,306 0 20895 0.6 0.005 104 16 41 6 1989 87,865 0 20791 0.6 0.005 104 16 41 6 1990 87,425 0 20687 0.6 0.02 414 62 161 25 1991 86,988 0 20273 0.6 0.02 405 61 158 24 1992 86,553 0 19867 0.6 0.02 397 60 155 24 1993 86,121 0 19470 0.6 0.02 389 58 152 23 1994 85,690 0 19081 0.6 0.02 382 57 149 23 1995, 85,262 500 19199 0.6 0.02 384 58 150 23 1996 84,835 0 18815 0.6 0.02 376 56 147 23 1997 84,411 0 18439 0.6 0.02 369 55 144 22 1998 83,989 0 18070 0.6 0.02 361 54 141 22 1999 83,569 0 17709 0.6 0.02 354 53 138 21 2000 83,151 0 17354 0.6 0.02 347 52 135 21 2001 82,735 0 17007 0.6 0.02 340 on 133 20 2002 82,322 0 16667 0.6 0.02 333 50 130 20 2003 81,910 0 16334 0.6 0.02 327 49 127 20 2004 81,501 0 16007 0.6 0.02 320 48 125 19 2005 81,093 0 15687 0.6 0.02 314 47 122 19 2006 80,688 0 15373 0.6 0.02 307 46 120 18 2007 80, 284 0 15066 0.6 0.02 301 45 118 18 2008 79,883 0 14764 0.6 0.02 295 44 115 18 2009 79,483 0 14469 0.6 0.02 289 43 113 17 2010 79 , 086 0 14180 0.6 0.02 284 43 111 17 TOTAL RETROFITS 21500 POTENTIAL; ACTUAL= 7604 1141 2965 456 NOTES:* 21000 residential units ("full unit equivalents") do not - currently use gas in Anchorage for space heating. * Rate of retrofits from personal communications with utilities and others. * Distribution of retrofits 0.25 single family 0.65 multifamily 0.1 mobile home I PE TT TE TT ct DOAN AT ITE CONT EEE ETD ITLL ES TE, 3 OTE PETITE i ATS | Table G.2: Projected Gas Conversions, Anchorage, High Penetration Electric to Gas Conversions Calculations Region: KENAI Case: BASE Penetration Total Units Total Units Gaining Potent- Electric Retrofit Total # of Retrofits in Access ial Percent Rate Retrofits *ELECTRIC TO GAS* Initial to Gas of (ALL CAL epee nennnn Haase === Year Gas Retrofits Pool Fuels) Fuels) Single Multi Mobile 1987 : 15,980 2000 2000 0.4 0.05 100 25 8 a 1988 : 15,900 0 1900 0.4 0.05 95 24 8 a 1989 : 15,821 0 1805 0.4 0.05 90 22 7 6 1990 : 15,741 0 1715 0.4 0.05 86 21 7 6 1991 : 15,663 0 1629 0.4 0.04 65 16 5 5 1992 : 15,584 0 1564 0.4 0.04 63 16 5 Di 1993 : 15,507 0 1501 0.4 0.04 60 15 5 4 1994 : 15,429 0 1441 0.4 0.03 43 m4 5 5) A995! |: 15,352 0 1398 0.4 0.03 42 10 5 5 1996) 2) 15,275 0 1356 0.4 0.03 41 10 5) 3 M9OT 15,199 0 1315 0.4 0.02 26 7 2 2 M998 4:77 45, 12s 0 1289 0.4 0.02 26 6 2 2 1999 : 15,047 0 1263 0.4 0.02 25 6 2 2 2000 14,972 0 1238 0.4 0.01 12 3 1 1 2001 14,897 0 1226 0.4 0.01 12 5 1 1 2002 : 14,823 0 1213 0.4 0.01 12 3 1 1 2003 14,748 0 1201 0.4 0.01 12 5 1 1 2004 14,675 0 1189 0.4 0.01 12 5 1 1 2005 : 14,601 0 1177 0.4 0.01 12 io 1 1 2006 : 14,528 0 1166 0.4 0.01 12 3 1 1 2007 : 14,456 0 1154 0.4 0.01 12 3 1 1 2008 : 14,383 0 1142 0.4 0.01 1 2 1 1 2009 : 14,311 0 1131 0.4 0.01 1 5 1 1 2010 : 14,240 0 1120 0.4 0.01 11 S 1 1 TOTAL RETROFITS 2000 POTENTIAL; ACTUAL= 892 221 71 64 NOTES:* Of 8000 units currently not served we assume 0 are vacation homes leaving all 8000 potentially available for service and gas retrofits. * We assume 8000 units currently have access to gas. There are 5300 residential ENSTAR accounts and, assuming 700 units reported in the commercial sector, this leaves 2000 units in the retrofit pool. * Electricity percentage from the residential survey. * Total retrofits by 2010 result in the percentage of gas-accessible buildings which are actually served as: 86% * Rate of retrofits from personal communications with utilities and others. * Distribution of retrofits: 0.62 single family 0.20 multifamily 0.18 mobile home MC IE LAT EET AE OEE a, A AA SNARE LET He IRL NELLA ALLELE ATs ALM AEE Rie etn JER em eee Table G.3: Projected Gas Conversions, Kenai, Base Penetration G-3 Electric to Gas Conversions Calculations Region: KENAI Case: HIGH Penetration Total Units Total Units Gaining Potent- Electric Retrofit Total # of Retrofits in Access ial Percent Rate Retrofits *ELECTRIC TO GAS* Initial to Gas of (ALL CAC rer ec nea Year Gas Retrofits Pool Fuels) Fuels) Single Multi Mobile 1987 : 15,980 2000 2000 0.4 0.05 100 25 8 7 1988 : 15,900 0 1900 0.4 0.05 95 24 8 rd 1989 : 15,821 0 1805 0.4 0.05 90 22 7 6 1990 : 15,741 2000 3715 0.4 0.15 557 138 45 40 1991 : 15,663 2000 5158 0.4 0.15 774 192 62 56 1992 : 15,584 0 4384 0.4 0.1 438 109 35 32 1993 =| 157507, 0 3946 0.4 0.1 395 98 32 28 1994 : 15,429 0 3551 0.4 0.08 284 70 23 20 1995. : 15,352 0 3267 0.4 0.08 261 65 21 19 1996 : 15,275 0 3006 0.4 0.06 180 45 14 13 A9S7 2) | tones 0 2825 0.4 0.06 170 42 14 12 1998 : 15,123 0 2656 0.4 0.04 106 26 8 8 1999 es 15,047 0 2549 0.4 0.04 102 25 8 7 2000 : 14,972 0 2447 0.4 0.02 49 12 4 4 2001 : 14,897 0 2399 0.4 0.02 48 12 4 3 2002 : 14,823 0 2351 0.4 0.01 24 6 2 2 2003 : 14,748 0 2327 0.4 0.01 23 6 2 2 2004 14,675 0 2304 0.4 0.01 23 6 2 2 2005 : 14,601 0 2281 0.4 0.01 23 6 2 2 2006 : 14,528 0 2258 0.4 0.01 2 6 2 2 2007 : 14,456 0 2235 0.4 0.01 22 6 2 2 2008 : 14,383 0 2213 0.4 0.01 22 5 2 2 2009 14,311 0 2191 0.4 0.01 22 5 2 2 2010 14,240 0 2169 0.4 0.01 22 5 2 2. TOTAL RETROFITS 6000 POTENTIAL; ACTUAL= 3853 955 308 277 NOTES:* Of 8000 units currently not served we assume 0 are vacation homes leaving all 8000 potentially available for service and gas retrofits. * We assume 8000 units currently have access to gas. There are 5300 residential ENSTAR accounts and, assuming 700 units reported in the commercial sector, this leaves 2000 units in the retrofit pool. * Electricity percentage from the residential survey. * Total retrofits by 2010 result in the percentage of gas-accessible buildings which are actually served as: 82% * Rate of retrofits from personal communications with utilities and others. * Distribution of retrofits: 0.62 single family 0.2 multifamily 0.18 mobile home 1 * We assume that gas becomes available in the Greater Homer area in 1990, which increases the pool of potential retrofits by 4000 over a 2 year period. Table G.4: Projected Gas Conversions, Kenai, High Penetration Electric to Gas Conversions Calculations Region: MATSU Case: BASE Penetration Total Units Total Units Gaining Potent- Electric Retrofit Total # of Retrofits in Access ial Percent Rate Retrofits *ELECTRIC TO GAS* Initial to Gas of (ALL CALs ee Year Gas Retrofits Pool Fuels) Fuels) Single Multi Mobile 1987 15,710 5000 5000 0.35 0.1 500 135 23 18 1988 15,631 0 4500 0.35 0.1 450 121 20 16 1969: :"-- 15,553 900 4950 0.35 0.1 495 133 23 17 1990 : 15,476 900 5355 0.35 0.1 536 144 24 19 4991: 153390 0 4820 0.35 0.075 361 97 16 13 1992.2 "15,321 0 4458 0.35 0.075 334 90 15 12 1993. =:> 157269 0 4124 0.35 0.075 309 83 14 1 1994 : 15,168 0 3814 0.35 0.075 286 We 13 10 oop 15,092 0 3528 0.35 0.075 265 71 12 9 1996 : 15,017 0 3264 0.35 0.05 163 44 ie 6 1997 : 14,942 0 3101 0.35 0.05 155 42 iz S 1998 : 14,867 0 2945 0.35 0.05 147 40 ie 5 1999: 14,793 0 2798 0.35 0.05 140 38 6 5 2000 : 14,719 0 2658 0.35 0.05 133 36 6 5 2001 : 14,645 0 2525 0.35 0.025 63 17 3 2 2002 : 14,572 0 2462 0.35 0.025 62 17 5 2 2003 : 14,499 0 2401 0.35 0.025 60 16 J 2 2004 : 14,427 0 2341 0.35 0.025 59 16 3 2 2005 : 14,355 0 2282 0.35 0.025 57 15 3 2 2006 : 14,283 0 2225 0.35 0.01 22 6 1 1 2007 : 14,211 0 2203 0.35 0.01 22 6 1 1 2008 : 14,140 0 2181 0.35 0.01 22 6 1 1 2009 : 14,070 0 2159 0.35 0.01 22 6 1 1 2010' = 135,999 0 2137 0.35 0.01 21 6 1 1 TOTAL RETROFITS 6800 POTENTIAL; ACTUAL= 4684 1262 213 164 NOTES * Of 8000 units currently not served we assume 2000 are vacation homes in the Big Lake area and elsewhere. Therefore 6000 are potentially available for service and conversion to gas. In this case we assume the LIDs get gas. This is about 1800 units. * We assume 9000 units currently have access to gas. There are 3400 residential ENSTAR accounts and, assuming 600 units reported in the commercial sector, this leaves 5000 units in the retrofit pool. * Electricity percentage from the residential survey. * Total retrofits by 2010 result in the percentage of gas-accessible buildings which are actually served as: 80% * Rate of retrofits from personal communications with utilities and others. Calculated as the number of "full time equivalent" units converted to gas. * Distribution of retrofits: 0.77 single family 0.13 multifamily 0.1 mobile home RCS CR RTT STAGE: EIT TEESE: SOAS LEOPARD TER! MeO a A | ts NE OS Ei A. ot NNER Table G.5: Projected Gas Conversion, MatSu,; Base Penetration G-5 Electric to Gas Conversions Calculations Region: MATSU Case: HIGH Penetration Total Units Total Units Gaining Potent- Electric Retrofit Total # of Retrofits in Access ial Percent Rate Retrofits *ELECTRIC TO GAS* Initial to Gas of (ALL A Year Gas Retrofits Pool Fuels) Fuels) Single Multi Mobile 1987s) || 15,710 5000 5000 0.35 0.1 500 135 23 18 A988 eitS65n 0 4500 0.35 0.1 450 121 20 16 NOBP si topooS 900 4950 0.35 0.1 495 133 23 17 1990 =: 15,476 900 5355 0.35 0.1 536 144 24 19 AOO1 sis 96 1000 5820 0.35 0.1 582 157 26 20 1992) db sen 1000 6238 0.35 0.1 624 168 28 22 1993S Tr eloyess 1000 6614 0.35 0.1 661 178 30 23 NOS ens 168 0 5952 0.35 0.075 446 120 20 16 eee e 15,092 0 5506 0.35 0.075 413 111 19 14 ASS eiia.ol7 0 5093 0.35 0.05 255 69 12 9 997 214, 962 0 4838 0.35 0.05 242 65 11 8 SSS Use, Beg, 0 4596 0.35 0.05 230 62 10 8 1999 14,793 0 4367 0.35 0.05 218 59 10 8 2000 Fr intae clo 500 4648 0.35 0.1 465 125 21 16 2001 14,645 0 4183 0.35 0.025 105 28 5 4 2002 is 14,572 0 4079 0.35 0.025 102 27 5 4 2003 : 14,499 0 3977 0.35 0.025 99 27 > 5 2004 : 14,427 0 3877 0.35 0.025 97 26 4 3 2005) 12/1/15 ,355 0 3781 0.35 0.025 95 25 1G) 3 2006 : 14,283 0 3686 0.35 0.01 37 10 2 1 2007/21) 16,211 0 3649 0.35 0.01 36 10 2 1 2008 : 14,140 0 3613 0.35 0.01 36 10 2 1 2009 : 14,070 0 3577 0.35 0.01 36 10 2 1 2010 ie tseoee! 0 3541 0.35 0.01 35 10 2 1 TOTAL RETROFITS 10300 POTENTIAL; ACTUAL= 6795 1831 309 238 NOTES:* Of 8000 units currently not served we assume 2000 are vacation homes in the Big Lake area and elsewhere. Therefore 6000 are potentially available for service and conversion to gas. In this case we assume the LIDs get gas. This is about 1800 units. In addition gas is extended in the early 1990s to Butte, Big Lake, and Meadow Lakes. Meadow Lakes. This adds 3000 to the pool. 500 more are added in 2000. * We assume 9000 units currently have access to gas. There are 3400 residential ENSTAR accounts and, assuming 600 units reported in the commercial sector, this leaves 5000 units in the retrofit pool. * Electricity percentage from the residential survey. * Total retrofits by 2010 result in the percentage of gas-accessible buildings which are actually served as: 75% * Rate of retrofits from personal communications with utilities and others. * Distribution of retrofits: 0.77 single family 0.13 multifamily 0.1 mobile home TEE ITE AEDT LE IL EEL ELLE ELLER PS EEE ELITE TED EA, SCAB: SAAS cc AT OT RTT Table G.6: Projected Gas Conversions, MatSu, High Penetration Appendix H Response to Public Comments This appendix presents the comments of and responses to the following parties: Matanuska Electric Association Chugach Electric Association U.S. Department of Energy, Alaska Power Administration Jerry Jackson Usibelli Coal Mine, Inc. Eric Myers Senate Advisory Council AMaAurAS This Page Intentionally Left Blank Martanuska Ecectric Association, Inc. P.O. BOX 2929 TELEPHONE PALMER, ALASKA 99645-2929 (907) 745-3231 December 13, 1988 Alaska Power Authority P. O. Box 190869 Anchorage, Alaska 99519-0869 Attention: Richard Emerman Gentlemen: SUBJECT: Comments on 19 November 1988 Draft Report Titled "Forecast of Electricity Demand in the Alaska Railbelt Region 1988 - 2010" Thank you for the opportunity to review this draft document and provide our comments and concerns. This is clearly a professional document which is well researched and written. The involved personnel should be credited for their dedicated efforts to provide this first of a kind "end use" forecast in Alaska. With the exception of the Matanuska Valley ("MAT") forecast, we find this fore- cast to be reasonably consistent with historic utility forecasts using Mi) "econometric" techniques. In the case of the MAT forecast, overstatement of certain critical load assumptions causes the equations to forecast a 20% decrease in residential electrical consumption between 1990 and 1995. Such a drastic reduction is not supported by previous utility forecasts or the sup- porting information used by this study. We believe the major problem occurs from overstatement of the number of MAT residences converting to natural gas heating. This coupled with overstated sup- porting assumptions such as all-electric home electric energy consumption and water heater conversions provides this drastic result. Specifically, the following critical assumptions should be re-examined for validity in the MAT and revised as appropriate. A. Conversion to Natural Gas (4.1.3) The study utilizes a "high" estimate of electric heat (3) buildings to natural gas which is speculative and not sup- ported by MEA or Study research. Page 4-5 of the Study pre- sents a "base" case and a "high" case. However, for unstated reasons, ISER elected to base all forecasts on the "high" case which significantly distorts the long range forecast. The effect of this is reviewed on page 4-37 titled "Sensitivity to Gas Penetration," but again, no changes were made to mitigate this unsupported assumption. This assump- m Le) ALASKA'S FIRST REC—iWCORPORATED 1941—ENERGIZ © f= nS) Alaska Power Authority Page 2 December 13, 1988 (45) tion also has the effect of lowering MAT average household electric consumption by roughly 30% over the 5 year period between 1990 and 1995. Final Residential’ EUI Values (2.4) The residential EUI units, as defined in the Study, represent the electrical KWHs of energy consumed annually by a typical end use type. You will note from pages 2-17 thru page 2-20 of the Study that MAT heat customers are assumed to use significantly more energy than any other region including Fairbanks where the heating requirements are significant. Paragraph three of page 2-14 indicates these "numbers were adjusted up" but no reasoning or justification provided. This assumption appears unrealistic. The MAT heat data should resemble that of an Anchorage household. Again, the effect of this value on the Study is to overstate the amount of reduced electrical consumption which will occur for each natural gas conversion. Electric Heat Share Inputs (2.2.4) Page 1-6 of the Study defines this as "market share is the fraction of the total market for a given end use that is actually served by a given fuel type." Page 2-9 of the Study indicates a survey finding that 47% of the MAT multi-family dwellings are electrically heated. This is based upon a sur- vey of eleven MAT residences, which we find to be questionable. Again, the effect of this assumption is to over-estimate the reduced electrical energy requirements of conversion to natural gas heating. Water Heat (2.3.1) Page 2-11 (Part 2.3.1) of the Study states "we assumed that all electrically heated homes have electric water heaters, and that all such water heaters are converted when the heating system is converted." The paragraph goes on to acknowledge that this assumption is incorrect as overstating the degree of actual water heater conversion but the Study does not correct the problem. Again, the effect is to overstate the reduced electrical consumption occurring from natural gas heat conversion. Alaska Power Authority Page 3 December 13, 1988 Electric Price Assumptions Section 4.2, "Residential Sales" (Page 4-6) states “utilization falls by approximately 5 percent as electric prices increase." As previously indicated by letter, MEA does not agree with the projected railbelt energy prices for 4 many valid reasons. We cannot find a basis stated to assume a 5% decrease in electric utilization. Moreover, the effect of incorrect pricing is to drive down the electric utiliza- tion which will in turn increase prices. This "death spiral" set of assumptions may have indirectly assisted in understa- tement of future MAT electrical consumption. F. Peak Load (4.8) The peak load analysis is an extremely weak part of the Study. This is unfortunate since peak load requirements are the major historic cost contributor to electric rates. The peak load values are based upon historic utility load factors (Nig) na not directly upon "end use" characteristics and are accordingly more of a historic forecast than an "end use" forecast. Since historic load factors are estimated to improve over time due to technology improvements, the peak load may be considered conservative (high). G. Comparison with Econometric Results (4.9) This section compares the forecast to existing ML&P and CEA econometric formulas. Although good correlation is apparently found, the Study does not provide sufficient discussion or data for a credible independent review. For instance, the Study projects an "end use" Anchorage electri- (Mil) city projection but compares this to ML&P and CEA future loads. This is difficult since the Anchorage area is served by CEA, MEA, and ML&P. It would be valuable to understand how the forecast for Anchorage was sorted by utility and why such techniques were not used in the final Study results. In any case, no attempt is made to compare current MEA forecasts with the MAT "end use" forecast since they apparently do not agree. As in any forecast, the final results are most important. End use forecasting is new to Alaska, APA, ISER, and also the United States in general. When this Gidpew technology produces results which differ significantly from the historic norm, the results should be scrutinized until a valid basis for the difference is found. Again, we cannot find such a basis to support the extremely low pro- jections of MAT loads. Alaska Power Authority Page 4 December 13, 1988 As a final point, we question the APA decision to utilize this "end use" fore- Gidcast as a basis for the critical railbelt intertie feasibility analysis since this technology and the contractor are "new" and Alaska has minimum experience with this form of forecasting. We would be comfortable with this decision if the Study results more closely resembled recent industry forecasts. Again, thank you for this opportunity. Sincerely, ames F. Palin General Manager MY :BB 358-1.1209.1 to .3 cc: Dave Hutchins, ARECA M1 M2 M3 M4 Response to Matanuska Electric Association The comment mentions a "20% decrease in residential electric consumption between 1990 and 1995." According to Appendix C, pages C-31 through C-35, we project the following percentage changes in residential electric sales: e Low Case: -12.5% e Middle Case: -10.3% ° High Case: -8.4% Evidently, then, the comment refers to usage per customer, for which we do forecast a 17.8% decline in the Middle case. This decline may not be completely consistent with previous utility forecasts, but it is entirely consistent with actual utility data, particularly MEA’s data. MEA’s historical data (1987 PRS, page III-6) show that during the 3 year period between 1983 and 1986, MEA use per residential customer dropped 24% (from 14224 to 10804). Similarly, between 1980 and 1985, residential use per customer in GVEA’s service territory dropped from 9767 kWh to 8103, a decline of 17%. Both of these declines reflect the shedding of electric heat load from the system. The ISER forecast of residential use per customer is also consistent with MEA’s own 1987 forecast. Over the period 1987-2005, MEA’s 1987 PRS forecasts a decline of 21.3% in residential use per customer for its service territory which includes Eagle River. The ISER Middle case forecasts a decline of 33% for the MatSu Borough. In comparing these projections it must be kept in mind that the ISER forecast covers the MatSu valley only, where the potential for significant gas conversion activity is greater than for the entire MEA service territory, which includes Eagle River. These concerns are addressed below on a point by point basis. The ISER HIGH conversion scenario is quite similar to MEA’s assumptions about future gas conversions: MEA’s 1987 PRS (p. II-10) projects a cumulative total of 2215 conversions through 2005. The ISER HIGH conversion scenario projects 2378 through 2010. We have added Appendix G to this report to present our calculations of gas conversion activity. The ISER estimate of "HIGH" natural gas penetration is based on detailed review of platting maps to identify potential pools of conversion candidates, coupled with specific assumptions about gas system expansion laid out in section 4.1 of the draft report. (These assumptions have been moved to section 2.2.5 of the final report). Additional text has been added in chapters 1 and 4 to clarify this issue. It is not true that "ISER elected to base all forecasts on the "high" case which significantly distorts the long range forecast." It is true that all three representative cases presented in this report are based on high gas penetration. However, these three cases are pulled from a probability distribution constructed from 72 possible outcomes, of which half are based on "low" gas penetration. The three were chosen to accurately represent H-1 M5 M6 M7 points on the cumulative probability curve. If only the "high" penetration assumption had been used in generating all 72 forecasts, the cases selected to represent the distribution would have been different and would have had a lower load than those presented here. The assertion of a 30% decline is not consistent with any of the numbers presented in the report. As discussed in M1 above, we project a decline of only 17.8% (15,430 to 12,687) in residential use per customer during this time period, based on the MIDDLE case which includes High gas penetration. Furthermore, using table 4.13 and Appendix C it is easy to calculate that if the High gas penetration assumption is replaced by Base penetration, we still project a decline of 11.6%. Therefore the decline attributable to the difference between High gas penetration and Base gas penetration is only 6.2%. The first paragraph of page 2-14 states that "we treated the EUI values as the main calibration levers of the residential model. The numbers were developed ... with the twin goals of (1) maintaining a reasonable foundation in engineering and end use data and (2) accurately reproducing control data on total sales by region." Every end use model must be explicitly calibrated to reproduce control totals. In the case of the MAT EUI numbers, initial estimates of EUIs produced a calculated total significantly lower than the control total. After calibration adjustments to the EUIs, calculated sales were still 10 GWh lower than controls, as shown at the top of table 2.14. This discrepancy was dealt with by a final calibration adjustment described in section 2.8. It is precisely because of the gas conversion issue that this final adjustment was made directly to total sales rather than being "loaded" onto the heating EUI. Electric Heat EUIs were used as calibration levers because they are among the most uncertain end use consumption levels. If overall calibration is to be achieved, but heat EUIs must be held at some arbitrary level, then either market shares or other EUIs must be adjusted. Other EUIs would have to be adjusted quite significantly to achieve overall calibration, in many cases well beyond the bounds of reasonableness. And an adjustment to heat market share would increase the potential for gas conversion. Finally, the difference between Anchorage EUIs and MatSu EUIs is not quantitatively significant for intertie analysis. If we were to replace the MatSu heating EUIs used with the Anchorage values, the resulting difference would be at most 6.4 GWh in 1995. This change would amount to a .4% (four tenths of one percent) upward revision for the Anchorage load center forecast used in intertie modeling. While a larger sample size would be desirable, the survey data used is nonetheless unbiased. Lacking any additional data on the subject, we can assert with equal vigor that the share may be too high, or that it may be too low. In any case, the quantitative significance of this question is also minor, due to the small number of multifamily homes (309) projected for conversion through 2010. If we were to cut H-2 M8 M9 M10 Mil the heat share in half, to 24%, the effect would be at most a .1% (one tenth of one percent) upward revision to the 1995 Anchorage load center forecast. As we state in the sentence following the text to which this comment refers, we did attempt to compensate for this problem by avoiding any adjustment to the water heater replacement share, in spite of the clear indication from the end use survey that the replacement share is lower than the average share. In response to this comment, we looked further into the issue. The following calculations suggest that we have in fact substantially overcompensated for the problem mentioned. The survey data presented in table 2.7 suggest that, were gas conversions not an issue, the electric share of all replacement water heaters should be set at 16%, rather than the current stock average value of 50%. (There are (4+8)=12 replacements planned by electric water heat owners, which constitute 50% of the market. Hence there are approximately ((4+8)/.5)=24 total water heater replacements planned in next 3 years in the sample. The electric replacement share is thus calculated to be 4/24 = 16%) Instead, we left the share at 50%. We project about 2000 gas heat conversions over the study period. We can also expect the entire existing water heater stock to turn over at least once during this period, resulting in at least 8,000 water heater replacements among people not converting to gas. Our failure to adjust the replacement share from 50% to 16% therefore results in (.50 -.16)*8,000 = 2,720 projected electric water heaters. Therefore, even if everyone converting to gas heat should leave their water heater electric, we will still have overcompensated for this problem by 720 water heaters, or about 3.6 GwWh. Under realistic assumptions, we have overcompensated by at least 2,000 water heaters, or about 10 GWh. The decrease in utilization is based on national econometric estimates of utilization elasticities. The price increases assumed for this study are based on negotiated gas contracts and the effects of bringing Bradley Lake on line, as well as a detailed generation expansion plan provided by CEA. Furthermore, we do not see how the flat demographic forecast used in this study provides any basis for population- related or density-related changes in the cost of distributing the power. There is no implicit death spiral in this study since price is not endogenous. We agree that the peak load analysis is a relatively weak part of this study. The first paragraph of section 4.8 is relevant to this issue. As is stated in that section, we do not feel that the data exist to assign valid end-use specific load forecasts. Nor can these data be obtained without metered end use load data. In making the comparisons reported in section 4.9, we did not attempt to disaggregate the end use forecasts by Anchorage utility due to lack of disaggregated end use data such as floorstock. We simply compared growth rates and price elasticities. We regret not having the resources to perform comprehensive utility-by-utility comparisons of our projections with econometric results. (See also M1 above, which “ H-3 M12 M13 does make some comparisons with MEAs work.) The purpose of the comparisons was to assess the plausibility for our results using one of many available benchmarks. The choice of the Anchorage econometric forecasts for comparison was not based on any pre-screening in search of favorable conclusions. It did seem logical, however, to focus on the largest load center. End use forecasting was first performed in Alaska in 1980, for the APA, by ISER. The ISER 1980 low case forecast for 1987 proved to be 1% below actual 1987 sales, while the middle case projection proved to be 8% high. End use methods were also used in part by Battelle in 1983 and 1985 for the Susitna project license application, and by Burns and McDonnell for CEA’s 1984 PRS. In the latter study, Burns & McDonnell used a rudimentary model of appliance turnover which led them to forecast a 20% drop in residential use per customer between 1987 and 1998. (Table V-4). As we have pointed out in M1 above, the ISER forecast for residential use per customer in the MatSu region is in fact not extremely different from MEA’s own projections. The same observation holds true for total residential sales: MEA projects a 9.3% decline in residential sales between 1987 and 1998; ISER projects a 20% decline for the portion of the MEA territory where all the gas conversion activity is concentrated. Turning to total energy requirements, MEA projects a 4% decline between 1987 and 1997; ISER projects a 12.7% decline in the MatSu borough and a 9.8% decline when Eagle River sales are roughly allocated to MEAs territory. Several final observations are relevant to this comment. First, we reiterate the point made in M1 and M12 that our results do not differ drastically from MEA’s own recent forecasts. Second, we acknowledge that the presence of the High gas conversion scenario in all three representative cases is "most noticeable" in the MEA forecast. The three representative cases were selected based on total Railbelt load. The cumulative probability distribution for the MatSu region might well dictate the use of different representative cases with higher loads if the exercise were confined to that one region. It wasn’t, and the cases had to be consistent across regions. © a ELECTRIC ASSOCIATION. INC. 5601 MINNESOTA DRIVE * PO BOX 196300 * ANCHORAGE. ALASKA 99519-6300 * PHONE 907-563-7494 / FACSIMILE: ee oe, = 907:562:0027 eyes December 22, 1988 eH Alaska Power Authority P.O. Box 190869 Anchorage, Alaska 99519-0869 Attention: Mr. Richard Emerman Subject: Forecast of Electricity Demand in the Alaska Railbelt Region: 1988 - 2010 Dear Mr. Emerman: Chugach Electric Association has reviewed the load forecasts prepared by ISER and offers the following comments. It should be noted that the technical memorandum allocating Eagle River sales to the Matsu region (described in footnote 2, page 1-1 of the report) was not made available to us until late yesterday, four weeks after the draft report was released and two days before the deadline the Power Authority has imposed on comments. We have not had sufficient time to review this information and therefore reserve the right to revise or add further comments in the future. ie Because the memorandum was not available, comparison of the load forecast to that of Chugach's was somewhat difficult. However, the ISER forecast for the Anchor- age/Matsu/Kenai region compared with a combined CD Chugach/AML&P forecast for the same regions shows that the ISER forecast is approximately 10-20 percent lower in the year 2010 depending on the case analyzed. Most of this discrepancy appears to be in the commer- cial/industrial sector. Zs The report assumes improvement of commercial appliance operating efficiencies of one percent per year (60 percent probability). Such an amount compounded over the study period equals nearly 25 percent efficiency improvement. We would caution against using such an aggressive assumption without further support. Sr ISER has done a good job in combining data from several @ different sources into this document. While we may Mr. Richard Emerman 2 December 22, 1988 look at the assumptions in a slightly different manner, the report will provide us with data that will be beneficial to us in the future. Again, we must reserve the right to modify or add-to our above comments until we have had adequate time to review the technical memorandum. =n 7@y Thomas A. Lovas, Manager Planning and Rates TAL/MDH/ts 894.TAL File 801.1 RF C2 C3 C4 Response to Chugach Electric Association Our forecast generally shows higher residential growth rates and lower commercial growth rates than those projected by CEA in their 1987 Power Requirements Study (PRS). Without knowing the specific forecasts to which CEA is referring to it is impossible to shed further light on the causes of these differences. In section 4.9 of this report, we have made some comparisons with econometric results from CEA’s 1987 PRS. In any event, we have tried to lay out our assumptions and modeling methodology in support of our results. As the comment implies, we only assume the ~1% annual improvement with 60% probability. Section 3.4 suggests some of the means by which efficiency gains of this magnitude are routinely being achieved in today’s market. As pointed out in section 4.1, the ~1% trend rate of efficiency improvement subsumes the legally mandated gains from federal ballast standards, which alone will provide ~5% overall reduction in Electric Intensity (EI). Further, as we report in section 3.3.1, the commercial end use survey data show a clear downward trend of 1.35% per year in overall building EI for new buildings since 1970. In light of this data, and our exclusion of EMCS systems from explicit consideration, we feel that the assumption and its associated probability are adequately supported. Comment noted with appreciation. We have tried throughout this study to emphasize primary data collection. H-5 This Page Intentionally Left Blank Department Of Energy Alaska Power Administration P.O. Box 020050 Juneau, Alaska 99802-0050 December 16, 1988 Alaska Power Authority ATTN: Richard Emerman P.O. Box 190869 Anchorage, AK 99519-0869 Dear Mr. Emerman: Thank you for sending down the draft of ISER’s new electrical forecasts. I have just two general comments: @ ° VD bP The study relies on just one year of historical data to correlate end use with electric energy sales. In my mind, this does not establish a reasonable confidence level in the correlations, and it introduces a major risk of error because of any peculiarities that may have been reflected in the base year data. The high, mid-range, and low forecasts are essentially identical for the first several years with remarkably small divergence thereafter. I suspect this reflects a set of study assumptions which, taken as a _ group, understate the range of likely futures. so Pie Ake hell Caen 5 aa aainiatewted This Page Intentionally Left Blank Al Response to U.S. Department of Energy, Alaska Power Administration We agree that it would be desirable to use a larger base of historical data to calibrate the end use models. However, no such data is available. Unlike an aggregate econometric load forecast, which requires a small number of historical series to estimate a set of regression equations, our end use models require base year data at the end use level. Specifically, we must know the level of residential customers by housing type and the level of commercial floorstock by building type for each region, as well as the market shares and average EUI levels for each commercial and residential end use in each of these building types. Most of our effort in this study went to collecting that data for 1987. Our data collection efforts were the first ever for the commercial sector and the first in several years for the residential sector. If and when more end use data becomes available for the Railbelt, it will be possible to explore the accuracy of both the model’s relationships and the exogenous assumptions used to drive it. To guard against excessive narrowness in the range of assumptions, the APA has included several sensitivity cases in its final evaluation of intertie projects. The three representative cases define a range of variation equal to 23% of the MIDDLE case level. These cases encompass only the variation from the 16th to the 84th percentiles of the cumulative distribution of outcomes. If the entire distribution is considered, however, the range of variation increases to 38%. This 38% variation is indicative of the range of study assumptions. There are also sound economic reasons why the variation in electric load forecasts may be less than the variation in demographic forecasts. Both Railbelt economic growth and the retail price of Railbelt electricity are largely determined by the world price of crude oil. It is likely, therefore, that under conditions of overall economic growth, load growth will be dampened by a rising electric price, and vice versa. Under the modeling framework employed here, electric load growth is still largely determined by employment and household growth. The range of demographic assumptions is taken directly from a previous report by Goldsmith which in turn employed assumptions and probabilities developed with the APA and its board of directors. The range of electricity prices was developed as a function of the range of projected crude oil prices which were also determined in a previous study effort. The range of assumptions about future natural gas availability is based on a detailed assessment of expansion scenarios, described in section 2.2.5. The range of assumptions for technological change was criticized as being too broad in comment C3. Finally, the range of future consumer discount rates, while not very great, was shown to have almost no effect on projected electric load. H - 6 December 19, 1988 COMMENTS ON FORECAST OF ELECTRICITY DEMAND IN THE ALASKA RAILBELT REGION: 1988-2010 Jerry Jackson Jerry Jackson and Associates P. O. Box 2466 Chapel Hill, North Carolina 27515 (919) 968-8454 INTRODUCTION I congratulate the Alaska Power Authority and their contractors on completing such an ambitious project. I believe that this effort provides a good baseline to address electricity demand issues in the Alaska Railbelt Region. As is the case with similar efforts conducted in other parts of the country, some areas stand out as reflecting weaker links in the chain of analysis than other areas. The objective of this paper is to focus on what I consider to be these weaker links in the hopes of fr) benefitting future Railbelt energy modeling and forecasting efforts. Because of my own time constraints, I focus my comments on commercial sector energy forecasting issues and on COMMEND, the commercial sector end use model used to develop these estimates. To some extent, my comments relate to issues which undoubtedly arose because of time and budget constraints which had to be met in this project. Even so, I believe that it is important to consider the implications of these issues for the forecasts. Since little detail is provided on technical data development and G2) model ing issues addressed in the project, my comments are also summary in nature. My comments are grouped below for five areas. DATA DEVELOPMENT Data development is, of course, one of the primary obstacles to developing a reliable commercial sector end use model. Both mail (33) surveys and onsite audits have been used in the past to develop reliable data on commercial establishments. This study attempted to utilize both approaches in developing characterizations of commercial buildings in the four regions. While this hybrid approach is an excellent idea, the number of usable responses (85 for the a and 196 for the mail survey) is so small that the resulting (NT SA COMMEND model parameters undoubtedly reflect relatively large standard errors. For instance the large office category, which is probably the most important building type, is represented by 6 onsite surveys and $ mail survey responses. The small sample sizes make inferences drawn on the relationship between old and new building EUI's tenuous, especially in light of the fact that New England and New York audit data shows that new buildings typically use more electricity than existing buildings even when saturation impacts are removed. In particular, new buildings typically have significantly higher lighting and miscellaneous EUI's than the existing stock. This initial data base can, of course, be supplemented with future data collection. Based on my analysis of audit and mail survey data, I suggest an onsite effort of at least 250 buildings and/or a mail survey response of at least 1000. EQUIPMENT EFFICIENCY DETERMINATION Efficiency improvements over the forecast period are an important determinate of future electricity use trends. Version 3.0 of COMMEND uses technical tradeoff elasticity measures which reflect the "percentage decrease in EUI from ai percent increase in capital cost." I used this conceptual representation in 1976 in the original ORNL commercial end use model; a variety of problems led to its replacement in the following year. Some of these problems are listed below. 1. Diminishing marginal returns mean that efficiency changes accompanying a given investment decline over the available options. If an elasticity is used, it should be a variable elasticity rather than a constant elasticity to reflect this physical phenomenon. 2. An elasticity concept ignores the discrete nature of market choices and the impacts of thresholds on decision-maker choices. 3. The smooth curve traced out by a constant or variable elasticity approach does not correspond with the curve one derives when connecting points represented by real technologies available in the market. Another weakness of this analysis is that all tradeoff elasticities reflect assumed rather than estimated values. In the case of lighting, which is by far the most important in this modeling application, technology data are used to compute elasticities of 5.23 and 7.25; however, an elasticity of 1.5 is the assumed value. a Gi) Thus, it appears that the assumed values do not bear a close correspondence to the physical relationships which they purport to represent. The report states that "COMMEND does not simply minimize life-cycle cost since empirical observation has for a long time shown that electric consumers do not fit into life-cycle cost-minimizing straightjackets. In a nutshell, equipment choice equations are calibrated to reflect initial observed EUI levels and future changes in EUI levels are computed based on changes in LCC from the base level and and assumed elasticity of responsiveness of life-cycle cost.” If the LCC approach does not work to estimate initial EUI's and cannot be used without some sort of ad hoc "responsiveness" adjustment and if the tradeoff elasticities do not reflect actual efficiency-cost relationships why is this LCC approach used? The LCC approach has been so constrained in this application to meet reality that it should either be replaced with a simpler judgmentally-based relationship (it is currently a complicated judgmentally-based relationship) or with judgmentally derived efficiency assumptions. My recommendation is to replace it with judgmentally-based assumptions. Several surveys of equipment choice in the market exist to guide such an effort. Another alternative is to develop a model component based on a more realistic representation of the way that commercial decision-makers select equipment efficiency. ESTIMATION OF UTILIZATION RELATIONSHIPS The report does not describe the development of the utilization relationship(s) (reflecting intensity of equipment use) in the commercial model; presumably, that estimation was not part of this project. These relationships are an important determinate of future energy demand and should be estimated statistically with historical data rather than assumed. VALIDATION One of the most revealing tests of model reliability is to conduct a forecast over the 1972 - 1987 historical time period. Relying on a model without validating it with ex post forecasting is like buying acar without driving it. These validation results should be presented along with other model documentation. 4 A lA CALIBRATION All end use models must be calibrated. The nature of this calibration effort can have a significant impact on model forecasts. Adjustment factors of 1.23 for Anchorage, 1.27 for Fairbanks and 1.15 for MatSu are substantial. Application of these factors to the total commercial forecasts handles this problem by ignoring it. The appropriate resolution of this discrepancy should be part of the commercial modeling effort. Since lighting and miscellaneous EUI's reported in this study are already on the high side of EUI's developed from other audits, they are not likely cause of the discrepancies. Estimation of floor space by building type using employment data and available floor space-per-employee estimates, number of pupils and estimates of pupils per-square-foot and other data would shed light on the accuracy of floor space estimates. To the extent that the greater control totals are a result of small manufacturing establishments or master-metered apartments, other building categories should be established with the markedly different EUI's which accompany these buildings. Ji J2 J3 J4 J5 J6 J7 Responses to Jerry Jackson We appreciate Jerry Jackson’s thoughtful comments. We generally share his concerns and echo his observation that many important areas of analysis were subject to binding time and resource constraints. We have tried to present sufficient detail for a wide audience to judge the overall integrity of the forecasting effort while maintaining the readability of the report. We welcome further discussions about technical matters and would be happy to provide additional details to anyone who requests them. We agree that the commercial sample sizes were less than ideal. Our workplan made a fundamental commitment of resources to on-site data collection. Once this commitment was made, we looked at as many buildings as we could. The workplan called for visits to about 130 buildings, and we met that goal. The reference to the number of useable responses is not completely accurate. He apparently refers to Table 3.3 "Measured Electric Energy Intensities," from which he concludes that only 85 on-site and 196 mail responses were useable. In fact, 135 on-site responses and over 470 mail responses were used to develop market shares, discount rates, and technology assessments such as the distribution of lighting technology. The lower figures apply only to the EI estimates and reflect the aggressive rejection of outliers from the electricity consumption data which we were able to collect from seven different utilities. Further exploration of these suspicious billing histories could probably increase the number of responses available for EI estimation at a small marginal cost. A glance at Figures 3.2 through 3.6 shows that the large office space type makes up only 11% of the Anchorage floorstock, and is virtually nonexistent in the Fairbanks, Kenai, and MatSu regions. This category was in fact oversampled relative to our energy-weighted proportional sampling plan, because we looked carefully at one of Anchorage’s few buildings over 10 stories. We agree that the new vs. average EUI relationship would benefit from more data. Although there has been almost zero commercial office construction in the Railbelt since 1985, the data for the four major office buildings built since 1982 clearly support the hypothesis that new lighting EUIs are lower than average. In addressing this issue we supplemented our official on-site survey data with personal communication with several office building managers, made informal visits to several buildings, and talked with the designers of some of the very few projects now on the drawing boards. We recognize that there are significant limitations to the technology curve approach. The COMMEND model is currently being adapted to a discrete technology framework. The use of a constant tradeoff elasticity seems to us to be fully consistent with H-7 J8 J9 diminishing marginal returns (measured in reduced EUI) to capital. To see this, one need only note that the technology curves are hyperbolas, which are concave functions. The first few dollars of additional investment yield more saved energy than the next few dollars. The second difficulty mentioned is a valid theoretical objection to the smooth curve approach. The same type of objection can be raised about consumer demand for automobile horsepower or any other product for which individual decisionmakers face discrete opportunity sets. This problem has not stopped economists from estimating and employing continuous demand curves for such goods or attributes. They can do this because although each decisionmaker may face a discrete opportunity set, the demand curve for all consumers taken together tends to approach a smooth curve if the total choice set and/or the number of decisionmakers is large. We think a good case can be made that these conditions are met in the commercial energy equipment market. The third difficulty mentioned is an empirical one which undoubtedly has some merit. Discrete choice models have difficulties of their own. Foremost among them is the task of defining a small set of technologies which can somehow represent the wide range of equipment actually deployed. For lighting, this task is less difficult, and we conducted some detailed lighting calculations to support the COMMEND results. The calculation presented was illustrative and assumed a particular set of circumstances favoring a higher elasticity value, including relatively short component lifetimes. We chose the lower value to produce a conservative estimate of technical potential consistent with the experience of other COMMEND users. We disagree with the characterization of the equipment choice algorithm in COMMEND as an "LCC approach" which has been "constrained." The probabalistic choice model used in COMMEND is described at some length in EPRI EM-5356, Models of Commercial Sector Equipment and Fuel Choice Decisions for the COMMEND Code. The model assumes that life cycle cost is one factor driving decisions. The importance of other factors such as convenience, reliability, habit, cash flow constraints, and limited information, can be captured in the two additional parameters which determine the choice probabilities. Rather than constraining the LCC algorithm, the probabilistic choice method introduces additional degrees of freedom. The relationship between life-cycle cost and equipment choice is not "complicated", but actually quite simple, involving a logit function. (What is arguably complicated is the use of simulated efficiency elasticities to reduce the computational and data development burden). The "ad hoc "responsiveness" adjustment" to which Jackson refers is the share elasticity. While share elasticities are currently estimated from judgment, they are estimable from data and efforts to do so are underway. In fact, the chief advantages of the probabilistic choice framework are its simplicity, estimability, and ability to model choices among discrete technologies. H-8 J10 Jil J12 We agree that if resources are available, utilization rates should be investigated with Alaska data. However, direct estimation of utilization would seem to require a data set containing either (1) good information on equipment EUIs for each observation, or (2) enough short-term variation in price to permit estimation over a time period during which EUIs could be assumed to be constant. We doubt that either condition can be satisfied with Alaska data. Indirect estimation of utilization elasticities as a residual similarly seemingly requires data on evolving EUIs in order to estimate efficiency elasticities. Certainly validation of any model is desireable. However, we are quite unsure how one would proceed with the ex-post forecasting exercise suggested here. At the very least, one would need a relatively complete set of end use data for 1972. To truly test the reliability of the model, as opposed to the exogenous inputs, one would also need data on floorstock by building type over the forecast period. As additional end use data become available for the Railbelt, we will gain the opportunity to conduct ex post forecasting exercises. In the meantime, it should be remembered that the COMMEND model is not new and is not mysterious. Unlike large econometric models, it has no simultaneous equation blocks and proceeds sequentially from an exogenous estimate of floorstock and a fresh estimate of new building EUIs every year. It has been in use for several years by many utilities around the United States. The complete resolution of discrepancies between calculated results and control totals is certainly a worthy goal. We attempted this resolution to the extent permitted by the combination of time, available data, and available data-gathering resources. For example, we looked very carefully at large customer billing histories and large customer floorspace records for the Kenai Peninsula and eventually moved a small number of establishments out of the commercial floorstock and into the industrial floorstock. This adjustment eliminated all of the initial discrepancy for that region. We disagree with the assertion that a direct adjustment to sales ignores the calibration problem. A direct adjustment to sales is mathematically equivalent to a direct adjustment to floorstock across all building types. It is also equivalent to assuming that there is an increment of light industry consumption which grows with the rest of the economy and follows overall trends in commercial sector EUIs. We did, of course, check the validity of our floorstock inventories against ratio estimates, as reported in Table 3.1. This check was consistent with the hypothesis that both the commercial floorstock was undercounted in all three northern regions. However, we had to use national estimates of the ratios since there are no estimates of alaska floorspace utilization ratios independent of our own data developed for this study. For some space types, our floorspace data quality was superb; for schools and colleges, we used a complete inventory of every square foot of space. As the report states, we know that our floorstock numbers are low because they come from a census rather than a sample. We agree with the comment that our EUIs for lights and miscellaneous seemed to be on the high side. The question then becomes: are there enough data to reject the H-9 hypothesis that a low floorstock is the cause of the problem? In answering this question we focused on two sources of load in the commercial control totals: miscellaneous light industry and residential use off commercial class meters. We looked at individual customer data for the MatSu and Fairbanks regions. We were unable to analyze the CEA tape due to company policy, but CEA staff prepared aggregate data which showed negligible industrial consumption beyond that already accounted for. We called several apartments and determined that some master- metering is still in place, but the utilities were unable to tell us how much. The final explanation for the discrepancies observed is dramatic underestimation of average market share for electric heat and water heat. Full-scale modeling of the end use structure of gas demand would shed light on this hypothesis, but was not possible within our constraints. Furthermore, a market share adjustment is equivalent to a sales adjustment unless the average share changes significantly over the forecast period. This would indeed happen if we were to adjust the initial average heat share dramatically upward, since the average share would then drop significantly and the resulting forecast would be significantly lower than under the direct sales adjustment. We took the more conservative approach. Our decision to make adjustments directly to sales was not taken cavalierly. Under all of the plausible scenarios which would explain the shortfall in calculated sales, the resulting adjustments were essentially equivalent to a direct adjustment, given the data at hand. § @ @® USIBELLI COAL MINE, INC. MARKETING 2173 University Avenue So. Suite 101 Fairbanks, Alaska 99709 (907) 479-2630 FAX 479-2793 December 20, 1988 Alaska Power Authority P.O. Box 190689 Anchorage, Alaska 99508 Attn: Richard Emerman, Senior Economist Re: Comments to draft report "Forecast of Electricity Demand in the Alaska Railbelt Region: 1988-2010" Dear Mr. Emerman; The comments herein will be limited to the general conclusions and assumptions articulated in the executive summary of the referenced document. Much of the detailed analysis contained’ in the body of the report is outside Usibelli Coal Mine’s (UCM) area of expertise and would be inappropriate for us to address. Generally, the gloomy forecast portrayed in the report is very hard to believe and a rather low appraisal of the resourcefulness of Alaskans. In particular, Table I shows an average load growth of only 0.2 percent annually through the year 2000, a rate which is much too low to be believable. For instance, the industrial demand growth is predicted at only 7 gigawatt-hours through 2000. One small industrial consumer, about half the size of UCM would account for this level of increase. Of the many projects in the railbelt on the drawing boards and the many that are yet to be born before 2000, a very small percentage of success would throw the flat estimate contained in this report way off base. Relative to the assumptions on page ES-4: a) A growth of 0 to 1 percent (say an average of 0.5 percent) will just barely cover the growth through 1992 resulting from the current deployment of additional @3) military personnel in the Fairbanks area. Surely, this will not be the only positive influence in the Railbelt economy for the next 12 years. Any significant economic boost, such as the exploration of ANWR, would result in a much higher growth rate. 2) The assumption that electricity prices will rise @4) "rapidly" at 2 percent is lower than what is known from ; current power and gas sales contracts. A report completed in August 1988 by ICF, Inc. for APA predicts oil prices at $18 per barrel for the projection scenario which APA has assigned a 60 percent OS probability of occurrence. This price is only 93 cents SS over the base oil price of current power sales contracts, which are fuel price driven. Even this modest increase in oil price would result in about a 2 percent annual increase in the cost of electricity. An ——article in the September 29, 1988 issue of the Anchorage Daily News stated that the new gas contracts Jb between Marathon Oil and Chugach Electric would result \ in about a 25 to 30 percent increase in electricity cost by the mid nineties from a phase in of the new gas price. Even with these much higher rates of price increase, southern railbelt power generation cost will / still be some of the cheapest in the nation and I see (7 no reason to expect that power cost will have any effect, unless positive, on economic development in the railbelt. 3) The assumption that the increase in efficiency of major appliances will offset the increase in use of more electric consuming devices may be true for a few households but seems a bit far fetched for the overall picture. Existing and committed intertie systems will improve prospects for delivery of low cost gas generated power to a wider consumer base. It seems more likely that the size of the electricity "fix" required by the average consumer will increase. 4) It is interesting to note that expansion of the gas distribution system is assumed but not the electrical distribution system. The net effect of the assumptions in the report seems to skew the results of the report towards lower growth than what would seem reasonable for responsible energy planning. Admittedly, the railbelt is currently awash in electric \ generating capacity. However, if one carries the projections (016) shea in this report, there will not be a need for increased generating capacity in the railbelt for over 60 years and there will still be a 30 percent reserve capacity in the year 2025. This kind of blissful planning environment seems like a dangerous foundation upon which to build railbelt energy policy. Thank you for the opportunity to comment on this report. I will be anxious to see the outcome of the final version. Sincerely yours; Usibelli Coal Mine, Inc. by Steve W. Denton Ul ey U2 \ t U3 U4 U5 U6 U7 Responses to Usibelli Coal Mine, Inc. The growth rates of total electric consumption are a function of the assumptions used in the forecasting process. We have tried to lay out these assumptions in this and previous reports. Without knowing which assumptions are less "believable" than others, it is difficult to respond to this general comment. As we have shown in section 4.9, our forecast projects higher residential sector growth than the 1987 CEA utility forecast. And our forecast trend in commercial sector use per customer is consistent with that obtained in GVEA’s 1987 forecast. Our industrial forecast is based on consultation with the Railbelt utilities and review of their expectations for industrial growth. As described in Appendix D, our MIDDLE case industrial forecast includes several unspecified activities in all regions of the Railbelt. Some of these activities are not projected to begin until the mid to late 1990s. The Fairbanks Light Infantry Division deployment is fully accounted for in the demographic forecasts supporting these load forecasts. In addition, the following other "positive influences" which contribute to Railbelt demographic growth in some or all scenarios include: ANWR and OCS oil development, TAGS development, Beluga Coal mining, Greens Creek and U.S. Borax mines in Southeast Alaska, unspecified mining, constant increases in agricultural employment, continuing expansion of tourism, expansion of bottomfishing (consistent with the resource base), Navy cruiser "homeporting," and unspecified increases in Federal civilian employment. The 2% figure is an average. This increase is fully consistent with the recently announced contracts signed by Enstar and CEA for the purchase of gas from Marathon oil co. The text of the executive summary has been changed to emphasize that the increase mentioned is in real 1987 dollars, adjusted for inflation. Comparable annual growth rates in nominal dollars would be in the 5 to 7% range. Apparently the comment refers to the "low price school" scenario presented on page ES-3 of ICF 1988, in which the price of oil (again in real dollars) rises from $14 in 1990 to $18 in 2000. The comment then goes on to say that this increase would result in a 2% increase in the price of electricity. This conclusion is consistent with our range of price growth rates. While the details of our price computations differed somewhat from the logic employed here, we seem to arrive at the same result. The comment "Even this modest increase..." implies that there are other major forces besides oil prices which are likely to push electric prices upward. We see no such forces, with the exception of Bradley Lake, which is fully accounted for in our price forecast. The figures quoted in this article are nominal dollar figures. When general inflation at 5% is removed from these increases, they are consistent with our price forecasts. We agree that the cost of electricity will not be a major negative influence on H- 11 U8 u9 U10 industrial location decisions. However, one only has to look at current planning for cogeneration and self-generation projects throughout the Railbelt (see eg, Decision Focus 1989, Chapter 5) to see that increases in electricity price are already having an apparent impact on Utility-supplied electric load, which is the subject of our report. We have nowhere stated, and do not mean to imply, that rising electric prices would be a negative influence on economic growth per se. Rather, rising electric prices can be expected to slow the growth in the demand for electricity, all other factors being equal. The price-responsiveness of electricity consumption is a clearly established economic fact which has been corroborated by several Railbelt utilities. (Sée, eg, CEA 1987 Power Requirements Study.) We have forecast unspecified increases in miscellaneous residential and commercial end uses of ~1.3% per year (see sections 2.6.3 and 3.5.3) . We have also taken account of existing electric market share levels and projected changes in these levels. Of particular importance is our projection of continuing erosion in the market share of electric heat due to expansion of the natural gas delivery system which can always deliver thermal energy which is far less expensive than gas-generated electricity. It takes a very large dose of gadgets and appliances to compensate for the load of one electric heat customer. For example, the loss of load from one single family home conversion from electric to gas heat is ~20,000 kWh per year. In order to replace this load, 200 customers would have to install and fully utilize an additional computer, toaster, and washing machine. The expansion of the gas distribution system is only assumed consistent with ongoing activity, and is assumed only with varying degrees of certainty. We fail to see where there are areas of unserved electric load in the Railbelt which might materially affect the load forecast. We have deliberately employed conservatisms at several points in the analysis which tend to bias the load forecasts upward rather than downward. These conservatisms are noted throughout the body of the report. It is, of course, a policy judgment as to whether an explicit upward or downward bias is appropriate to this analysis. The answer depends on what courses of action are being evaluated and on the costs of both over- and underforecasting. (See, eg, comments of Eric Myers below). As we stated in section 1.4.2, our approach was to err slightly on the high side. We certainly hope that this forecast will not be used as the basis for a 60 year moratorium on new power plants, and that in future years updated forecasts will be prepared which incorporate the latest available data. H - 12 he Cb ie = 2 DES 28-1988 SLASKA POWER AUTHORITY December 22, 1988 Mr. Richard Emerman Senior Economist Aiaska Power Authority PO Box 190689 Anchorage, Alaska 99508 Dear Richard, Please accept this letter in response to the opportunity to comment on tne draft report "Forecast of Electrical Demand in the Alaska Railbelt Region: 1988-2010." In general, the report is comprehensive, well written and represents a pioneering (at least for Alaska) effort to analyze electrical demand with l appropriate consideration and evaluation of end uses. However, as indicated ey by the following comments, several important assumptions may lead to a potentially significant overstatement of future electrical demand for the Railbelt. Equally of concern, the analysis gives little or no consideration to the possibility of “extreme events” (ie, another collapse in oi] prices) that has the potential to make large capital-intensive power investments EZ) economically unattractive. Especially given the state's precarious fiscal condition, it would be imprudent to substantiate the "need" for a large Capital commitment on the basis of a narrow range of uncertainty. cifi mmen 1. The report assumes that with a probability of 60%, oil prices will follow the "Low" scenario outlined by ICF. There is apparently no consideration of oil prices below this level. This seems difficult to justify in light of the past few months when tense OPEC negotiations held the potential for a major price collapse (prices in the range of $5.00 or less was seriously discussed). Effectively assuming a $14 “floor” on prices undermines the overal] credibility of the report. 2. A fundamental problem concerns the concept of so-called “conservatisms” as used in the report (page 1-12). ISER states that the “conservatisms” employed are biases toward higher load projections which is a conventional utility practice. while conventional utility practice can be understood if tne issue being analyzed is additional capacity requirements and the associated risk of a shortfall in power supplies. However, that risk is not relevant in the case of the intertie analysis because there already exists abundant excess Capacity In the Railbelt. Alternatively, the appropriate conservatism should be a standard that recognizes the potential risk of using scarce state resources to overbuild Railbelt power systems to an even greater degree than at present. This kind of risk analysis would employ biases that lower load forecasts. It is interesting that the APA Board apparently adopted such , an approach in light of their assigning a 90% of the total probability to the (ES /"Low" and "Consensus Low” oil price scenarios. The apparent inconsistency “— between the ISER concept of conservatism and the APA Board's concept of conservatism should be resolved. 3. Specific examples of significant upward bias In the load projections that may not be justified include the following: a) wy $ @ e@ the assumption that commercial sector miscellaneous load grows at 3% annually (p. 3-30) and at lesser rates in other sectors appears arbitrary and unsubstantiated (the only discussion of this issue concerns computers while the report suggests that computer consumption may drop, not rise); e the assumption that no residential consumers will purchase appliances more efficient than the minimum federal standards (a “significant conservatism”, p. 2-21) yet data mentioned in the report indicates that average efficiencies in California have been below required minimums; @ the assumption that there will be significant levels of electric heat in new buildings seems at odds with information in the report concerning the market for electric heat; e@ the report appears to rule out the possibility of water heater retrofit activity notwithstanding what appears to be significant potential for this activity; @ the section on “exogenous demand growth" appears completely unsupported... while it is possible that the size of new houses wil] grow at 0.6%/year, a contrary assumption also seems possible... at a minimum, house size should be related to income in some fashion; @ the assumption that miscellaneous end-use demand will grow at 1.3%/year with certainty irrespective of prices and the possibility that future appliances will completely substitute for older ones (eg, microwave ovens). | appreciate the opportunity to comment on the report and look forward to your continuing work on this project. Sincerely, Fae Eric F. Myers 6710 Potter Hghts Anchorage, Alaska 99516 This Page Intentionally Left Blank en a El E2 E3 E4 ES E6 E7 E8 E9 E10 Responses to Eric Myers While the conservatisms employed have undoubtedly increased the load forecast somewhat, we are confident that the increase has not been "significant" compared with the overall level of uncertainty surrounding the forecasts. For example, all of our forecasts were apparently below the level of the combined-Railbelt utility forecasts subsequently used for sensitivity testing in the intertie analysis. See response to comment A2, above. The Division of Policy raised similar concerns in their comments regarding the ICF draft report on oil and fuel prices. In response, the APA has stated that the lowest oil price scenario used to drive the demographic forecast was not considered a boundary case, and that any project found feasible according to the lowest ICF oil price scenario would be further tested under lower boundary oil price assumptions, such as the Alaska Department of Revenue (DOR) "Mid" and "Low" price estimates published in the fall of 1988. See response to E2 above. We believe this point has some merit. However, we feel that the change introduced into the forecasts by our use of the conservatisms mentioned is insignificant compared with the range of uncertainty encompassed by the five critical assumptions which were combined to produce a range of forecasts. The concern expressed in this comment would be of greater importance if it was decided to only produce one or two point estimates of demand, rather than a range of forecasts. See also the response to E2 above. We do not feel that there is any conflict in need of resolution. The APA Board selected the oil price probabilities as their best guess of what the future will hold. There was no attempt to introduce conservatism in that exercise. The assumption regarding miscellaneous equipment growth is consistent with past trends and is substantiated by citation of a Brookhaven National Laboratories Study. The text discusses this issue. We have assumed that there will be some long run demand for electric heat. Our reasoning has been presented in section 3.2.1. Presumably the comment refers to the replacement of electric water heaters with gas units. We do not rule out such water heater retrofits. As indicated in section 2.3.1, we have assumed that every house which converts its space heat system to gas also converts its water heater. The assumed rate of growth in the size of new houses is supported by historical trend, as indicated in section 2.6.3. We agree that tying new house size to income makes theoretical sense. However, this was not possible in the model employed. In H - 13 Ell addition, the link between new house size and average income is more tenuous than that between new house size and new home-buyer income. For example, during the past two years in Anchorage, the only new homes which have gone up have been high priced, custom built, large single family homes. During this period average per capita income has been stagnant. Presumably the income of the new home buyers was not. As stated in section 2.6.3, the assumed increase in miscellaneous residential use is introduced to reflect our fundamental ignorance about new uses of electricity in the home. It also reflects the lack of an explicit link in the model between percapita income and the purchase of unspecified appliances. In addition, we should note that there is significant potential for growth in waterbeds, saunas, jacuzzis, VCRs, and home computers as revealed by our end use survey. Senate Advisory Council PO. Box V State Capitol Juneau, Alaska 99811 Phone: (907) 465-3114 MEMORANDUM TO: Richard Emerman Senior Economist Alaska Power Authority FROM: Kurt S. Ozinich 69 Senior Advisor Senate Advisory Council DATE: December 23, 1988 SUBJECT: Review of Draft Report -- Long-Term Forecast of Electricity Demand in the Railbelt Based on a review of your draft report "Long-Term Forecast of Electricity Demand in the Railbelt" the following comments are pertinent. In view of the historical electric load growth in the Railbelt, I remain concerned at the rather low growth being projected to the year 2010 i.e. 85 percent chance that it will be less than 1.3 percent for the period 1988-2000. - In order for readers to be able to better evaluate this projection in the context of the historical record, I would recommend expanding Figure 1 backwards in time to around 1967. The statement in the second paragraph of section 2.2.4 that assumes that nay. nousenoids will never switch to electric heat when replacing worn out fossil fuel units seems somewhat arbitrary and in view of increasing concern about air quality not very likely. It is fortunate indeed that the Railbelt area has a substantial surplus generation capacity at this time because it would surely be difficult to Gayiustify building new generation facilities on the basis of the draft report forecast. I would presume that the forecast model and methodology is such as to allow for easy updating and check of its validity few years down the road. KSD: bp This Page Intentionally Left Blank om om S1 S2 S3 Responses to Senate Advisory Council ’ Figure I of the Executive Summary and Figure 4.13 have been expanded to show historical data back to 1980. A switch from fossil fuel heat back to electric heat is extremely unlikely, in part because of the asymmetry of conversion costs. A homeowner pondering a switch from electric to fossil trades off the costs of conversion against the substantial benefits from lower heating bills. The homeowner pondering a switch from fossil to electric faces no such tradeoff: she looks at the prospect of sacking her fossil furnace and its associated ductwork, paying for the installation of electric heating equipment, and then being rewarded for her effort by substantially higher long-run heating expense. Air quality is clearly degraded by a switch from direct fossil fuel use to electric heat since the Railbelt will continue to use gas and oil as its marginal generation sources and since 3 times as much fuel is burned in these power plants as would be burned to make the heat directly in the customers home. The reason the Railbelt has so much excess generation capacity at this time is that we have had dramatic downward revisions in overall economic growth. These revisions are reflected in our forecast. We too hope that as time goes on our work will be checked for validity and that any forecast used for planning purposes will be updated using the latest available data and techniques. This Page Intentionally Left Blank Appendix I: Additional Fairbanks Loads: Military Installations and the University of Alaska This appendix contains a research memorandum analyzing potential changes in the amount of Military and University of Alaska Fairbanks load served by the Railbelt utilities. This page intentionally left blank UNIVERSITY OF ALASKA ANCHORAGE 3211 Providence Drive Anchorage, Alaska 99508 (907) 786-7710 INSTITUTE OF SOCIAL AND ECONOMIC RESEARCH Additional Fairbanks Load Impacts: Military Installations and the University of Alaska Date: January 12, 1989 Project Manager: O. Scott Goldsmith Principal Investigator: Alan Mitchell 1. Summary of Conclusions This report investigates the electrical load impacts of potential changes in military power purchase decisions in the Fairbanks area. Eielson Air Force Base, Fort Wainwright, and Fort Greely currently purchase no electricity from civilian Fairbanks utilities (they do trade some). A future scenario that assumes that all military electrical requirements will be purchased from civilian utilities is very unlikely. A more likely scenario, and one that is currently being pursued by GVEA, is the supply of a portion of the military electrical requirements. The primary military generation facilities are cogeneration plants that produce both electricity and heat for space and hot water heating needs. Cogeneration facilities efficiently produce electricity and heat in particular proportions. If electrical needs outstrip this balance, the additional electricity is more costly to produce. It is these electrical needs in excess of the balance point that GVEA is attempting to sell the military. These needs are estimated in this report to be about 41 GWh/year, about 32% of the current military electrical demand, and are projected to grow at about 1.4% per year through 2010. Little data as to the size of the imbalance exists, so the estimates are substantially uncertain. The civilian utilities are estimated to have a 15-20 mill/kWh cost advantage in the supply of the military electricity beyond the balance point. Department of Defense policy supports the purchase of electricity from civilian utilities when it is economical to do so. However, local military personnel do have concerns about the total magnitude of the savings possible and the effects of purchases on the reliability of their power system. Also addressed in this report are potential increases in electricity purchases by the University of Alaska at Fairbanks (UAF). UAF cogenerates electricity and heat for their campus but does purchase approximately 1 GWh/year from GVEA (about 3% of their demand). When the GVEA-UAF contract expires in two years, a more favorable purchase price for UAF would prompt them to purchase more electricity from GVEA and avoid the installation of additional condenser capacity. The additional load would be 1.2 GWh/year occurring in the months of May through July, bringing total purchases to 2.2 GWh/year. This is taken as a most-likely scenario. A more unlikely scenario is that the power purchase price would drop enough so that all of the UAF needs beyond their steam balance point 1 A DIVISION OF THE UNIVERSITY OF ALASKA STATEWIDE SYSTEM OF HIGHER EDUCATION would be supplied by GVEA. This load would amount to at least 9 GWh/year according to the UAF power plant supervisor, Gerald England. However, he assigns a low probability to such favorable purchase economics. The UAF plant produces electricity beyond the balance point more efficiently than the military plants (because of higher inlet turbine pressures). 2. Military Electrical Load 2.1 Characteristics of Military Electrical Load and Generation Table 1 summarizes the electrical load and generation characteristics of the three major military installations in the Fairbanks area. The total electrical consumption for the three bases is currently about 125 gigawatt-hours per year (measured at the bases) with a peak demand of about 26 MW. The Fairbanks military bases currently trade some electricity with GVEA, but no monetary compensation is involved. Total trades during 1988 amounted to less than 0.08 GWh.' Eielson and Fort Wainwright are the primary producers of electricity for the military. Fort Greely purchases most of its electricity needs from Fort Wainwright, with the exception of some peaking power supplied on-site. GVEA is compensated for wheeling the electricity from Fort Wainwright to Fort Greely. The primary generation facilities at Eielson and Fort Wainwright are coal-fired steam turbines. Steam from the generation cycle can be extracted to supply a steam heating distribution system for the military base facilities. This steam distribution supplies the space heating and hot water needs of the facilities. Thus, the generation facilities at Eielson and Fort Wainwright are cogeneration facilities. Steam from coal-fired steam turbines is extracted to supply the heating and hot water needs of military base facilities. 2.2 Factors Affecting Increased Sales to the Military Civilian utilities, primarily GVEA, are interested in selling additional electricity to the Fairbanks military installations. GVEA’s primary objective is to supply that portion of the military’s electrical demand that exceeds the "steam balance point". GVEA and the military assign little to no probability to a scenario where the military will buy all their electricity from GVEA or other civilian utilities.” Both the coal plant at Eielson and the coal plant at Fort Wainwright are cogeneration facilities. The steam that drives the electrical generation turbines can be extracted from the outlet of the turbines and used for space and water heating. If all of the steam passing "Personal communication with Mr. Marvin Riddle of GVEA, December 19, 1988. >This was indicated during communication on December 8, 1988 with the GVEA consultant, Mr. Robert Huffman, who is pursuing military sales agreements. Current and Historical Electrical Load Load Projections Generation Plant Characteristics Eielson AFB 1988 Load was ~53 GWh. Winter Peak = 11.5 MW, Summer Peak = 7 MW. Load has grown 10 - 15% in past few years because of new construction. Military projects a 20% increase in next few years because of construction. ISER projects 1.3%/year growth from then on. Primarily use coal-fired steam turbines. Steam is extracted from the turbines to supply a steam distribution system for base heating and hot water needs. Have backup diesel generators. Have a new 10 MW steam turbine, a new 5 MW turbine, another 5 MW turbine, and two old 2.5 MW turbines. Fort Wainwright 1988 Load was ~55 GWh. Winter Peak = 10.55 MW, Summer Peak = 6.5 MW. Load has grown ~10% during last couple years. ISER projects 25% increase in next few years because of increased construction. ISER projects 1.3%/year growth from then on. Primarily use coal-fired steam turbines. Steam is extracted from the turbines to supply a steam distribution system for base heating and hot water needs. Have 5 turbines: one 5 MW _back-pressure turbine, three 5 MW condensing/extracting turbines, and one 2 MW condensing/ extracting turbines. They generate most of Fort Greely’s electricity and wheel over GVEA lines. Fort Greely 1987 Load was 17.2 GWh. Winter Peak = 4 MW, Summer Peak = 3 MW. Load has been flat in past few years. ISER projects 3%/year decline in load for next four years because of reduction of 100 active duty personnel. ISER projects 1.3%/year growth from then on. Buy most electricity from Wainwright wheeled over GVEA lines. Use diesel generators for back-up and peaking. Have a steam distribution system for heating and hot water needs of base facilities, but supply system from oil boilers. No cogeneration occurs. Army has considered installing a coal cogeneration plant, but project is in long range planning. DA I LOA EI TL TT TE TEE ATE ALN ARLE T LOE LE OEE EE Table 1 - Characteristics of Military Electrical Demand and Generation through the turbines can be extracted and used for heating purposes,’ a large fraction of the energy in the steam is productively used, either for heating or generating electricity. However, if there is steam in excess of heating requirements, this additional steam is not extracted for heating purposes but instead is condensed with a cooling pond. The energy ae of this process is poor relative to the process where the outlet steam is used for eating. When the outlet steam is being fully utilized for heating, the electricity produced is 3, i , : ; f 5 Technical constraints for condensing/extracting turbines require that a maximum of 90% of the steam be extracted for steam heating purposes. effectively being produced at a heat rate of about 5,500 Btus per kilowatt-hour.’ Substantially more coal is actually being consumed per kilowatt-hour produced, but much of the coal would have been required anyway to supply the heating needs. However, when the steam from electrical production exceeds heating needs, only electricity is being produced with the extra coal consumption. The entire incremental coal use must be assigned to the production of electricity, since no additional heat load is being supplied. The effective heat rate for this additional electricity is about 18,400 Btus/kWh. It is these electrical needs in excess of the "steam balance point" that GVEA is attempting to supply to the military. Incremental Cost of Electricity Cogeneration Plant Incremental Cost (mills/kWh) 60 50F 30F 10F ° L L al. 1 n hy eyes fee Poe L L 1 iL o 12 3 46 6 6 78 § 0 nN 12 «13 «14 Electrical Load (MW) 250,000 pounds/hour of Steam Heat Demand Balance: 25 pounds of steam/kWh Figure 1 - Incremental Cost of Electricity as a Function of Load for a Cogeneration Plant Figure 1 illustrates the balance point concept. The figure shows the short-run incremental cost of producing a kilowatt-hour of electricity when the steam demand for heating the military base is 250,000 pounds per hour. For up to 10 MW of electrical demand, full extraction is occurring, and the effective cost per kWh is about 16 mills/kWh. For electrical demands beyond this point, the steam required is not utilized for heating purposes and is condensed in a cooling pond. The cost for each additional kWh beyond the balance “the formula for determining the effective heat rate (Btus/kWh) is: (Rx Hz - Rx Tg + 3,413) / Hg, where Hy = the efficiency of a heating-only system (typically 0.80), T, = the total efficiency of the cogeneration system counting the electricity and the useful heat produced (typically 0.75), and R = the total amount of coal used per kilowatt-hour of electricity produced (typically 20,000 Btus per kWh). For the typical values listed, the effective heat rate of electrical production is 5,500 Btu/kWh. 4 point is about 50 mills/kWh. The following sections examine some of the issues that might affect the probability of the military entering into a contract with civilian utilities for the purchase of power in excess of steam requirements. Economic factors, reliability issues, impacts on ice fog, and regulatory/institutional issues are discussed. 2.2.1 Economic Issues Figure 2 shows the short-run variable cost of power produced by a military coal plant, as a function of the effective heat rate of the plant. Also shown is the range of electricity purchase prices paid by GVEA for electricity from the existing Anchorage-Fairbanks intertie. If additional electricity is sold by GVEA to the military, the majority of the additional electricity will be delivered by the existing Anchorage-Fairbanks intertie, since the intertie has available capacity for much of the year. If the military purchases electricity from GVEA, they will generate less electricity from their coal plants and save the cost of the avoided fuel and avoided operation and maintenance. Thus, if the avoided fuel and O&M cost of the military coal plants (the sloped line on the figure) exceeds the purchase price off the intertie (the shaded bar), there is economic incentive for a transaction to occur. For the electricity below the steam balance point, the effective heat rate of the military generating facilities is about 5,500 Btu/kWh, shown by Band A in the figure. The variable generation cost at this heat rate is about 16 mills/kWh.° This cost is substantially less than the price paid by GVEA for intertie electricity, so it is clear that this segment of electricity demand will not be met by civilian generation in the short run. The high efficiency of the military’s cogeneration facility produces inexpensive electricity. A long run economic comparison needs to consider the capital costs and fixed operation and maintenance costs of military and civilian generation facilities in addition to the above variable costs. Such a comparison determines whether economic factors might lead the military to retire their electrical generation facilities in the future. Additional civilian capacity will probably be supplied by combustion turbines with capital and fixed O&M costs of about 11 mills/kWh (levelized real cost). Future replacements of military generation facilities will probably be coal-fired steam turbines. Coal generation facilities typically have capital and fixed O&M costs of 50 mills/kWh. However, the capital and fixed O&M costs attributable to the production of electricity at military cogeneration facilities is less. A substantial fraction of the capital cost and fixed O&M cost would be required for the operation of their central steam heating system even if electricity is not generated. Incremental capital and fixed O&M costs are closer to 30 mills/kWh. The 15 - 20 mill/kWh variable cost advantage held by the military is cancelled by a 19 mill/kWh capital plus fixed O&M advantage by the civilian utilities. However, natural gas and oil (civilian fuels) are expected to escalate faster in price than coal (military fuel), thus increasing the SThe figure assumes a coal cost of $2.60/MMBtu. Eielson and Fort Wainwright reported prices between $2.60 and $2.70 per MMBtu. Coal Plant Electricity Cost Versus Heat Rate Variable Short-Run Cost, mills/kWh 70; 60F so} , | 40 Intertie Purchase Price | go KAKI (UM oo | | 20+ f oe | | | A B | ot | ——— 0 2 4 6 8 1 12 14 #16 18 20 22 24 Coal Plant Heat Rate (000 Btu/kWh) $2.60/MMBtu Coal Cost, 2 mills/kWh Variable O&M Figure 2 - Coal Plant Short-Run Variable Cost as a Function of Heat Rate military’s variable cost advantage over time. Although the long term cost of military electrical generation (below the steam needs of military facilities) is roughly comparable to the cost of civilian generation, the factors discussed in the subsequent sections (e.g. reliability) indicate that the military will continue to operate their plants. As stated before, communication with both GVEA and the military confirm this conclusion. The costs of military power production beyond their steam needs are substantially different than costs below the steam balance point. Each kilowatt-hour above the balance point requires a substantial amount of fuel, because no heating credit can be deducted from the fuel use. The effective heat rate for this electricity is about 18,400 Btu/kWh (see Box 1). At this heat rate, Figure 2 (Band B) indicates that the military’s variable generation cost exceeds the price paid by GVEA for electricity from the Anchorage-Fairbanks intertie by 15-20 mills/kWh. Thus, there is economic incentive in the short-run for the military to purchase electricity from GVEA for needs above their steam balance point. In the long- run, the advantage may be larger because the military might substitute less expensive (capital cost) diesel generation capacity for more expensive coal capacity. Increased power purchases will decrease the utilization of military coal capacity. Below some threshhold utilization, a switch to diesel capacity is economical. It should be noted that Fort Greely pays Fort Wainwright from 60 to 70 mills/kWh for their electricity. Both facilities are army bases and the charge rate is only for accounting Electricity in excess of the steam balance point is produced by generating additional steam with the plant boilers and passing this steam through all stages of the steam turbine; none of the steam is extracted for heating purposes. The amount of electricity generated with the steam is determined by the steam turbine’s "water rate", expressed as pounds of steam required per kWh of electricity produced. The water rate is largely a function of the inlet and outlet pressure and temperature conditions of the turbine. Both the Eielson and Wainwright steam turbines operate at an inlet pressure of about 400 psi and an inlet temperature near 650 degrees F. They have outlets operating at about 2 inches of mercury absolute pressure. The power plant operators claim the full load water rates are 11 pounds per kWh, and this figure is consistent with that determined from a thermodynamic chart assuming an 80-85% isentropic turbine efficiency (100% would characterize the theoretically best turbine operating between the inlet and outlet states). For the purposes of calculating the cost of electricity beyond the steam balance point, the incremental water rate is the important parameter, i.e. how much additional steam is required to produce an additional kWh. Gerald England, the UAF plant supervisor, says that his 10 Ib/kWh turbine has an incremental water rate of 9.08 Ib/kWh, because efficiency improves as full-load is approached. Using the same ratio for the military turbines implies an incremental water rate of 10 Ib/kWh. Producing 650 degree F 400 psi steam from boiler feed water requires about 1,250 Btus of heat per pound of steam. To produce 1,250 Btus of heat with an 80% efficient boiler (Eielson and Wainwright boiler efficiencies) requires 1,563 Btus of Coal. Thus, a kWh of generation requires 10 Ib x 1,560 Btu/Ib = 15,630 Btus of Coal. However, machinery within the generation plant typically consume 15% of the generated electricity. Therefore, the heat rate per net kWh is 15,630 Btu/kWh / 0.85 = 18,400 Btus/kWh. We have 90% confidence that the value is between 16,000 Btu/kWh and 21,000 Btu/kWh. | Box 1 - Heat Rates Beyond Balance Point for Eielson AFB and Fort Wainwright purposes. For the purposes of determining economic savings from purchasing power from civilian utilities, the army would compare purchased power prices with Fort Wainwright generation costs, since Greely’s power is generated at Wainwright.° 2.2.2 Reliability Issues The military is quite concerned about the reliability of their power system. The military power plant operators claim that their generation facilities are substantially more reliable than their connection with the civilian utility grid. The reliability of their civilian connection is affected by both generation failures and transmission and distribution failures. They feel that increasing reliance on civilian utilities will degrade the reliability of their power supply. However, no change in reliability would occur for Fort Greely, because the power they receive from Fort Wainwright is currently wheeled over GVEA lines. The extent to which reliability is degraded through further purchases from civilian utilities “Personal communication with John Toenes, Deputy Director of the Department of Engineering and Housing, Fort Rich, December 15, 1988. does not depend only on the relative reliability of military generation facilities and the civilian interconnection. The ability of the military to shed non-critical loads in response to loss of civilian power is important, and the speed with which standby generation can be brought on line is also relevant. Because of reliability concerns, both the power plant operator at Eielson (Ed Bargar) and the power plant operator at Fort Wainwright (Gary Brewster) said they would not buy more than 3 MW at any given time from GVEA (total of 6 MW). Even at this level of take there would be some loss of reliability, but the loss of reliability would be primarily for non- critical loads. Fort Greely could supply their entire load from civilian utilities without loss of reliability because of their existing reliance on the civilian grid. 2.2.3. Ice Fog Concerns At times during the winter ice fog is generated by the coal plants at Eielson and Fort Wainwright. The ice fog at Eielson can affect runway visibility and the fog from the Wainwright plant affects visibility on an adjacent highway. Increased power purchases by the military would reduce the ice fog to some extent. Not all of the military-generated ice fog is produced by coal facilities. Exhaust from vehicles on base is also a significant source. Both Robert Huffman (GVEA Consultant) and Ed Bargar (Eielson Engineer) claim that some benefits would accrue from ice fog reduction, but they state that the economic cost savings from increased power purchases is the more important issue. 2.2.4 Institutional/Regulatory Issues The stated Department of Defense policy is to purchase power from civilian sources when it is less expensive than producing it on base, assuming the civilian power meets suitable levels of reliability. Supplying the military electrical needs in excess of their steam requirements appears to be consistent with this policy, since the civilian utilities have a cost advantage for this portion of the military electrical demand. Mr. John Toenes, the Deputy Director of the Department of Engineering and Housing at Ft. Rich, confirmed that their policy is to pursue power purchases if there is potential for economic savings.’ Toenes says that the Army was first to initiate negotiations with Anchorage Municipal Light and Power for the purchase of economy power for Fort Rich. Toenes also indicated that they have spoken with GVEA about purchases for Fort Wainwright and Greely. He expressed concerns about GVEA’s reliability, which he claims has been degraded by reliance on the Anchorage-Fairbanks intertie. He also indicated that the potential savings are not that large when all costs of the transaction are considered. However, he expressed no reluctance to talk further about possible economy purchases. GVEA has hired Robert Huffman to pursue such purchase contracts with the military. Huffman is currently gathering background information and will begin negotiations shortly. "Personal communication, December 15, 1988. The Alaska Federal-Civilian Energy Efficiency Swap Act of 1980 (Public Law 96-571) is a Federal law that addresses some aspects of military-civilian electricity transactions. One intent of the act was to make legal the sale of electricity from military-owned coal-fired generators to a civilian utility, if such a sale involves economic savings to the military and the civilian utility, and the sale results in a reduction in use of oil and gas. The specific situation that motivated the bill was the potential for sales from Fort Wainwright and Eielson Air Force Base to GVEA during the period when oil prices were high. Another section of the bill, Section 4, addresses sales from civilian utilities to the military, the direction relevant in this analysis. The section states: Sec. 4. For purposes of economy and efficiency and conserving oil and natural gas, whenever practicable and consistent with other laws applicable to any agency and whenever consistent with the requirements applicable to any agency, such agency shall endeavor to purchase electric power from any non-Federal person for consumption in Alaska by any facility of such agency where such purchase— (1) will result in a savings to other consumers of electric energy sold by such non-Federal person without increasing the cost incurred by any agency for electric energy, or (2) will result in a cost savings to such agency of electric energy without increasing costs to other consumers of electric energy, taking into account the remaining useful life of any facility available to such agency to generate electric energy for such agency and the cost of maintaining such facility on a standby basis. The specific situations motivating this section of the bill were the existence of self- generating military installations located near Kotzebue and Naknek, Alaska. The military installations and the neighboring communities were using oil to generate electricity. By supplying the military facilities from the civilian generation system, economies of scale could be achieved that would result in a net generation cost savings and a net reduction in the use of oil. As is current Department of Defense policy, the Swap Act encourages sales when economic savings are likely to occur. However, the act does place some emphasis on the conservation of oil and natural gas. The transactions considered in this report are ones that would actually increase the use of natural gas. The increase in civilian generation required to meet the military load would probably be natural gas-fired generation purchased from the Anchorage-Fairbanks intertie. The displaced military generation would be coal-fired. Mr. Toenes at Fort Rich indicated that Department of Defense policy is to favor the use of coal over oil and gas. However, he maintained that the policy is not strict. 2.3 Estimation of Electrical Load Beyond Steam Balance (This is a Technical Section: Skip to section 2.3.3 if desired) The amount of electrical load that occurs beyond the steam balance point for the coal generation plants at Eielson Air Force Base and Fort Wainwright is estimated in this section. This information was not directly available from the military, so it is calculated here from other data that was supplied. For Fort Wainwright, monthly data giving the amount of steam that was condensed in the cooling pond was provided. From this data, the amount of condensed steam in excess of the minimum required was determined. Then, 9 the electricity produced with this excess steam (which is the electricity in excess of the steam balance point) was estimated. For Eielson, monthly data for the electricity and steam requirements of the base and technical information on the electrical/steam generation plant was provided. The ratio of steam production to electrical production for the generation plant was estimated, assuming maximum steam extraction. From this ratio, the amount of electrical production that results from exactly meeting the base steam requirements can be calculated. The actual electrical use in excess of this electrical production is the electricity demand beyond the steam balance point. 2.3.1 Eielson Steam Imbalance The Eielson generation plant has recently undergone an $18 million renovation that has involved the installation of a new 10 MW turbine. This unit combined with an equally efficient 5 MW unit will supply the bulk of the base electrical needs in the future. Past electrical load has been met by substantially-less efficient turbines, so historical data concerning steam imbalance is of little value. The following calculations concerning steam imbalance are based on the technical characteristics of the new generation configuration. Table 2 summarizes the steam imbalance calculation. Inputs to the calculation include the "Turbine Water Rate, All Stages", a technical parameter of the electrical generation turbine that indicates the amount of steam required to generate a kWh of electricity when the steam flows through all stages of the turbine (i.e. no early extraction occurs for heating purposes) and the turbine is fully-loaded. This figure was supplied by Mr. Ed Bargar, the Eielson Utilities Engineer. "Turbine Water Rate, Pre-Extraction Stages" in the amount of steam required to generate a kWh if the steam only passes through the initial stages of the turbine and then is extracted for heating purposes. This value was calculated from the known inlet conditions of the turbine (400 psi, 650 degrees F), an assumed extraction pressure of 60 psi, and an assumed isentropic turbine efficiency of 85%. Condensing/extracting turbines require that not all of the steam be extracted for heating purposes. Some steam must pass through all stages of the turbine. "Maximum Steam Extraction Percentage" indicates the maximum amount of steam that can safely be extracted from the turbine for heating purposes. Bargar indicated that 90% was typical for their turbines. "In-Plant Electricity Consumption" gives the fraction of each generated kWh that is consumed by equipment within the generation plant. Turbines are typically less efficient at part-load than at full-load. The "Part-Load Water Rate Adjustment" indicates the percent that the average turbine water rate exceeds the full-load water rate. "Maximum GWh/month Purchase" is the maximum electricity that we estimate Eielson would buy in one month because of reliability concerns. It is based on a peak purchase of 3 MW and a 70% load factor. In the calculation section, the turbine water rate is calculated assuming maximum steam extraction is occurring, i.e. 90% of the steam is extracted and 10% flows through the entire turbine. This water rate is adjusted upward to account for part-load inefficiencies and to account for the fact that only 85% of a generated kWh is delivered to the base because of in-plant use. Thus, 27.4 pounds of steam are required at the inlet of the turbine to deliver 1 kWh to the base when maximum steam extraction is occurring. Only 90% of this steam 10 as INPUTS: Turbine Water Rate, All Stages = 11.0 1b/kWh Turbine Water Rate, Pre-Extraction Stages = 24.4 1b/kWh Maximum Steam Extraction Percentage = 90% In-Plant Electricity Consumption = 15% Part-Load Water Rate Adjustment = 7% Maximum GWh/month Purchase = 1.53 GWh/month CALCULATIONS: Turbine Water Rate, Max Extract = , 21.8 1b/kWh Adj. for Part-Load Eff. and In-Plant Use = 27.4 1b/kWh Extracted Steam per kWh to Base, Max Extract = 24.6 1b/kWh lbs steam GWh GWh to to Base on Steam GWh GWh Seasonal Base (million) Curve Deficit Take Distrib. Oct 87 4.41 88.87 3.61 0.80 0.80 7 Nov 4.91 96.27 3,91 1.00 1.00 8.8% Dec 5.96 147.83 6.00 -0.04 -0.04 -0.3% Jan 88 6.00 145.23 5.89 0.11 0.11 0.9% Feb 5.21 110.32 4.48 0.73 0.73 6.4% Mar 4.51 iG 25 4.15 0.37 0.37 3.2% Apr 4.13 80.33 3.26 0.87 0.87 7.7% May 3.86 56.34 2.29 1.58 1.03) 13.4% Jun 3.39 39.26 1.59 1.80 1:53) 13.4% Jul 3.16 36.42 1.48 1.68 253 13.4% Aug 3.80 39-21 1.59 2.21 B53 13.4% Sep 88 3.94 61.96 2.51 1.42 1.42 25% 533: 1,004 40.8 12.5 11.4 100% 21.4% of Total Use na eR ee ee Table 2 - Eielson Electrical Load Beyond Steam Balance Point is actually extracted, so 24.6 pounds of steam are delivered to the base for each 1 kWh of electricity delivered. The table below the calculation section gives actual steam and electricity consumption of Eielson for October 1987 through September 1988. The 4th column, "GWh on Steam Curve", shows how much electricity would have been produced if only enough steam was fed to the turbines to meet the military base steam needs (1 kWh per 24.6 pounds of steam 11 requirement).* The 5th column, "GWh Deficit", shows how much the actual electricity requirements exceeded this hypothetical generation amount. This is the portion of the electrical demand that is in excess of the steam balance point. The 6th column duplicates the 5th column except the monthly GWh values are capped at 1.53 GWh/month. This column estimates what purchases from a civilian utility would have been if operation of the military generation plant were determined by steam needs, and civilian purchases were limited to 1.53 GWh/month to preserve reliability. The total purchases for the year would have been 11 GWh or about 21% of the total Eielson electrical load. The "Seasonal Distrib." column shows the fraction of the total sales that would have occurred in each month. The estimate of total purchases is very sensitive to the "Turbine Water Rate, Pre-Extraction Stages" parameter. A +/- 15% variation in this parameter causes a +30% and -45% variation in total purchases, i.e. purchases vary from 15 GWh/year to 6 GWh/year with this variation in the pre-extraction water rate. Considering other sources of uncertainty in the calculation, a 90% confidence interval for the estimated purchases is 4 - 18 GWh/year with a most likely estimate of 11 GWh/year. Eielson’s electrical load is expected to grow about 20% in the next few years because of new construction on base.’ We assume that the steam load grows proportionately, so the amount of electricity in excess of the steam balance point should also grow at 4.5% per year for the next four years. Beyond that, we assume that the steam and electrical demand will grow at the same rate as the Fairbanks civilian electrical demand, 1.3% per year through 2010 (Railbelt End Use Demand Forecast, middle case, 1995 - 2010 growth rate). As well as requiring proportionate growth in steam and electrical demand, the above growth rate projection also requires that the technical characteristics of the generation plant not change over time. In particular, if the pre-extraction water rate changes at some point in the future, a significant change in the purchased electricity will occur. The Eielson generation plant currently operates at an inlet turbine pressure of 400 psi. If an upgrade of the plant in the future increases the inlet pressure to 600 psi (the inlet pressure of the current University of Alaska, Fairbanks turbines) the pre-extraction water rate could drop to 17 Ib/kWh. At that water rate, net purchases over a year would probably drop to zero (Eielson would sell to the civilian utilities in the winter and the civilian utilities would sell to Eielson in the summer). 2.3.2 Fort Wainwright Steam Imbalance Fort Wainwright provided data for monthly steam flows within their electrical/steam generation plant from which flows to the cooling pond could be determined. The steam 5The generating plant actually used during this year had characteristics substantially different than the one modeled here. The objective of the calculation is to estimate future steam imbalances, so characteristics of the plant to be used in the future are used in the calculation. Personal communication with Ed Bargar, Eielson Utilities Engineer, January 11, 1989. 12 #2 flow to the cooling pond consists of two components: the minimum steam required to maintain safe operation of the turbine, and the steam used to generate electricity in excess of the steam balance point. The minimum flow required to maintain safe operation of the turbine was estimated and subtracted from the total flow to determine the steam used to generate electricity in excess of the balance point. With an estimated water rate for the turbine, the amount of electricity produced with this steam was calculated. Maximum Extracted Steam Fraction = 90% In-Plant Use = 15.0% Turbine Water Rate, All Stages = 11.0 lb/kWh Water Rate per Distributed kWh = 12.9 lb/kWh Maximum Wainwright Purchase, GWh/month = 1.53 GWh/month Annual Greely Load = 17.2 GWh/year Total Wainwt. Steam Steam Excess Beyond Greely Beyond Wainwt. GWh GWh to Base Condens Condens Balance Load Balance Purchase Month Gener. Distrib ----- million pounds ----- GWh GWh GWh GWh 1 2 3 4 5 6 7 8 9 10 sesssssssssssssssssssssssssssssssssssssssssssssssssssssesssssssssssssesssssss=2s25=2=2== Dec 87 8.2 7.0 142.6 54.3 38.5 2.97 1.68 1.29 1.29 Jan 88 8.5 7.2 148.5 57.2 40.7 3.15 1.74 1.41 1.41 Feb 8.0 6.8 132c7 59.0 44.2 3.42 1.63 1.79 1.53 Mar 7.6 6.4 122.1 54.3 40.7 3.15 1.54 1.60 1.53 Apr 6.7 Bar 98.5 48.5 37.5 2.90 1.37 1.55 4355 May 6.2 5.3 77.0 48.5 40.0 3.09 1.27 1.83 1.53 Jun 5.7 4.8 veut 44.2 36.2 2.80 1.16 1.64 1253) Jul 5.8 4.9 72.9 44.3 36.2 2.80 1.18 1.62 1.53 Aug 6.0 5.1 70.7 50.1 42.3 3. 1 2 ie Sep 6.3 5.4 85.2 49.5 40.1 35 1 1 ts Oct 7.2 6.1 115.4 52.1 39.3 3. 1 1 1. Nov 88 8.0 6.8 151-5 60.1 45.5 3. 1 1. sassssssssssssssssssssssssseses2e2s52===5===> = = = 84.2 71.6 1,268.9 622.2 481.3 5 2 18. Table 3 - Fort Wainwright and Fort Greely Electrical Load Beyond Steam Balance Point - Calculation Method 1 Table 3 summarizes the calculation of the steam imbalance for the Fort Wainwright coal plant. The 2nd column of the table gives the gross electrical generation of the plant, data supplied by the Wainwright plant operator, Gary Brewster. The 3rd column estimates the amount of electricity delivered to the base and the GVEA distribution system for sale to Greely. In-plant electricity consumption of 15% is assumed to derive the column. The 4th column gives the amount of steam delivered to the military base steam distribution system (raw data). The total steam produced by the boiler was supplied as raw data but is not shown in the table. The Sth column is the amount of steam that condensed in the cooling pond, which is determined by subtracting the steam delivered to the base heating system from the total steam produced by the boiler. The portion of this steam that is in excess of minimum requirements in shown in column 6 (i.e. the steam in excess of that which would be sent to the cooling pond if the plant just followed the base heating needs. It is assumed that a maximum of 90% of the steam can be extracted from the turbine for heating purposes.) This excess steam is the steam that was used to generate electricity beyond the steam balance point. 13 To calculate the amount of electricity generated with this steam beyond the balance point, the water rate of the turbine (all stages) must be known. A number of methods were used to estimate the water rate of the Wainwright turbines. All generated approximately the same results. First, the power plant operator, Gary Brewster, confirmed that 11 pounds per kWh of gross generation was as reasonable number. Second, this figure is consistent with operation between a 400 psi, 650 degree F turbine inlet, a 2 psi (absolute) outlet, and a 80% isentropic turbine efficiency. Finally, a statistical regression was performed of gross generation versus (steam delivered to the base) and (steam condensed in the cooling pond). The reciprocal of the coefficient of (steam condensed in the cooling pond) provides an estimate of the water rate of the turbine. The regression gave a water rate of 10.9 Ib/kWh with a 90% confidence interval of 9.8 to 12.3 lb/kWh. An estimate of 11 lb/kWh was used in the analysis. This gross water rate was converted to a net water rate of 12.9 Ib/kWh by assuming a 15% in-plant electrical consumption (typical for the Wainwright plant). Thus, for every 12.9 pounds of steam passing through all stages of the turbine and then condensed in the cooling pond, 1 kWh of electricity is generated. (This is also a critical figure for calculating the plant heat rate, as discussed earlier in this report). Using the water rate of 12.9 lb/kWh and the excess cooling pond steam in column 6, the electricity beyond the balance point can be calculated in column 7. Assuming no constraints on the amount of electricity purchased by the military bases, this column would represent the total sales to Wainwright and Greely if all electrical needs beyond the steam balance point were met by civilian utilities. However, we assume that Wainwright will not buy more than 1.53 GWh/month (3 MW at a 70% load factor) because of reliability concerns. Columns 8 through 11 are used to analyze the impact of that assumption. First, the Greely load is subtracted from column 7 to produce column 9. The total annual consumption at Greely for 1987 was used, and it was distributed across months according to the distribution of total Wainwright and Greely consumption. Column 9 gives the portion of the electricity beyond the balance point attributable to Wainwright electrical demand. Column 10 duplicates this column except that monthly amounts are capped at 1.53 GWh/month, the reliability constraint. This calculation method shows that the total sales to Wainwright would be 18 GWh/year. The largest source of error in this estimate is probably the inaccuracy of the steam meters used to provide the data on steam flows. The critical steam flow for the calculation is the flow to the cooling pond. It was derived from the difference between two large steam flows, the total boiler flow and the flow to the base. Plant operator Brewster says that the steam meters are 1953 vintage and are no better than 5% accurate. Since the cooling pond flow is the difference between two measured flows, the difference is much less accurate than 5%. We calculate a +/- 9 GWh/year inaccuracy in the electricity beyond the balance point due to a 5% steam meter accuracy. The 18 GWh/year result for the Wainwright imbalance (in addition to the 17 GWh/year Greely imbalance) is high relative to informal estimates by the Wainwright Power plant operator and GVEA personnel. Also, the data used in the calculation imply that the "pre- extraction water rate" of the Wainwright plant is very high, about 40 - 50 pounds per kWh. (The pre-extraction water rate indicates how much extracted steam is required to produce a kWh; it is always higher than full turbine water rate because extracted steam only flows 14 through the initial stages of the turbine.) Brewster claims that this is not so. Our suspicion is that the data for the flow to the condenser is quite inaccurate because of steam meter inaccuracies. An alternative calculation similar to the one done for Eielson was performed as a second estimate of the imbalance. The Eielson calculation requires no condenser steam flow data, but does require more estimates as to technical parameters of the generation plant. For the Wainwright load and plant, the calculation yielded a 3 GWh/year Wainwright imbalance, in addition to the 17 GWh/yr Greely imbalance. As a compromise estimate for the Wainwright imbalance we choose 12 GWh/year with a 90% confidence interval of 2 to 25 GWh/year. Sales to Greely are estimated to be 17 GWh/year with a confidence interval of 14 - 18 GWh/year. Wainwright expects 10 - 15% electrical load growth in the next few years due to additional light infantry personnel according to Brewster. However, personnel increases are expected . to be about 40%. Brewster states that much of the growth will be absorbed in existing and remodeled facilities. Also, some the housing for the troops will be off-base. We assume a slightly higher growth rate than the military’s estimate and project electrical needs beyond the balance point to grow at 8% per year for the next 3 years. From then on, we project 1.3% per year growth (the mid-case Fairbanks growth) until 2010. Once again, this assumes that electrical needs and steam needs increase proportionately, and that the reliability constraint grows with general load growth. As was mentioned for Eielson, an upgrade of the Wainwright plant to higher inlet turbine pressures (e.g. similar to those at UAF) would reduce the imbalance significantly. No such upgrade is being planned in the near term according to Brewster. Recent decisions by the army indicate that approximately 100 active duty personnel will be transferred out of Fort Greely. This represents a 16% drop in the active duty personnel on the base. Assuming a somewhat less than proportionate drop in electrical demand means that the Greely load will decrease by 3% per year for approximately the next four years. Beyond then, we project the load to grow at 1.3% per year until 2010 (Fairbanks mid-case). The army has studied the possibility of installing a coal fired generation plant at Fort Greely to supply both the electrical and steam needs of the base. If such a plant were installed, any purchases from civilian utilities would drop significantly. The project 10 currently is in "long-range planning". 2.3.3. Summary of Potential Military Purchases Table 4 summarizes the potential purchases of the Fairbanks-area military bases. The seasonal distribution of the load and an estimate of its peak demand are given in Table 5. The seasonal distribution was derived from the individual seasonal distributions in the previous tables. The peak purchase demand for 1988 was estimated by assuming a 3 MW reliability constraint for both Eielson and Wainwright, and a Greely peak in May of 3.2 MW. Personal communication with Lee Gayle, Department of Engineering and Housing, Fort Rich, December 15, 1988. 15 % of POTENTIAL PURCHASES (GWh) Annual Measured at Military Bases Month Purchases 0% = 8.4% . - ~ 7.4% 1989 11.9) 13.0 6 8.1% 1990 12.4 14.0 16.2 -6 9.3% * 1991 13.0 554 15.7 8 9.1% 1992 13.6 15.3 15.2 | 9.1% 1993 13.8 15.5 15.4 ae 9.2% 1994 14.0 15.7 15.6 or) 9.1% 1995 14.1 15.9 15-8 ao 8.1% 1996 14.3 16.1 16.0 a] 8.9% 1997 14.5 16.3 16.2 ot Dec 6.3% 1998 14.7 16.5 16.5 of 1999 14.9 16.8 16.7 oO ** - Month of largest 2000 15.1 17.0 16.9 9 purchase 2001 15.3 ifce Ue 6 2002 15.5 17.4 17.3 2 Estimated Peak Purchase 2003 15.7 17.7 17.6 9 Demand for 1988 = 9.2 MW 2004 15.9 17.9 17.8 oo (May) 2005 16.1 18.1 18.0 2 2006 16.3 18.3 18.2 9 TT a 2007 16.5 18.6 18.5 -6 Table 5 - Seasonal 2008 16.7 18.8 18.7 3 © tet . * 2009 16.9 19.1 19.0 0 Distribution of Potential 2010 17.2 19.3 19.2 7 Military Purchases Average Growth = 1.4%/yr Table 4 - Potential Purchases of Civilian Electricity by the Military 3. University of Alaska, Fairbanks Electricity Purchases The University of Alaska in Fairbanks supplies their electricity needs with coal-fired and garbage-fired steam turbines that allow for extraction of steam to supply the campus heating, hot water, and some cooling needs (absorption cooling). The total electricity demand for FY88 was 35 GWh. They have had an economy energy contract for a number of years with GVEA to buy and sell power. In FY88 net purchases from GVEA were about 1 GWh. Purchases from GVEA are primarily for electricity beyond their steam balance point, which costs them approximately 38 mills/kWh when using coal and 29 mills/kWh when using garbage pellets (fuel cost only). However, not all of their requirements beyond their balance point are purchased from GVEA. GVEA’s rate varies and is only sometimes below the 38 mills/kWh coal production cost (GVEA current rate is 45 mills/kWh). Electricity below the UAF steam balance point costs them about 15 mills/kWh (fuel cost). Gerald England, the UAF plant supervisor, states that their contract with GVEA will expire in 1991. If a more favorable purchase price is obtained from GVEA, more electricity will be purchased. In particular, a favorable purchase price would allow the University to avoid the installation of additional condenser capacity for summer loads. Purchased power would 16 displace the need for this condenser. England estimates that they would buy 1 MW from May through July, but only on weekdays when the outside temperature is in excess of 65 degrees F. We estimate this load to be about 1.2 GWh/year in addition to the existing 1 GWh/year purchase, and use this as a most-likely scenario. The purchases are escalated at 1.3%/year thereafter, the mid-case Fairbanks growth rate. Table 6 summarizes the expected UAF purchases over time. We estimate that 80% of the purchases would occur during the summer months. England states that if the purchase price from GVEA were low enough to be competitive with all of UAF’s needs beyond the balance point, about 25% (conservatively low estimate) of their needs would be supplied by GVEA. For their FY88 load this amounts to 9 GWh/year. However, England says that such a favorable purchase price is unlikely. He says that their coal supplier, Usibelli, would probably match any decreases in purchase power prices. When asked about the potential for significantly increased sales to UAF, Marvin Riddle, the GVEA dispatcher, did not expect any substantial increases. 17 EXPECTED UAF PURCHASES (GWh) Measured at UAF YEAR PURCHASE 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 NNN NNN NN NNN NNN NN NNN NN RMPOONNNAKBKUUUF HE RWWWW DYNO cal i SAB nN Table 6 - Expected UAF purchases This page intentionally left blank Appendix J: Additional Loads to be Served by the Northeast Intertie This appendix contains a load forecast for areas in the Copper Valley Electric Association service territory which would be served by the proposed Northeast intertie project. This page intentionally left blank UNIVERSITY OF ALASKA ANCHORAGE 3211 Providence Drive Anchorage, Alaska 99508 (907) 786-7710 INSTITUTE OF SOCIAL AND DATE: 11 January 1989 ECONOMIC RESEARCH TO: Dick Emerman FR: Alan Mitchell RE: NE Intertie Load Forecast Attached is the load forecast for the non-Fairbanks portions of the NE intertie load. The first three pages include summary graphs for both energy and peak demand and an associated summary table. Low, Mid, and High cases are presented. The tables on the next three pages summarize the load components that comprise the three cases. An "Include" row in the table indicates whether a particular load component is included in the case summarized by the table. A'"1" indicates that the load is included and a "0" indicates that the load is not included. Some of the load components are not included in any cases. They are shown for information purposes. For each load component, a load factor was estimated and is also shown in a row in the table. These load factors were used to develop the peak demand estimates. Providing service to some of the load components is contingent upon construction of the NE intertie (e.g. Alyeska Pump Station 10, Paxson, and Tok). For these components the capital cost of serving these loads was estimated (i.e. the cost for necessary distribution lines, transformers, etc.). Since these costs will only be incurred if the intertie is built, these need to be included as costs in the system analysis of the NE intertie. Operation and maintenance expenses for the equipment should also be included in the system analysis. More detail for each of the load components is provided after the summary charts. You mentioned previously about the possibility of some relatively large electrical loads for mining operations north of Glennallen. We investigated further and found no such loads. Doug Bursey mentioned the Valdez Creek mine; however, this mine is located on the Denali Highway, closer to the Cantwell side than the Paxson side. There has been some mining activity at the Slate Creek mine in the past few years, but the activity was small, and the mine is located 30 miles off the intertie route. The main difference between the load forecast cases is the estimate for the amount of grid- supplied electricity for the US Air Force Backscatter Radar facility. The estimate in the low case is 0 GWh/year and in the high case is 32 GWh/year. Much of this uncertainty will be eliminated when the USAF selects a contractor for their power requirements on March 24, 1989. There are about 5 bidders in competition, and their designs vary in the amount of grid-generated power they are likely to use. The two bumps in the High Case forecast are due to the employment boom associated with two major projects, the potential APRI refinery in Valdez (1993) and the TAGS gasline (2005). 1 A DIVISION OF THE UNIVERSITY OF ALASKA STATEWIDE SYSTEM OF HIGHER EDUCATION 160 140 120 100 1988 1993 1998 2003 2008 Year —— Low Case ~—+—Mid Case ~*~ High Case Loads Include Distribution Losses NE Intertie Load Forecast System Peak (MW) 30 1988 1993 1998 2003 2008 Year —— Low Case ~—+— Mid Case ~*~ High Case Loads Include Distribution Losses \| \| NE INTERTIE LOAD SUMMARY Loads Include Distribution Losses YEAR ae LOW) pa ee | (MELD) coon ===) |B LGH i=<— GWh MW GWh MW GWh MW 1988 49 10.2 49 10.2 49 10.2 1989 49 10.1 49 10.1 50 10.3 1990 51 10.6 51 10.6 52 10.8 1991 53 10.9 53 aelenL 54 11.2 1992 52 10.8 72 13.5 79 14.5 1993 50 10.4 a? 14.1 90 16.2 1994 51 10.6 78 14.3 107 19.5 S95: Sa 10.6 83 14.9 111 20.2 1996 52 10.7 83 14.9 95 16.9 1997 52 10.8 84 Siok 96 a7 aos 1998 53) 10.9 84 15.2 97 17.2 1999 53) 11.0 84 Lies 98 Lio 2000 54 Lied 85 15.4 99 17.6 2001 54 11.2 85 15.5 100 Tie: 2002 54 11.2 86 15.6 100 17.9 2003 54 3 86 15.7 103 18.3 2004 55 11.4 87 15.8 105 19.0 2005 55 11.5 87 15.9 121 22.2 2006 56 11.6 88 16.0 129 23.9 2007 57 be ee 88 16.1 139 26.0 2008 57 11.8 89 16.2 110 19.9 2009 57 1iz.9 89 16.3 110 19.9 2010 58 12.0 90 16.4 Ee 20.0 NE INTERTIE LOAD FORECAST SUMMARY *** LOW CASE *** ALL Load Values are in Gigawatt-Hours (GWh) Distribution Losses = 9.0% |PEAK w/ TOTAL w/ SALES | 1 2 3 4 5 6 7 8 9 10 "1 12 13, | LOSSES LOSSES TOTAL | ---- BASE FORECAST ---- DISTRIB ------- USAF RADAR ------- --APRI REFINERY-- ALYESKA YEAR | MW GWh GWh | Low Medium High EXPANSION LOW MID HIGH = Popul Indust Popul Ps10 PAXSON TOK INCLUDE (1-Yes, O-No) --> 1 0 - 0 1 1 0 0 1 0 0 0 1 0 LOAD FACTOR 55% 55% 55% 55% 85% 85% 85% 55% 70% 55% 90% 55% 55% CAPITAL COST TO SERVE ($ mil) $1.5 $0.15 $10.6 1988 | 10.2 49 45 | 44.6 44.6 44.6 0.00 0 0 0 0.0 0 0.0 0.0 0.00 0.0 1989 | 10.1 49 44 | 44.3 44.5 45.1 0.00 0 0 0 0.0 0 0.0 0.0 0.00 0.0 1990 | 10.6 51 46 | 44.4 44.8 45.4 0.70 0 0 0 1.2 0 0.0 0.0 0.00 0.0 1991 | 10.9 53 48 | 45.0 45.6 46.1 0.71 0 0 0 2-5 0 0.0 0.0 0.00 0.0 1992 | 10.8 52 47 | 44.8 45.6 45.8 0.71 0 17.4 24.0 1.7 0 0.0 0.0 0.00 0.0 1993 | 10.4 50 46 | 44.6 45.6 45.6 0.71 0 23.2 32.0 0.5 0 5-5 0.0 0.00 0.0 1994 | 10.6 51 46 | 44.9 46.2 46.4 0.72 0 23.2 32.0 0.5 0 17.2 0.0 0.20 0.0 1995 | 10.6 51 47 | 45.2 46.6 46.8 0.73 0 23.2 32.0 0.5 0 17.1 3-9 0.20 8.0 1996 | 10.7 52 47 | 45.6 46.6 47.2 0.73 0 23.2 32.0 0.5 276 2.3 3.9 0.20 8.0 1997 | 10.8 52 48 | 46.1 47.6 48.1 0.74 0 23.2 32.0 0.5 276 2.3 3.9 0.21 8.2 1998 | 10.9 53 48 | 46.4 47.9 48.7 0.75 0 23.2 32.0 0.5 276 2.3 3.9 0.21 8.2 1999 | 11.0 53 48 | 47.0 48.2 49.3 0.75 0 23.2 32.0 0.5 276 2.3 3.9 0.21 8.3 2000 | (11.1 54 49 | 47.4 48.6 50.2 0.76 0 e522, 32.0 0.5 276 2.3 3.9 0.21 8.3 2001 | 11.2 54 49 | 47.5 49.1 51.0 0.77 0 23.2 32.0 0.5 276 2.3 3.9 0.21 8.4 2002 | 11.2 54 49 | 47.8 49.4 51.7 0.77 0 23.2 32.0 0.5 276 2.3 3.9 0.21 8.5 2003 | 11.3 54 49 | 48.0 49.8 53.6 0.78 0 23.2 32.0 0.5 276 2.2 3.9 0.22 8.6 2004 | 11.4 55 50 | 48.3 50.3 56.3 0.78 0 23.2 32.0 0.5 276 2.2 3.9 0.22 8.6 2005 | 11.5 55 50 | 48.8 50.8 70.4 0.79 0 23.2 32.0 0.5 276 (4 3.9 0.22 8.7 2006 | 11.6 56 51 | 49.3 51.2 78.1 0.80 0 23.2 32.0 0.5 276 2.2 3.9 0.22 8.8 2007 | (11.7 57 510] 50.0 51.8 87.2 0.81 0 23.2 32.0 0.5 276 2.2 3.9 0.22 8.9 2008 | 11.8 57 52 | 50.4 52.2 60.6 0.81 0 23.2 32.0 0.5 276 2.2 3.9 0.23 9.0 2009 | 11.9 57 52 | 50.8 52.6 60.3 0.82 0 23.2 32.0 0.5 276 2.2 3.9 0.23 9.0 2010 | 12.0 58 53) | 51.1 DSs.4 61.1 0.83 0 23.2 32.0 0.5 276 2.2 5-9, 0.23 O14 ALL Load Values are in Gigawatt-Hours (GWh) Distribution Losses |PEAK w/ TOTAL w/ | Losses YEAR | LOSSE GWh INCLUDE (1-Yes, 0-No) --> LOAD FACTOR CAPITAL COST TO SERVE ($ mil) Ss 9.0% SALES TOTAL GWh 1 NE INTERTIE LOAD FORECAST SUMMARY 2 *** MID CASE *** 3 ---- BASE FORECAST ---- Low Medium 4 DISTRIB High EXPANSION 6 7 USAF RADAR -- MID HIGH 1 0 85% 85% 9 10 --APRI REFINERY-- Indust Popul 1 ALYESKA Ps10 12 PAXSON 10.2 10.1 10.6 11.1 13.5 14.1 14.3 14.9 14.9 15.1 15.2 15.3 15.4 15.5 15.6 15.7 15.8 15.9 16.0 16.1 16.2 16.3 16.4 SSSGRRKRAGKRRRRAAIANUS 89 90 S a KAN DAARAAN AA AAU A S ) - 49.8 50.3 50.8 51.2 51.8 52.2 52.6 pS. 4 (—— — — — — — — — — — — — —— — — — — — — —) cooeoeoeao eee eo eo oOo oO oO oO oO oO oO oO oO OO NNNNKANHNNNANHNNNNNNN ROCCO 32.0 32.0 32.0 32.0 32.0 32.0 32.0 32.0 32.0 32.0 32.0 32.0 32.0 32.0 32.0 32.0 32.0 32.0 cooooeoec eco oa NH COO viuuuuuuuunwnoondo ooo www 0.5 0.5 0.5 0.5 0.5 0.5 ooooooco 276 276 276 276 276 276 276 276 276 276 276 276 276 276 276 “nNuUsoo000 n w 2.3 2.3 2.3) 2.3 2.3 2.2 2.2 222 2.2 2.2 2.2 2c2 2.2 NWWWWnNWnWnNWnNWnNWnwWwWUWWwWwWwWwwoaoaoaoaoand eoovovveveoava ovo eaWDeDoDD DDO OOOO CO oO — 2 NE INTERTIE LOAD FORECAST SUMMARY *&** HIGH CASE *** ALL Load Values are in Gigawatt-Hours (GWh) Distribution Losses = 9.0% |PEAK w/ TOTAL w/ SALES | 1 2 3 4 5 6 7 8 9 “10 1 12 13 | LOSSES LOSSES TOTAL | ---- BASE FORECAST ---- DISTRIB = os: USAF RADAR: = s-<—<= ~-APRI REFINERY-- ALYESKA YEAR | MW GWh GWh | Low Medium High EXPANSION Low MID HIGH Popul = Indust Popul Ps10 PAXSON TOK INCLUDE (1-Yes, 0-No) --> 0 0 1 1 0 0 1 1 0 1 1 1 0 LOAD FACTOR 55% 55% 55% 55% 85% 85% 85% 55% 70% 55% 90% 55% 55% CAPITAL COST TO SERVE ($ mil) $1.5 $0.15 $10.6 1988 | 10.2 49 45 | 44.6 44.6 44.6 0.00 0 0 0 0.0 0 0.0 0.0 0.00 0.0 1989 | 10.3 50 45 | 44.3 44.5 45.1 0.00 0 0 0 0.0 0 0.0 0.0 0.00 0.0 1990 | 10.8 52 47 | 44.4 44.8 45.4 0.70 0 0 0 1.2 0 0.0 0.0 0.00 0.0 1991 | {1-2 54 49 | 45.0 45.6 46.1 0.71 0 0 0 2.3 0 0.0 0.0 0.00 0.0 1992) [96-5 79 72 | 44.8 45.6 45.8 0.71 0 17.4 24.0 1.7 0 0.0 0.0 0.00 0.0 1993 )\[ 6.2 90 82 | 44.6 45.6 45.6 0.71 0 23.2 32.0 0.5 0 3-5. 0.0 0.00 0.0 1994 | 19.5 107 97 | 44.9 46.2 46.4 0.72 0 23.2 32.0 0.5 0 ive 0.0 0.20 0.0 1995 | 20.2 111 101 | 45.2 46.6 46.8 0.73 0 ge 32.0 0.5 0 A7ed 3.9 0.20 8.0 1996 | 16.9 9 87 | 45.6 46.6 47.2 0.73 0 as-2 32.0 0.5 276 25 5-9 0.20 8.0 1997 | (17.1 96 88 | 46.1 47.6 48.1 0.74 0 23.2 32.0 0.5 276 225 i) 0.21 8.2 1998 | 17.2 97 88 | 46.4 47.9 48.7 0.75 0 23.2 32.0 0.5 276 2.5 3.9 0.21 8.2 1999 | 17.3 98 89 | 47.0 48.2 49.3 0.75 0 es.2 32.0 0.5 276 2.3 3e9. 0.21 8.3 2000 | 17.6 9 90 | 47.4 48.6 50.2 0.76 0 23.2 32.0 0.5 276 2-3 3.9 0.21 8.3 poi | 17-7 100 91] 47.5 49.1 51.0 0.77 0 235.2 32.0 0.5 276 2.3 3.9) 0.21 8.4 2002 | 17.9 100 1) 47.8 49.4 51.7 0.77 0 sce 32.0 0.5 276 2.3 S97. 0.21 8.5 2003 | 18.3 103 93 | 48.0 49.8 53.6 0.78 0 23.2 32.0 0.5 276 2.2 5.9. 0.22 8.6 2004 | 19.0 105 % | 48.3 50.3 56.3 0.78 0 23.2 32.0 0.5 276 2.2 Fh) 0.22 8.6 2005 | 22.2 121 110 | 48.8 50.8 70.4 0.79 0 2.2 32.0 0.5 276 2.2 SE) 0.22 8.7 2006 | 23.9 129 118 | 49.3 51.2 78.1 0.80 0 23.2 32.0 0.5 276 ace 3.9) 0.22 8.8 2007 | 26.0 139 127 | 50.0 51.8 87.2 0.81 0 23.2 32.0 0.5 276 ene 3.9 0.22 8.9 2008 | 19.9 110 100 | 50.4 wee 60.6 0.81 0 23.2 32.0 0.5 276 2.2 3.9 0.23 9.0 2009 | 19.9 110 100 | 50.8 52.6 60.3 0.82 0 os.e 32.0 0.5 276 ee 3.9 0.23 9.0 2010 | 20.0 111 101 | 51.1 53.1 61.1 0.83 0 23,2 32.0 0.5 276 card aie, 0.23 ea | LOAD #1, #2, #3: Base Forecast This component of the forecast contains the current Copper Valley Electric (CVEA) load. Reflected in the projection are general growth in the economy, operation of the Alyeska pipeline at its current level, and construction of the TAGS gasline in the high case (#3) only. Other potential additions to the CVEA load and potential loads that will be accessed by the proposed NE intertie are included as separate components of the forecast (Loads #4 -#13). WHERE LOCATED: Copper Valley Electric Service Territory: Glennallen, Copper Center, Valdez, and surrounding area. SERVICE CONTINGENT ON NE INTERTIE?: No INCLUDED IN WHICH CASES?: #1 - Low Case, #2 - Mid Case, #3 - High Case CALCULATIONS: The three attached worksheets detail the load projections for the Low, Mid, and High cases. The projection of residential electric customers for CVEA was patterned after the MAP model household projections for the Valdez/Cordova Census area. The particular MAP projections selected for the CVEA Low, Mid, and High cases were exactly the cases used in the Low, Mid, and High cases of the Railbelt end-use electricity forecast. The household growth rates implied by the Low and Mid cases from 1988 through 2010 for CVEA are almost the same, about 1.44% per year. Although these two cases are associated with different household growth rates for the Railbelt as a whole, they are associated with similar customer growth rates for the CVEA service territory because tourism growth is high in the Low Case and low in the Mid case. Because the economy of this census area has a high tourism component, this tourism assumption causes the Low case to produce as much growth as the Mid case. Thus, the demographic scenarios for this forecast may not span the full range of possibilities for the CVEA service territory, but the forecast remains consistent with other components of the Railbelt Intertie Upgrade study. The substantial increase in the number of customers in the years 2005 through 2007 in the High Case is due to construction of the TAGS gasline. The total residential load in a given year is calculated by multiplying the number of residential customers by a use per customer. The use per customer for 1988 was taken from actual data. The change in use per residential customer over time was derived from the Railbelt end-use forecasts for other regions in the Railbelt. Specifically, the Mid case forecasts for Fairbanks, Kenai, and Mat-Su were used. The implied use per customer for these forecasts was calculated, except that the electric space and water heater end uses were deleted. This was done because almost no electric heating is found in the service territory, according to CVEA engineer, Mike Easley. The absolute use per customer in the service territory (~520 kWh/month) is consistent with this statement. The average change in use per customer (with electric space and water heating deleted) was -0.21% per year for Fairbanks, Mat-Su, and Kenai. This rate was used in the Mid forecast for CVEA. For the Low CVEA forecast -0.33% per year was used, and -0.08% per year was used in the High forecast. These Low and High values were developed by first calculating the standard deviation of the Fairbanks, Kenai, and Mat-Su use per customer growth rates. 1.5 standard deviations were subtracted and added to the average value to develop Low and High values. All of CVEA’s public building, public street lighting, small commercial, and large commercial uses (except industrial) were combined into one commercial electricity category. From this total commercial usage, a commercial use per residential customer was calculated. This commercial use per residential customer was then forecast as before by using the Fairbanks, Kenai, and Mat-Su end-use forecasts as a pattern. The Mid Case growth rate of commercial use per residential customer was calculated as -0.82% per year, and Low and High values of -1.08% per year and -0.56% per year were used. Three industrial loads were identified: Alyeska Main Line Refrigeration (MLR) Stations 1 and 2, Alyeska Pump Stations (PS) 11 and 12, and Fish Processing facilities. The 1988 figures represent the current consumption of those facilities. 120 kW of load is expected to be added to MLR2 at the beginning of 1990. The current load factor of MLR2 is about 58%. Assuming a similar load factor for the load addition, the addition amounts to a 0.6 GWh/year increase. Loads for PS11 and PS12 are projected to remain constant over the forecast period. Alyeska states that these loads will not change with decreased flow through the pipeline. The fish processing load is projected to increase from the current 1.0 GWh/year to 1.6 Gwh/year in 1991 (CVEA supplied the 1.6 GWh/year estimate). All industrial load projections in the forecast are the same in the Low, Mid, and High cases. BASE FORECAST: wae INPUTS: Initial Values: Residential Customers (000s) = Total Residential Sales (GWh) = Total Commercial Sales (GWh) = Residential Use/Customer Growth = LOW CASE *** 1.90 12.0 2525, -0.33% Commercial Use / Residential Customer Growth = -1.08% 1988 Residential Customers (000s) 1.90 kWh/month per Customer 528 Residential Use (GWh) 12.0 3 Commercial kWh/month per Residential Customer 1,120 1 Commercial Use (GWh) 25.5 Alyeska MR1 & MR2 (GWh) 3.50 Alyeska PS11 & PS12 (GWh) iee59) Fish Processing (GWh) 1.00 Total Industrial (GWh) 7.05 TOTAL SALES (GWh) 44.6 1989 1.90 526 12.0 1990 1991 1992 1.91 521 12.0 1.89 525 1.9 1.90 523 12.0 108 25.2 1,096 1,084 24.9 24.8 24.6 3.50 2.55 1.00 4.10 2.55 1.00 4.10 2.55 1.60 4.10 2.55 1.60 7.05 7.65 8.25 8.25 44.3 44.4 45.0 44.8 1993 1.92 520 12.0 1,072 1,060 24.4 4.10 2.55 1.60 8.25 44.6 1994 1995 1996 1.98 516 12.3 1.95 518 12.1 514 12.5 1,049 24.5 1,038 24.7 24.9 4.10 2.55 1.60 4.10 2.55 1.60 4.10 2.55 1.60 8.25 8.25 8.25 44.9 45.2 45.6 2.02 2.07 1997 513 12.7 1,026 1,015 25.2 4.10 2.55 1.60 8.25 46.1 1998 2.10 511 12.9 1,004 25.3 4.10 2.55 1.60 8.25 46.4 1999 2.15 509 13.1 994 25.6 4.10 2.55 1.60 8.25 47.0 2000 2.19 508 13.3 983 25.8 4.10 2.55 1.60 8.25 47.4 2001 2.21 506 13.4 972 25.8 4.10 2.55 1.60 8.25 47.5 2002 2.25 504 13.6 962 25.9 4.10 2.55 1.60 8.25 47.8 2003 2004 2.28 2.32 503 501 13.7 13.9 951 26.0 941 26.2 4.10 2.55 1.60 4.10 2.55 1.60 8.25 8.25 2005 2.36 499 14.1 931 26.4 4.10 2.55 1.60 8.25 921 26.7 4.10 2.55 1.60 8.25 48.0 48.3 48.8 49.3 2007 2.47 496 14.7 911 27.0 4.10 2.55 1.60 8.25 50.0 2008 2.52 494 14.9 901 27.2 891 27.4 4.10 2.55 1.60 4.10 2.55 1.60 8.25 8.25 50.4 50.8 882 27.5 4.10 2.55 1.60 8.25 51.1 LL BASE FORECAST: **k* MID CASE *** INPUTS: Initial Values: Residential Customers (000s) = 1.90 Total Residential Sales (GWh) = 12.0 Total Commercial Sales (GWh) = 25.5 Residential Use/Customer Growth = -0.21% Commercial Use / Residential Customer Growth = -0.82% 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Residential Customers (000s) 1.90 1.90 1.90 1.92 1.94 1.95 1.99 2.03 2.04 2.10 2.14 2.16 2.20 kWh/month per Customer 528 527 526 525 524 523 522 520 519 518 517 516 515 Residential Use (GWh) 12.0 12.0 12.0 12.1 12.2 12.2 12.5 12.7 12.7 13.1 13.3 13.4 13.6 Commercial kWh/month per Residential Customer 1,120 1,110 1,101 1,092 1,083 1,074 1,066 1,057 1,048 1,040 1,031 1,023 1,014 Commercial Use (GWh) 25.5 25.4 25.2 25.2 25.2 25.1 25.5 25.7 25.6 26.2 26.4 26.5 26.8 Alyeska MR1 & MR2 (GWh) 3.50 3.50 4.10 4.10 4.10 4.10 4.10 4.10 4.10 4.10 4.10 4.10 4.10 Alyeska PS11 & PS12 (GWh) 2.55 2.55 2.55 2.55 2.55 2.55 2.55 2.55 2.55 2.55 2.55 2.55 2.55 Fish Processing (GWh) 1.00 1.00 1.00 1.60 1.60 1.60 1.60 1.60 1.60 1.60 1.60 1.60 1.60 Total Industrial (GWh) 7.05 7.05 7.65 8.25 8.25 8.25 8.25 8.25 8.25 8.25 8.25 8.25 8.25 TOTAL SALES (GWh) 44.6 44.5 44.8 45.6 45.6 45.6 46.2 46.6 46.6 47.6 47.9 48.2 48.6 2001 1,006 27.0 4.10 2.55 1.60 8.25 49.1 2002 2003 2004 27.2 4.10 2.55 1.60 8.25 49.4 27.4 4.10 2.55 1.60 8.25 49.8 981 27.6 4.10 2.55 1.60 8.25 50.3 ors 2739 4.10 2.55 1.60 8.25 50.8 2006 2007 2.43 2.48 509 508 14.8 15.1 965 = 957 28.2 28.4 4.10 4.10 2.55 2.55 1.60 1.60 8.25 8.25 51.2 51.8 2008 950 28.6 4.10 2.55 1.60 8.25 52.2 2009 942 28.8 4.10 2.55 1.60 8.25 52.6 934 29.1 4.10 2.55 1.60 8.25 53.1 BASE FORECAST: INPUTS: Initial Values: Residential Customers (000s) = Total Residential Sales (GWh) = Total Commercial Sales (GWh) = Residential Use/Customer Growth = *** HIGH CASE EK 1.90 12.0 25.5 -0.08% Commercial Use / Residential Customer Growth = -0.56% 1988 Residential Customers (000s) 1.90 kWh/month per Customer 528 Residential Use (GWh) 12.0 = nN Commercial kWh/month per Residential Customer 1,120 Commercial Use (GWh) 25.5 Alyeska MR1 & MR2 (GWh) 3.50 Alyeska PS11 & PS12 (GWh) 2.55 Fish Processing (GWh) 1.00 Total Industrial (GWh) 7.05 TOTAL SALES (GWh) 44.6 1989 1990 1991 1.94 1.93 527 12.2 1.93 1.92 528 527 12.2 12.2 1,113 1,107 1,101 25.8 25.6 25.6 1,095 25.4 3.50 2.55 1.00 4.10 4.10 2.55 2.55 1.00 1.60 4.10 2.55 1.60 7.05 7.65 8.25 8.25 45.1 45.4 46.1 45.8 1,089 25.2 4.10 2.55 1.60 8.25 45.6 1,083 25.7 4.10 2.55 1.60 8.25 46.4 1995 1,076 25.9 4.10 2.55 1.60 8.25 46.8 1,070 1,064 26.1 4.10 2.55 1.60 8.25 47.2 26.7 4.10 2.55 1.60 8.25 48.1 1,058 27.0 4.10 2.55 1.60 8.25 48.7 1,053 27.4 4.10 2.55 1.60 8.25 49.3 2000 2001 1,047 1,041 28.0 28.4 4.10 2.55 1.60 4.10 2.55 1.60 8.25 8.25 50.2 51.0 1,035 28.8 4.10 2.55 1.60 8.25 51.7 2.44 522 15.3 2.59 521 16.2 1,029 1,023 30.1 4.10 2.55 1.60 8.25 53.6 31.8 4.10 2.55 1.60 8.25 56.3 2006 2007 2008 1,018 1,012 41.1 4.10 2.55 1.60 8.25 70.4 46.1 4.10 2.55 1.60 8.25 78.1 4.31 520 26.9 2.87 520 17.9 1,006 1,001 52.1 4.10 2.55 1.60 8.25 87.2 34.5 4.10 2.55 1.60 8.25 60.6 995 34.2 4.10 2.55 1.60 8.25 60.3 989 34.7 4.10 2.55 1.60 8.25 61.1 LOAD #4: CVEA Distribution System Expansion WHERE LOCATED: Chistochina, Lake Louise, Sheep Mountain, Misc. SERVICE CONTINGENT ON NE INTERTIE?: No INCLUDED IN WHICH CASES?: Low, Mid, and High The cost of the distribution expansion to the first three areas is approximately $3.3 million. CVEA has applied for state grants to fund the expansion, but they have not been approved as of yet. The annual cost of the distribution system (capital levelized at 4.5% real for 30 years, O&M assumed to be 1.5% of capital cost per year) is about 35 cents per kWh of load supplied. Because the areas currently rely on small diesel generators, the cost of generation is relatively expensive--however, possibly less than 35 cents/kWh + CVEA generation cost. The expansion to these particular areas may not be that probable, but other miscellaneous expansion of the CVEA system may occur. Because of this not-easily-identified miscellaneous expansion, we include the load in all three cases. CALCULATIONS: CVEA estimates the Chistochina and Lake Louise loads to be about 0.6 GWh/year. We add 0.1 GWh/year for Sheep Mountain and other miscellaneous loads. LOAD #5, #6, #7: US Air Force Backscatter Radar Installation - Industrial Load The US Air Force is currently in the process of designing a Backscatter Radar transmitter facility north of Glennallen. There is a possibility that some of the electrical requirements of the facility will be met from the electrical grid as opposed to onsite generation. Load components 5, 6, and 7 represent 3 possibilities as to the amount electricity supplied from electrical grid. The load impact due to the construction of the facility and the increased population from additional jobs is covered in Load Component #8. The power supply for the project was put out to competitive bid by the Air Force. Copper Valley Electric has bid on the project as well as ~4 other private firms. The winner of the power supply bid will be decided on March 24, 1989. The uncertainty in the estimate of the power draw from the grid will be narrowed substantially when the winner is known. WHERE LOCATED: North of Glennallen SERVICE CONTINGENT ON NE INTERTIE?: Maybe. INCLUDED IN WHICH CASES?: #5 - Low, #6 - Mid, #7 - High The power supply for the project was put out to bid. If Copper Valley wins the bid, they will supply some of the needs of the Backscatter facility from their grid even if the NE Intertie is not built. If another bidder wins the contract, it is unlikely that local CVEA generation will be priced low enough (with CVEA margins) to induce a contract (unless a special deal for Solomon spill power can be arranged). If a NE intertie were built, lower 13 cost power would be available either directly from southern utilities or through CVEA. CALCULATIONS: Two transmitters will be located at the site. Both transmitters will typically be operating concurrently. Each transmitter has two modes of operation, and depending on the operation modes of the transmitters, the total transmitter load will vary from 4.8 to 7.2 MW.’ Other miscellaneous equipment at the facility will add 0.1 - 0.2 MW. In addition, a refrigeration system may be required to the stabilize the ground underneath the transmitting antennae. Frank Moolin and Associates of Anchorage are designing the refrigeration system. Joe Perkins states that the design of the system is not complete because geotechnical testing at the site has not been completed. The site may require from 0 to 1 MW (peak) of refrigeration depending on the results of the testing and the design approach taken. He estimates the load factor the refrigeration system to be 15%. Stan Lawrence of the USAF at Elmendorf has been involved in the review of the power supply bids. They have forwarded their selection of a contractor to national offices for review and a decision. He was unable to say who their preliminary selection was. However, he did provide information on the likelihood of the use of grid-generated electricity. Strict reliability requirements were written into the bid specifications with large monetary penalties for non-compliance. Thus, all bidders will build enough generation and reserve on-site to meet the full load. Any electricity drawn from a grid would be for the purposes of avoiding variable fuel and O&M costs and possibly providing some additional "stiffness" to the generation supply system. However, different designs have differing abilities to reliably take electricity from the grid, according to Lawrence. Sufficient spinning reserve must be available to compensate for the loss of the grid. Also, the design must be able to accommodate rapid changes in load from the radar facility. The facility load will occasionally vary 3-4 MW ina matter of seconds. When asked for a range of estimates as to the amount that various designs could reliably take from the grid on a continual basis, Lawrence estimated 1 to 3 MW. He could conceive of some steam turbine designs taking as much as 5 MW. In addition, refrigeration and miscellaneous loads could be supplied from the grid. Doug Bursey, CVEA general manager, states that their design would take 3 MW at roughly 90% load factor from the grid. In addition they could serve all of the refrigeration load from the grid. For the Low, Mid, and High estimates, 0, 2.7, and 3.7 MW average load from the grid were chosen. Because of the $1 - $3 million capital cost associated with connecting to the grid, the less favorable purchase prices associated with low purchase quantities, and design constraints, it is quite possible that some private contractors will choose to take no power from the grid. The radar is expected to start partial operation on January 1, 1992 and full operation six months later. ‘Information supplied by Stan Lawrence, USAF Elmendorf (552-5185), and Roland Robb of General Electric (315-456-7007), the supplier of the radar equipment. 14 LOAD #8: USAF Backscatter Radar - Construction and Employment Impact This load component covers the electricity used during the construction of the Backscatter Radar project. It also accounts for the electricity associated with the increased population induced from Backscatter Radar employment. WHERE LOCATED: Glennallen SERVICE CONTINGENT ON NE INTERTIE?: No INCLUDED IN WHICH CASES?: Low, Mid, and High CALCULATIONS: The two worksheets on the following pages calculate the electrical load effects of Backscatter radar project aside from its potential direct draw from the utility grid. The first worksheet estimates the electrical load impact during construction of the project. It is assumed that CVEA’s distribution line, located 0.5 mile from the site, will be used during the construction phase. The line is capable of supplying about 500 kW according to Bursey. At a 60% load factor the line could supply 2.6 GWh/year. We estimate that this would be sufficient for most of the construction needs. The construction electricity estimate is developed as follows. The number of jobs are estimated during the construction period and listed in the "Primary Workers" row of the — spreadsheet. These estimates were provided by Stan Lawrence of the USAF and include jobs associated with the construction of the radar facility and the associated power plant. The primary workers spend money in the local economy creating secondary jobs. Typically, the spending creates another 0.5 jobs for every primary job. However, given the temporary nature of the work we estimate the local spending effect to create 0.3 secondary jobs. "Primary Workers" and "Secondary Workers" are combined into "Total Workers". "Total Workers" are converted into households by assuming 2 Workers per household. The average for the Valdez/Cordova census area is 1.67 workers per household. Because temporary workers bring fewer dependents, and the project will cause a slight increase in the labor force participation rate for existing residents, the value of 2 workers per household was used for the incremental effect. The residential use per household values were taken from the Mid Case of the Base Forecast (Component #2). Residential use was calculated from these use per household figures and the induced number of households. Commercial use was determined by estimating the amount of electricity used by each primary worker and the amount used by each secondary worker. The current CVEA average is about 670 kWh/month per worker. This figure is high relative to Fairbanks (480 kWh/month), Kenai (346 kWh/month), and Anchorage/Mat-Su (600 kWh/month). Doug Bursey, the CVEA general manager, suggests that this is due to a high tourism component in their economy. For example, the electricity use of a hotel per worker is high. Since the jobs induced by the construction of the radar project will not be tourism jobs, we used a lower figure of 500 kWh/month. This figure was used for both primary and secondary 1S workers. It was projected to change over time by -0.82% per year, the growth rate of ~ commercial use per household in the Mid case of the Base Forecast. The second worksheet estimates the electrical load impact from the long-term employment generated by the project. The USAF estimates that the operations phase of the radar will involve 60 jobs. Secondary jobs are estimated at 0.5 secondary jobs per primary job. Total workers were converted to households by using the average for the area, 1.67 workers per household. The rest of the calculation parallels the previous worksheet except no commercial electricity use is attributed to the primary workers. The electricity they use at their job is accounted for in previous load components, Components 5, 6, and 7. 16 Zt USAF Backscatter Radar Construction Phase Electrical Load Impact INPUTS: Secondary Workers / Primary Workers = 0.3 Workers / Household 2.00 Initial kWh/month per Household = 528 kWh/month Growth Rate = -0.21% per year Initial kWh/month per Primary Worker = 500 kWh/month Growth Rate = -0.82% per year Initial kWh/month per Secondary Worker = 500 kWh/month Growth Rate = -0.82% per year 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Primary Workers Secondary Workers Total Workers Households 0 0 65 130 65 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 kWh/month per Household 528 527 526 525 524 522 521 520 519 518 517 516 515 514 513 512 511 509 508 507 506 505 504 Residential Use (GWh) 0.0 0.0 0.4 0.8 0.44 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Commercial kWh/month per Primary Worker 500 496 492 488 484 480 476 472 468 464 460 457 453 449 446 442 438 435 431 428 424 421 417 Commercial kWh/month per ; Secondary Worker 500 496 492 488 484 480 476 472 468 464 460 457 453 449 446 442 438 435 431 428 424 421 417 Commercial Use (GWh) 0.0 0.0 0.8 1.5 0.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total Use (GWh) 0.0 0.0 1.2 2.3 1.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 8L USAF Backscatter Radar Operations Phase Electrical Load Impact INPUTS: Secondary Workers / Primary Workers = 0.5 Workers / Household 1.67 Initial kWh/month per Household = 528 kWh/month Growth Rate = -0.21% per year 0 kWh/month -0.82% per year Initial kWh/month per Primary Worker Growth Rate = Initial kWh/month per Secondary Worker = Growth Rate = 500 kWh/month -0.82% per year 1988 1989 1990 1991 1992 Primary Workers Secondary Workers Total Workers Households 0 0 0 0 54 kWh/month per Household 528 527 526 525 524 Residential Use (GWh) 0.0 0.0 0.0 0.0 0.3 Commercial kWh/month per Primary Worker 0 0 0 0 0 Commercial kWh/month per Secondary Worker 500 496 492 488 484 Commercial Use (GWh) 0.0 0.0 0.0 0.0 0.2 Total Use (GWh) 0.0 0.0 0.0 0.0 0.5 1993 54 522 0.3 480 0.2 0.5 1994 54 521 0.3 476 0.2 0.5 54 520 0.3 472 0.2 0.5 1996 468 0.2 0.5 1997 1998 464 0.2 0.5 54 517 0.3 460 0.2 0.5 1999 2000 2001 457 0.2 0.5 515 0.3 453 0.2 0.5 514 0.3 449 0.2 0.5 2002 2003 2004 446 0.2 0.5 54 512 0.3 442 0.2 0.5 511 0.3 438 0.2 0.5 2005 2006 2007 2008 54 54 54 54 509 508 507 506 03 0.5 0.3 0.3 0 0 0 0 435 431 428 «= 424 0.2 0.2 0.2 0.2 O05 0.5 0.5 0.5 54 505 0.3 421 0.2 0.5 54 504 0.3 417 0.2 oss LOAD #9: Alaska Pacific Refining Inc. (APRI) Refinery - Industrial Load APRI is proposing to build a 120,000 barrel per day refinery in Valdez costing about $800 million. This load component is the electrical requirements of the refinery itself. The refinery would sell refined products to the U.S. and overseas. Maritime interest introduced federal legislation to block the sale of refined products overseas, but the legislation was defeated. APRI thinks that maritime interests will try further to block their main market. APRI has stopped the engineering of the project temporarily. They recently purchased a refinery in Los Angeles and are waiting until profits from that project generate cash for the Valdez project before proceeding. WHERE LOCATED: Valdez SERVICE CONTINGENT ON NE INTERTIE?: No INCLUDED IN WHICH CASES?: None The refinery load is included in no cases because APRI has stated that they would generate their own power from refinery gas (methane, ethane, and carbon monoxide). CALCULATIONS: The refinery needs are estimated at 40 - 50 MW by APRI. 45 MW with a 70% load factor assumption implies a 276 GWh/year use, approximately 6 times the entire 1988 CVEA load. APRI also states that they would like to sell electricity to CVEA if the refinery is built. They would have 30 MW of generating capacity available for sales to the grid (total installed capacity of 105 MW). No estimate of the production cost of APRI’s electricity could be developed. LOAD #10: APRI Refinery - Construction and Employment Impact This load component accounts for electrical load impacts of the APRI refinery other than its direct use of electricity. Electricity used during construction of the facility is accounted for, and electricity use due to the population increase in the Valdez area is accounted for. WHERE LOCATED: Valdez SERVICE CONTINGENT ON NE INTERTIE?: No INCLUDED IN WHICH CASES?: High CALCULATIONS: The calculations use the same model and basic assumptions as used for Load Component #8, the USAF Backscatter Radar project. The following two worksheets detail the 19 calculations. APRI estimates that the refinery will open in mid-1992. We use a later opening date of early 1996. APRI estimates the main construction period to be 18 months long with an average of 2000 workers on-site (converted to 1500 workers over a two year period for the purposes of analysis). Some civil work will occur prior to the peak construction period, and it is estimated to create 300 jobs. We make the assumption that CVEA can supply the bulk of the on-site electricity needs. APRI stated that they would purchase power from CVEA for construction purposes if it were available. The operations phase of the refinery will require 275 workers according to APRI. The effect of this employment on electrical load is estimated by the second worksheet. 20 LZ APRI Refinery Construction Phase Electrical Load Impact INPUTS: Secondary Workers / Primary Workers = (ke § Workers / Household 2.0 Initial kWh/month per Household = 528 kWh/month Growth Rate = -0.21% per year Initial kWh/month per Primary Worker = 500 kWh/month Growth Rate = -0.82% per year Initial kWh/month per Secondary Worker = 500 kWh/month Growth Rate = -0.82% per year 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Primary Workers 0 0 0 0 0 300 1500 1500 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Secondary Workers 0 0 0 0 0 90 450 450 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Total Workers 0 0 0 0 0 390 1950 1950 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Households 0 0 0 0 0) 1195) | 975.) 975 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 kWh/month per Household 528 527 526 525 524 522 521 520 519 518 517 516 515 514 513 512 511 509 508 507 506 505 504 Residential Use (GWh) 0.0 0.0 0.0 0.0 0.0 1.2 6.1 6.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Commercial kWh/month per Primary Worker 500 496 492 488 484 480 476 472 468 464 460 457 453 449 446 442 438 435 431 428 424 421 417 Commercial kWh/month per : Secondary Worker 500 496 492 488 484 480 476 472 468 464 460 457 453 449 446 442 438 435 431 428 424 421 417 Commercial Use (GWh) 0.0 0.0 0.0 0.0 0.0 2.2 11.1 11.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total Use (GWh) 0.0 0.0 0.0 0.0 0.0 3.5 17.2 17.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0, 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ec APRI Refinery Operations Phase Electrical Load Impact INPUTS: Secondary Workers / Primary Workers = 0.5 Workers / Household 1.67 Initial kWh/month per Household = 528 kWh/month Growth Rate = -0.21% per year Initial kWh/month per Primary Worker = 0 kWh/month Growth Rate = -0.82% per year Initial kWh/month per Secondary Worker = 500 kWh/month Growth Rate = -0.82% per year 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Primary Workers 0 0 0 0 0 0 0 0° 275) |) (275) _ 275) (275) 275 | 275) 275) | 1205 | 275) (275) (2f> | 205 | 2%) V2r5 | 275 Secondary Workers 0 0 0 0 0 0 0 0 138 «#138 «#6138 «©6138 «©1138 «©6138 «©9138 «©6138 «8138 «= 138 = 138 S138) 138) 138 = 138 Total Workers 0 0 0 0 0 0 0 O 413 413 413 413 413 413 413) 413 413 413 413) 413) 413) 413413 Households 0 0 0 0 0 0 0 0 2467 267 267 247 2467 2467 2467 2467 247 «247 = 247) 247-247) KT 47 - kWh/month per Household 528 527 526 525 524 522 521 520 519 518 S17 516 515 514 513 512 511 509 508 507 506 505 504 Residential Use (GWh) 0:0) 0:0; 0:0) 0:0' 0:0 0/0 (0.0) 0:0 1-5: 4-5) 45° 4:5 }1.5 455) 4-5 | 1.5) 3-5) 9U-5 125) 1-5 | 1-5) 1-5 9 1-5 Commercial kWh/month per Primary Worker 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Commercial kWh/month per Secondary Worker 500 496 492 488 484 480 476 472 468 464 460 457 453 449 446 442 438 435 431 428 424 421 417 Commercial Use (GWh) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.8 0.8 0.8 0.8 0.7 0.7 O.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 Total Use (GWh) (0.0) |0-0) | (0.0;) (020) | (0/0) jo%0) | 0:0) (0-0 | 12-3)) (223) [/72:3)) (223) |.2:3) | (2-3') 2-3 | 222) (252) \i2.2)) (2-2) | 2-2) 222) \i2.2) } 2.2 LOAD #11: Alyeska Pump Station #10 WHERE LOCATED: Between Glennallen and Delta Junction SERVICE CONTINGENT ON NE INTERTIE?: Yes. The current CVEA distribution system does not extend to the pump station but the NE intertie would pass nearby. CVEA estimates that the capital cost of serving the pump station will be about $1.5 million (a 5 MVA transformer would be required to serve for the load). The levelized cost of the transformer (4.5% real discount rate, 30 year life, 1.5%/yr maintenance) would be about 2.9 cents/kWh. INCLUDED IN WHICH CASES?: Mid, and High Although CVEA currently serves Pump Station 11 and 12, service to Pump Station 10 is not as certain. Pump Station 10 produces fuel for itself and other Pump Stations. Therefore, reliability needs are higher because of the value of the functions being performed at the station. Also, the effective cost of fuel for electrical generation at Pump Station 10 is lower than non-fuel-producing pump stations, because there is no transportation cost. Fuel is produced on-site. The transformer necessary to serve the load would cost about 3 cents/kWh (levelized real), also detracting from the economics of power purchase. CALCULATIONS: Engineer Bill Frichtl of Alyeska estimates the average load at Pump Station 10 at 900 kW. However, Doug Bursey of CVEA says that they would only serve about half the load because of Alyeska’s reliability concerns. A 450 kW average load amounts to 3.9 GWh/year. The load factor is estimated to be 90%. Since the intertie is not expected to be complete until 1994, we show the pump station load as starting in 1995. LOAD #12: Paxson and Summit Lake WHERE LOCATED: Between Glennallen and Delta Junction SERVICE CONTINGENT ON NE INTERTIE?: Yes. A Potential Voltage Transformer would be used to serve the load at a cost of $0.15 million. The levelized cost per kWh of the transformer is about 6 cents/kWh. INCLUDED IN WHICH CASES?: Low, Mid, and High CVEA generation cost plus the 6 cent/kWh cost of the transformer is probably less than the cost of the generation from the small diesel generators serving this area now. CALCULATIONS: The 1986 load for Paxson was about 0.2 GWh. The estimate is escalated over time by using the Mid case of the Base Forecast as a pattern. 23 LOAD #13: Tok Some have suggested that if the NE intertie is built, an additional line may be built from Delta Junction to Tok. According to CVEA general manager Bursey there is currently a capacity constraint on the line from Fairbanks to Delta Junction, hindering the construction of a line from Delta Junction to Tok. The NE intertie would eliminate this constraint. WHERE LOCATED: Tok SERVICE CONTINGENT ON NE INTERTIE?: Yes. The estimated cost of a line from Delta Junction to Tok is about $11 million ($80,000/mile for 130 miles). The levelized cost per unit of load supplied is about 10 cents/kWh. INCLUDED IN WHICH CASES?: None. The differential in cost between Tok generation ($0.76/gallon fuel, 13.7 kWh of generation per gallon --> 5.5 cents/kWh fuel cost + ~1.5 cents/kWh variable O&M = 7.0 cents/kWh) and intertie electricity (~3 cent/kWh) is not sufficient to justify the cost of the line. If this load were included in the High case, it would probably penalize the economics of the NE intertie. The cost savings from serving the load would be less than the $11 million construction cost of the Delta-Tok line. Therefore, it was not included in any cases. CALCULATIONS: The FY88 load for Tok was 7.2 GWh. With some load growth and additional loads along the route, the estimate for 1995 (first year of the intertie) is 8.0 GWh. 24