Loading...
HomeMy WebLinkAboutwind biodiesl co-genMcMinnville Electric System i Agricultural Bio -Fueled Generation of Electricity and Development of Durable and Efficient NOx Reduction Project Title: Agricultural Bio-Fueled Generation of Electricity and Development of Durable and Efficient NOx Reduction Covering Period: February 1, 2004 through May 1, 2007 Date of Report: April 30, 2007 Recipient: McMinnville Electric System 200 Morford Street McMinnville, TN 37111-0608 Award Number: DE-FG36-04GO14250 Subcontractors: EmeraChem Steve DeCicco – Vice President of Operations – (865) 246-3000 Lisa Mitchell – Project Engineer – (865) 246-3000 Dr. Albin Chernichowski – GlidArc Technologies - +33-680-232-643 Stowers/Caterpillar Tom Stanzione – Manager, Distributed Generation – (865) 675-2869 Dave Martin – Power Generation Project Manager – (865) 546-1414 Matt Kirkpatrick – Commercial Engine Sales – (865) 546-1414 Chris Kiczaja – Sales Manager Electric Power Generation – (615) 341-3215 Tennessee Valley Authority Ralph Boroughs – Project Manager – (423) 751-4644 Chevales Ward – Environmental Engineer – (423) 751-7316 Duane Brigman – Environmental Scientist – (423) 876-4202 McMinnville Electric System ii Other Partners: American Public Power Association – (202) 467-2900 National BioDiesel Board – (800) 841-5849 Tennessee Soybean Promotion Board – (931) 668-2772 Helen Hennon – Director of Governmental Services, Quantum Environmental & Engineering Services, LLC – (865) 689-1395 Ryan Strickland – COO, Agri-Energy Management (931) 270-8129 Kelly Strebig – University of Minnesota Center for Diesel Research - (651) 330-0450 Will Ayers – V.P. Engineering, Cim-Tech Filtration – (217) 678-2511 Dr. Thomas Reed – Biomass Energy Foundation – (303) 279-3707 William Ayres – AgBioEnergy, LLC – (913) 341-7114 Dave Brown – Phillips Sales and Service – (931) 473-2450 Project Contact(s): Technical: MES Manager, Engineering and Operations – Ralph Dunn, (931) 473-3144, rdunn@mesystem.net Business: MES Manager, Finance and Accounting/Assistant General Manager – Neal Cox, (931) 473-3144, ncox@mesystem.net McMinnville Electric Project Team: MES General Manager – Rodney Boyd, (931) 473-3144, rboyd@mesystem.net MES Manager, Finance and Accounting/Assistant General Manager – Neal Cox, (931) 473-3144, ncox@mesystem.net MES Manager, Engineering and Operations – Ralph Dunn, (931) 473-3144, rdunn@mesystem.net MES Electrical Engineer – Huel Martin, (931) 473-3144. hmartin@mesystem.net DOE Project Team: DOE-HQ contact – Valerie-Sarisky Reed, 202-586-1507 DOE Field Project Officer – Fred Gerdeman, 303 275-4928, fred.gerdeman@go.doe.gov McMinnville Electric System iii DOE Contract Specialist – Pamela Brodie, 303-275-4741, pamela.brodie@go.doe.gov DOE Project Monitor – Renae Binstock, 303-275-4772 , renae.binstock@go.doe.gov McMinnville Electric System iv Summary Abstract Project Title: Agricultural Bio-Fueled Generation of Electricity and Development of Durable and Efficient NOx Reduction Objectives: The objective of this project was to define the scope and cost of a technology research and development program that will demonstrate the feasibility of using an off-the-shelf, unmodified, large bore diesel powered generator in a grid-connected application, utilizing various blends of BioDiesel as fuel. Furthermore, the objective of project was to develop an emissions control device that uses a catalytic process and BioDiesel (without the presence of Ammonia or Urea)to reduce NOx and other pollutants present in a reciprocating engine exhaust stream with the goal of redefining the highest emission reduction efficiencies possible for a diesel reciprocating generator. Process: Caterpillar Power Generation adapted an off-the-shelf Diesel Generator to run on BioDiesel and various Petroleum Diesel/BioDiesel blends. EmeraChem developed and installed an exhaust gas cleanup system to reduce NOx, SOx, volatile organics, and particulates. The system design and function was optimized for emissions reduction with results in the 90-95% range; especially for NOx. TVA measured the emissions and reviewed the environmental effects. Outcome to Date: Objective #1 - using a 3516B Caterpillar generator, McMinnville Electric System has successfully generated 1,629,024 kWh’s of renewable electric power (1008 hours of operation) using soybean based, American made BioDiesel as fuel. In the process McMinnville Electric System has used 126,126 gallons of BioDiesel which equates to 84,080 bushels of soybeans. After examining the internal rotating engine parts and combustion chamber, Caterpillar engineers report the “Test was very Successful”. (See attached report BioDiesel Demonstration with SCONOx NOx Removal, Attachment “C”) Objective #2 - we were able to achieve a 96.6% reduction in NOx without the use of Ammonia or Urea as reductants utilizing EmeraChem’s exhaust gas cleanup system and BioDiesel as a reductant. The NOx emission reduction results were independently measured and verified by the Tennessee Valley Authority. (See attached report BioDiesel Demonstration with SCONOx NOx Removal, Attachment “G”) McMinnville Electric System v Diagram of the Project: Equipment: 3516BDITA Caterpillar Generator, EmeraChem EMx Prototype Emissions System, 30,000 gallon fuel tank, 2000 kVA power transformer, EMCP II+ Control Panel, NexGear Series 1 Advanced Paralleling Switchgear, PointGuard on-site remote-monitoring hardware. What we have learned: We have success proven that a large-bore stationary diesel generator can utilize 99.9% Biodiesel as fuel for a prolonged period of time, that the BioDiesel has no effect on engine durability and performance and that BioDiesel can be successfully substituted for petroleum diesel in warmer climates. We have successfully proven that NOx emissions can be reduced by > 96.6% in a large bore stationary diesel engine without the use of ammonia or urea as reductants by reforming BioDiesel into hydrogen. We have learned that you can produce >21% Hydrogen by reforming BioDiesel. We also have learned that underground mining operations and generators in non- attainment areas could benefit from the experience gained from this report. We have not been successful in operating the BioDiesel to hydrogen reformer for extended periods of time (beyond 30 hours) without operational issues. For this type of project to be successful, the reformer would need to operate for at least 200-400 hours without maintenance, breakdown or failure. Much research is being done in this area and it is our intent McMinnville Electric System vi to continue our research until we have successfully overcome this obstacle. Much more detail and additional research findings can be found in the attached report entitled BioDiesel Demonstration with SCONOx NOx Removal. History: This project has been funded by the Department of Energy as a Congressionally Directed Project. Additional funds were provided by the American Public Power Association, National BioDiesel Board and the Tennessee Soybean Promotion Board. In-kind help was provided by EmeraChem, Stowers Caterpillar and the Tennessee Valley Authority. Our original planned start date was February 02, 2004 with an original planned completion date of August 1, 2005. Circumstances beyond our expectations, including cold weather and total destruction of the original BioDiesel hydrogen reformer necessitated MES requesting and DOE graciously granting two (2) no-cost extensions which resulted in an actual completion date of April 30, 2007. Disclaimer: “This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect these of the United States Government or any agency thereof.” FOR A MORE COMPLETE EXPLINATION OF THE FINDINGS CONTAINED IN THIS EXECUTIVE SUMMARY AND ADDITIONAL INFORMATION ABOUT THE PROJECT, SEE THE ATTACHED REPORT BioDiesel Demonstration with SCONOx NOx Removal. BioDiesel Demonstration with SCONOx NOx Removal Award No. DE-FG36-04GO14250 Prepared for: U.S. Department of Energy Golden Field Office 1617 Cole Boulevard Golden, Colorado 80401-3393 Prepared by: McMinnville Electric System 200 Morford Street McMinnville, Tennessee 37110 April 30, 2007 McMinnville Electric System ii Contents Project Title: .................................................................................................................................... 1 General Overview: .......................................................................................................................... 1 Purpose: .......................................................................................................................................... 1 Utility Name and Address: .............................................................................................................. 2 Other Participants: .......................................................................................................................... 2 Utility Description: .......................................................................................................................... 2 Key Personnel and Phone Numbers: .............................................................................................. 3 Description: ..................................................................................................................................... 3 Diagram of the Project: ................................................................................................................... 4 Dates: .............................................................................................................................................. 4 Alternatives: .................................................................................................................................... 4 Results to Date: ............................................................................................................................... 5 Status: ............................................................................................................................................. 5 Applicability: ................................................................................................................................... 5 Future Plans: ................................................................................................................................... 5 Equipment: ...................................................................................................................................... 6 Performance: .................................................................................................................................. 6 Injector Analysis after 500 Hours .............................................................................................. 11 Injector Analysis after 1007 Hours of Operations .................................................................... 15 Budget: .......................................................................................................................................... 19 Additional Notes: .......................................................................................................................... 19 References: ................................................................................................................................... 19 Attachment “A” ............................................................................................................................. 21 Attachment “B” ............................................................................................................................. 32 Introduction .................................................................................................................................. 33 Demonstration Goals ................................................................................................................ 33 Background ............................................................................................................................... 33 Experiment Design ........................................................................................................................ 33 Engine Test Platform ................................................................................................................. 33 Fuel and Lubricants ................................................................................................................... 34 Testing Protocol ........................................................................................................................ 34 Catalyst System ......................................................................................................................... 34 Catalysts .................................................................................................................................... 35 Regen System ............................................................................................................................ 36 Data Acquisition ........................................................................................................................ 39 Original Test Plan ...................................................................................................................... 39 Installation and Commissioning .................................................................................................... 40 Short Term Fuel Blend Testing ...................................................................................................... 47 McMinnville Electric System iii ULSD Test .................................................................................................................................. 47 B2 Test ....................................................................................................................................... 49 B5 Test ....................................................................................................................................... 51 Observations on ULSD, B2 and B5 Short Term Fuel Blend Tests .............................................. 52 B20 Test ..................................................................................................................................... 55 B50 Test ..................................................................................................................................... 58 B100 Test................................................................................................................................... 60 Fuel Affects ............................................................................................................................... 62 Long Term Test .............................................................................................................................. 65 Limiting Factors ......................................................................................................................... 65 Conclusions ................................................................................................................................... 72 Summary of Results .................................................................................................................. 72 Future Work .............................................................................................................................. 73 Attachment “C” ............................................................................................................................. 74 Attachment “D” ............................................................................................................................ 88 Attachment “E” ............................................................................................................................. 91 Attachment “F” ............................................................................................................................. 96 Attachment “G” .......................................................................................................................... 167 TVA Activities in Support of the McMinnville BioDiesel / SCONOX Project ........................... 168 Introduction ......................................................................................................................... 168 Test Plan .............................................................................................................................. 168 Emissions Test Equipment ................................................................................................... 169 Results ..................................................................................................................................... 171 Baseline Tests ...................................................................................................................... 171 B2 BioDiesel Blend ............................................................................................................... 173 B5 BioDiesel Blend ............................................................................................................... 174 B20 BioDiesel Blend Testing ................................................................................................ 175 B50 BioDiesel Blend Testing ................................................................................................ 176 B100 BioDiesel Testing ........................................................................................................ 177 Fuel Characterization Tests ................................................................................................. 179 McMinnville Electric System iv Disclaimer The purpose of this Report is to share information about the use of BioDiesel, and a prototype NOx emissions reduction system, with all interested parties. It is furnished with the understanding that McMinnville Electric System, the City of McMinnville, Tennessee, and McMinnville Electric System’s provider of power, and their respective directors, officers, agents, representatives, assigns, subcontractors, suppliers, and employees, and the McMinnville Electric System Board of Public Utilities, shall not be held liable for any claims, demands, causes of action, costs, or losses for personal injuries, property damage, or loss of life or property, arising out of or in any way connected with the testing or operation of a similar or similar Project(s), including claims based upon Breach of Contract, Breach of Agreement, Breach of Warranty, strict liability or negligence, or any other loss, damage, or injury caused by or relating to the design, manufacture, selection, delivery, condition, operation, use, maintenance or repair of a similar or similar Project(s). UNDER NO CONDITION OR CAUSE OF ACTION SHALL MCMINNVILLE ELECTRIC SYSTEM BE LIABLE FOR ANY LOSS OF ACTUAL OR ANTICIPATED BUSINESS OR PROFITS OR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES. Neither MES, the City of McMinnville, Tennessee, MES’S provider of power, nor their respective directors, officers, agents, representatives, assigns, employees, subcontractors, nor suppliers, nor McMinnville Electric System Board of Public Utilities, shall be liable for any direct, indirect, general, special, incidental, exemplary, or consequential loss or damage of any nature, including loss of life or injury, arising out of their performance or non-performance of the information provided hereunder. The provisions of this Section shall apply whether such liability rises in contract, agreement, tort (including negligence), strict liability or otherwise. “This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect these of the United States Government or any agency thereof.” McMinnville Electric System v Acknowledgements A project of this magnitude cannot be undertaken without the help of many individuals to whom I would like to express my sincere thanks. The McMinnville Electric System Board of Public Utilities is the best utility board that a utility manager could ask for. They are among the rare leaders of an organization that will allow for and encourage “out-of-the-box” thinking. They have held me accountable throughout this Project, not in a critical way, but as an encouragement. They believe that we must protect our environment for future generations and that reliance on foreign governments to supply our fuel needs is questionable judgment. They are the true leaders of this project. Jeff Golden (chairman), Sam Martin (vice-chairman and R&D Committee member), Jeff McKinley (R&D Committee chairman), Bobby Kirby (aldermanic representative and R&D Committee member) and Sally Brock (board member) should be thanked by our community for the commitment they, individually and as a Board, have to this organization. Robert Newman (general counsel) has also been an invaluable resource when it came to negotiating Purchase Power Agreements and Parallel Operating Agreements with TVA. Congressman Lincoln Davis cannot be thanked enough for his sponsorship of this Project and his interest during the entire process. He and his staff (Beecher Fraser, Brandi Lowell and Cicely Simpson) have been fantastic to work with, during the last two years. Congressman Davis has a commitment to American farmers that is second to none in Congress. To me, he is an American hero in our Nation’s capitol. Tennessee owes him many, many thanks! Steve DeCicco and Lisa Mitchell (EmeraChem) have been by my side through the good times and bad. We have both raised our hands in victory during this Project and suffered agonizing moments when it appeared the obstacle before us was insurmountable – many times in the same day – sometimes within the same hour. Had it not been for their spirit of adventure and thirst for inventiveness, this Project would not have achieved the success that we have enjoyed. To them, a place in my heart and life will always abide. I would also like to thank Tom Girdlestone (EmeraChem) for allowing Steve and Lisa to continue to work on this project and to see it through to success. Ralph Dunn and Huel Martin (McMinnville Electric System) have been the hands and engineering minds of this Project since it was first conceived back in 2001. They have literally worked day and night to make it successful; and without them beside me helping, and behind me pushing, I could not have made this Project a reality. My substation maintenance crew (Jonathan Womack, Troy Sauls and Jeremy Womack) worked throughout the night on many occasions to wash the catalyst and have them ready to reinstall and operate the next morning. The McMinnville Electric System line crew (Billy Pitmon, Gene Rhea, Tony Foster, Frankie Rains, Phillip Rowland, Craig Foster, Dwight Jones, Neal Smith and Jonathan Jacobs) installed the generator, all of the switchgear and wiring. What a remarkable group of craftsmen. McMinnville Electric System vi Neal Cox (McMinnville Electric System) has been the finance and accounting mind behind the Project. Neal has worked tirelessly behind the scenes to make sure all of the accounting “i’s” are dotted and “t’s” are crossed and has been responsible for compiling the mounds of paperwork that must be supported during the Project. Neal is also a diligent champion of alternative fuels and believes that helping American farmers is in our nation’s best interest. Dave Brown and Joey Wilcher (Phillips Sales and Service) were the brawn and brains for the fuel delivery and filtration system. You can have the best fuel and the best storage facility, but if you can’t get the fuel from the tank to the engine, you don’t have anything. They were available at a moment’s notice, 24-hours a day. When we needed help – anytime – they were there quickly. Their help and guidance through all the fuel blend testing was immeasurable – I give them thanks. Ralph Boroughs (TVA) was a skeptic at first and later one of my most wholehearted supporters. His first and foremost objective is environmental protection; and at that, he is great. The first time Ralph and I met, you could tell from his non-verbal communication that there was no way that he was going to support a project like this. After he found out our goal was >90% NOx reduction, we had his attention and his arms-crossed demeanor changed as well The day we achieved the 96.6% NOx reduction target, he was the one with the biggest smile on his face. Ralph Boroughs is a great guy and a good friend. Chevales Ward and Duane Brigman (TVA) were also of utmost help to the Project. During one particular emissions test, it was raining strait down and no one would have faulted Chevales and Duane for stopping the test and coming in out of the rain; however, these two committed environmental champions worked throughout the rain storm because they knew how tight our testing schedule was and that stopping at that point would have meant a week’s setback. Thank you Chavales and Duane for a moment that I will not soon forget. Others I would like to thank include Parks Wells (Tennessee Soybean Promotion Board), Tom Stanzione (Stowers/Caterpillar), Dave Martin (Stowers Caterpillar), Kelly Strebig (University of Minnesota Center for Diesel Research), Helen Hennon (Quantum Environmental and Engineering), Jonathan Overly (East Tennessee Clean Fuels Coalition), Ryan Strickland (Agri-Energy), Bill Ayers (AgBioEnergy), Gary Ewing (SECOR International), Mike Kossey (USDA), the American Public Power Association and Dr. Albin Chernichowski (GlidArc Technologies) who flew here from France to be part of this Project. Also, I would like to thank Ms. Jane Flatt (McMinnville Electric System) for all her hard work and help. Jane is a one-in-a-million help and confidant. Among those who possess great literary skills, she stands as a giant; and among other traits, has the ability to make my documents look and read much better than I can write. Finally, I would like to thank my wife who has endured countless trips to the generation substation in the middle of the night to bring me items which I need; forgiven me for the numerous times when all I had time to do was run in the house, change clothes, and run back out the door again because I was needed at the Project jobsite; and for all the nights that she has been kind enough to leave a light shining so I could see to get in the door after a long day of emissions testing. McMinnville Electric System Page 1 of 182 Award Project Final Report Project Title: BioDiesel Demonstration with SCONOx NOx Removal General Overview: The purpose of the study was to demonstrate the ability of an EMx catalyst to clean the emissions from a large stationary diesel reciprocating engine. EMx had been previously demonstrated on a 50 kW diesel reciprocating engine using regeneration via direct fuel injection into the catalyst. EMx has been in commercial practice on gas turbines. This demonstration project involved scaling up the reciprocating engine experience and scaling down the gas turbine experience. It also involved a new, emerging regeneration technology capable of utilizing liquid fuels (LSD, ULSD, BioDiesel, etc.) and a plasma reformer. The catalyst used for the test was standard EMx, utilizing K2CO3 sorber on a barium-alumina washcoat. The catalysts were on 200 cpsi cordierite substrates. Four rows of EMx catalyst were installed in each chamber, with two 18” x 42” x 6” modules per row. The total catalyst volume for each catalyst chamber was 21 ft3. Sulfur management was handled by frequent washing of the catalyst; no ESx (sulfur) catalyst was installed. The catalyst regeneration gas was supplied by a plasma reformer fueled by the parent (native) fuel. The engine utilized in the Project was an off-the-shelf 3516B EPG Caterpillar engine (build date 8/20/2004) with 2293 bhp coupled to a Caterpillar SR4B generator and capable of producing 1640 kW @ 60 Hz continuous. The engine has a 4.46 g/bhp-hr NOx emission rating. This particular engine was selected by Caterpillar because it is their most popular large-bore engine for stationary applications. This engine can be found around the country in back-up and standby power generation applications. Engine performance was remotely monitored by Caterpillar and emissions performance was monitored by the Tennessee Valley Authority across a wide spectrum of fuels including: 100% ULSD (ultra-low sulfur diesel), a blend of 2% soybean biodiesel and 98% ULSD, a blend of 5% biodiesel and 95% ULSD, a blend of 20% biodiesel and 80% ULSD, a blend of 50% biodiesel and 50% ULSD, and 100% biodiesel. Purpose: To demonstrate the feasibility of using an off-the-shelf, unmodified, large bore diesel powered generator in a grid-connected application, utilizing various blends of BioDiesel as fuel. In addition, a first-of-its-kind emissions control device that uses a catalytic process and BioDiesel (without the presence of Ammonia or Urea) was developed to reduce NOx and other pollutants present in a reciprocating engine exhaust stream. McMinnville Electric System Page 2 of 182 This Project was initiated with the belief that America must become less reliant on foreign sources of fuel (i.e., petrol diesel) and become more aware of the effect that engine exhaust gasses have on our health and our environment. Soybeans represent a large segment of the world’s agriculture, and in terms of gross production, soybeans are the dominant oilseed crop1. They absorb light from the sun (solar energy), water from the earth (hydro energy) and CO2 from the air (wind energy) and convert that stored energy into natures own battery (stored energy). Soybeans therefore become the greatest fuel treasure that we produce and harbor on American soil and should also be recognized as one of our greatest sources of renewable energy. This Project is intended to provide information that fuel can be domestically produced, refined and utilized to provide clean, renewable power for American homes and businesses. All this, while providing economic incentives for farmers to make the most of fallow farmland, to provide a return on their investment and to invest that money back into our Nation’s economy. Utility Name and Address: McMinnville Electric System 200 Morford St. P.O. Box 608 McMinnville, TN 37110 Other Participants: EmeraChem Stowers/Caterpillar Tennessee Valley Authority American Public Power Association National BioDiesel Board Tennessee Soybean Promotion Board Agri-Energy, LLC Utility Description: Size: 7860 Electric Customers Annual Load: 243,682,957 kW Services Offered: Surge Protection Generation Resources: 22 MW peaking plant Other: 217 miles of line, 36 Customers/mile, 2 delivery points 1 Bajjalieh, N., 2002, Proteins from Oilseeds, Integrative Nutrition, Inc. (Research note) McMinnville Electric System Page 3 of 182 Key Personnel and Phone Numbers: McMinnville Electric System Rodney Boyd – General Manager/CEO – (931) 473-3144 Ralph Dunn – Manager, Engineering and Operations – (931) 473-3144 Neal Cox – Manager, Finance and Accounting – (931) 473-3144 Huel Martin – Electrical Engineer – (931) 473-3144 EmeraChem Steve DeCicco – Vice President of Operations – (865) 246-3000 Lisa Mitchell – Project Engineer – (865) 246-3000 Dr. Albin Chernichowski – GlidArc Technologies - +33-680-232-643 Stowers/Caterpillar Tom Stanzione – Manager, Distributed Generation – (865) 675-2869 Dave Martin – Power Generation Project Manager – (865) 546-1414 Matt Kirkpatrick – Commercial Engine Sales – (865) 546-1414 Chris Kiczaja – Sales Manager Electric Power Generation – (615) 341-3215 Tennessee Valley Authority Ralph Boroughs – Project Manager – (423) 751-4644 Chevales Ward – Environmental Engineer – (423) 751-7316 Duane Brigman – Environmental Scientist – (423) 876-4202 Other Helen Hennon – Director of Governmental Services, Quantum Environmental & Engineering Services, LLC – (865) 689-1395 Ryan Strickland – COO, Agri-Energy Management (931) 270-8129 Kelly Strebig – University of Minnesota Center for Diesel Research - (651) 330-0450 Will Ayers – V.P. Engineering, Cim-Tech Filtration – (217) 678-2511 Dr. Thomas Reed – Biomass Energy Foundation – (303) 279-3707 William Ayres – AgBioEnergy, LLC – (913) 341-7114 Dave Brown – Phillips Sales and Service – (931) 473-2450 Description: The objective of this project was to define the scope and cost of a technology research and development program that will result in ammonia free, pollution reduction system with the highest emission reduction efficiencies possible for the electric industry. Caterpillar Power Generation adapted an off-the-shelf 3616 BDITA Diesel Generator to run on BioDiesel and various Petroleum Diesel/BioDiesel blends. EmeraChem developed and installed an exhaust gas cleanup system to reduce NOx, SOx, volatile organics, and particulates. The system design and function was optimized for emissions reduction with results in the 90-95% McMinnville Electric System Page 4 of 182 range; especially for NOx. TVA measured the emissions and reviewed the environmental effects. See Attachment “B” for a thorough discussion on the EMx Prototype test, Attachment “C” for Caterpillar’s engine analysis results and comments and Attachment “D” and “E” for information on GlidArc Technology. Diagram of the Project: Dates: The DEED Grant Agreement between McMinnville Electric System and the American Public Power Association was signed in February of 2005 and the project was completed in April 2007. Alternatives: Alternatives to the Project include: operating the generator engine without the use of external pollution controls; the use of a conventional Selective Catalytic Reduction (SCR, ammonia or urea injection system); the use of other NOx emission control technologies (lean NOx catalyst, three-way catalyst, etc.); the use of lesser blends of BioDiesel (B-2, B-5, B-10, B-20, etc.). McMinnville Electric System Page 5 of 182 Results to Date: Using a 3516B Caterpillar generator, McMinnville Electric System has successfully generated 1,629,024 kWh’s of renewable electric power using soybean based, American made BioDiesel as fuel. In the process McMinnville Electric System has used 126,126 gallons of BioDiesel which equates to 84,080 bushels of soybeans. In addition, we were able to achieve a 96.6% reduction in NOx without the use of Ammonia or Urea as reductants. Status: Complete Applicability: Other utilities, especially those in non-attainment areas and environmentally sensitive areas, could use the results of this Project to site diesel powered generations in their area without the adverse environmental impact of untreated diesel exhaust emissions or the environmental impact of ammonia slip in the exhaust stream. The underground mining industry would also benefit from the environmental findings of this Project; both on the use of BioDiesel in underground stationary generation and from the use of a catalytic NOx trap. Utilities outside of the United States could use the results of this Project to site generation in areas that are remote; and thus, hard to deliver fuel to the site, by producing BioDiesel from palm oil, rape seed oil, canola oil, etc. local to the generation facility. One example is an inquiry that I have had from a location deep in the heart of Africa that has an abundance of palm oil. Transportation of petrol diesel to this area is difficult and expensive but the ability to site a BioDiesel refinery and install diesel generation is within the ability of the area. Future Plans: Future studies should include long-term studies of the effects of BioDiesel on the catalyst. All future studies hinge on the availability of reliable regeneration technologies, TVA’s acceptance of BioDiesel fueled generation and their willingness to purchase the energy output of a BioDiesel fueled generation facility. If TVA is willing to purchase electric energy produced by the combustion of BioDiesel from this Project, McMinnville Electric System will work with EmeraChem and other partners toward development and automation of a durable plasma regeneration technology that will result in a robust, efficient and clean NOx removal technology. Future plans should also include the development and construction of a more efficient and stable plasma regeneration technology possibly utilizing GlidArc-III technology. (see Attachment McMinnville Electric System Page 6 of 182 “E”) and a long term durability test (>10,000 hours) on the Caterpillar motor/generator, in a grid-connected application, using B-99.9 as fuel. Equipment: Generator: Engine: Caterpillar 3616B Duty: CONTINUOUS Connection: SER STAR Generator Frame: 826 Type: SR4B No. of Bearings: 2 Generator Arr: 1441826 Housing: 00 Winding Type: FORM WOUND Genset Rating (kW): 1640 (kVA): 2050.0 Sync Speed: 1800 Voltage: 277/480 3 phase Frequency: 60 Hz Pwf. Factor: 0.8 Rated Current: 2465.8 Gen. Pitch: 0.7143 No. of Leads: 6 EmeraChem EMx Prototype Emissions System, 30,000 gallon fuel tank, 2000 kVA power transformer, EMCP II+ Control Panel, NexGear Series 1 Advanced Paralleling Switchgear, PointGuard on-site remote-monitoring hardware. Performance: Caterpillar Generator: No downtime due to engine, internal moving components in excellent condition, hose/seal material acceptable for <B-30 but needs to be modified for operating on B- 100, engine test was very successful. EmeraChem EMx Prototype: The EMx catalyst system performs at greater than 90% NOx removal, even with very high inlet NOx concentrations and operating temperatures of 750˚F. The EMx system eliminates the visible plume and significantly silences the engine exhaust. One significant discovery was that as the fuel blend progressed from ULSD to B-100 the ability of the Caterpillar generator to export power to the grid went down accordingly. Upon Caterpillar’s review it was noted that BioDiesel had approximately 10,000 less Btu’s per gallon than ULSD and that the engine was in essence starving for fuel. When we attempted to adjust kW output beyond the load kW noted below, the engine would shut-down and bring testing to a halt. Caterpillar dispatched a technician to the job site and made a change to the throttle position sensor that allowed the engine to operate beyond the throttle limit and we were able to generate at a full 1640 kW. On several occasions we operated at 1650 kW to see how the engine would respond. Also of interest was the fuel consumption as we progressed from ULSD to B-100. As you can tell from the chart below, fuel consumption increased as the fuel blend decreased. Later, fuel consumption was checked on many other occasions at 1640 kW with B-99.9 and was found to be +/- 2% of 123 gph. Fuel consumption was calculated using a stopwatch and the change in fuel level in the generator day-tank. McMinnville Electric System Page 7 of 182 Date Fuel Load GPH Fuel kW/Gallon kW Consumption 7-Jul-05 ULSD 1625 120.00 13.54 8-Jul-05 B-2 1625 118.79 13.68 8-Jul-05 B-5 1625 120.60 13.47 9-Jul-05 B-20 1560 122.61 12.72 9-Jul-05 B-50 1545 123.33 12.53 10-Jul-05 B-100 1495 126.00 11.87 Fuel blending for all tests from B-2 through B-50 was conducted by an engineer and carefully calibrated using a certified Seraphin Model FS282 Field Standard Test Measure. Fuel quality issues plagued us throughout the testing process and resulted in McMinnville Electric System using fuel from three different BioDiesel refineries. Test results from samples taken during the test showed anywhere from high levels of methanol (see Attachment ”A”) to high levels of glycerin. Pictures depicting some of our fuel quality struggles are shown below. McMinnville Electric System Page 8 of 182 Fuel batch testing results are listed below: ASTM # test name Unit Max Min batch1 batch2 batch3 batch4 batch5 batch6 batch7 batch8 D6584 Free Glycerin % mass 0.02 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% D6584 monoglycerides 0.814% 0.792% 0.793% 0.671% 0.554% 0.633% 0.814% 0.693% D6584 diglycerides 1.872% 1.563% 1.450% 0.851% 0.816% 0.633% 1.076% 0.952% D6584 triglycerides 9.530% 8.422% 6.781% 2.566% 2.546% 2.031% 4.227% 3.568% D6584 Total Glycerin % mass 0.24 1.484% 1.317% 1.129% 0.568% 0.531% 0.478% 0.812% 0.694% D93 Flash Point °C 130 172 174 174 166 167 172 171 177 D2709 Water & Sediment vol% 0.05 0.050% 0.050% 0.040% 0.040% 0.050% 0.040% 0.120% D874 Ash, Sulfated % mass 0.02 D5453 Total Sulfur % mass 0.05 0.0001% 0.0000% 0.0000% 0.0001% 0.0001% 0.0001% 0.0000% D664 Acid Number mg KOH/g 0.8 0.390 0.250 0.250 0.220 0.280 0.170 0.250 0.280 Cc17- 95 Soap ppm 19 15 D130 Copper Corrosion #3 1a D445 Kinematic Viscosity mm/s 6 1.9 5.643 5.426 5.164 4.677 4.412 4.320 4.463 4.583 D524 Carbon Residue, Ramsbottom % mass 0.05 0.060% 0.200% 0.040% 0.030% 0.042% 0.040% 0.044% 0.010% As the above table illustrates, fuel quality is a major concern going forward with the Project. Of significant note was engine oil consumption. All large-bore stationary diesel engines are expected to consume a certain amount of oil during the engine duty cycle. Because a diesel powered generator in a grid-connected application operates under high load conditions and near the top of its horsepower range, some oil will naturally make its way past the piston rings and be ignited in the combustion process. This is natural and expected. Our experience with the other 11 large-bore stationary diesel generators that McMinnville Electric System owns and operates has shown that a diesel generator will consume approximately one (1) gallon of oil for every six (6) hours of operation under normal load (generation) conditions. It was noted early in the operation of the Project that the generator engine was not using the same amount of oil as would have been expected. After 1008 hours of operation, the Caterpillar 3516B engine consumed 25 gallons of oil which is 0.025 gph as compared to our experience consumption of 0.167 gph. This represents an 85% reduction in oil consumption comparing expected with actual. The test 3516B was equipped with a self-oiler from the Caterpillar factory and there was some concerned that the oiler was not operating properly or that some other issue was happening that we were unaware of. After review by a Caterpillar technician, there was no operational reason found why the engine was not consuming oil at the same rate as would be expected. It was theorized that the difference in lubricity between diesel fuel and BioDiesel might be a determining factor to the difference or that oil was bypassing the piston rings and making its McMinnville Electric System Page 9 of 182 way down into the crankcase resulting in lube-oil dilution. Without quantifiable data (operation of the engine for >10,000 hours) it will remain a theory and a mystery. Caterpillar took oil samples and had analysis performed at 1, 235, 500, 800 and 1007 hours of operations. Results are shown below: McMinnville Electric System Page 10 of 182 In the early planning stage of the project, one major concern regarding the use of B-100 was injector coking. According to the National BioDiesel Board, fuel injector coking can occur as a result of fuel that is of a higher viscosity that is allowed under ASTM D975 or ASTM D6752 resulting in poor fuel atomization and fuel degradation2. Because of such concern, Caterpillar removed several of the injectors at 500-hours runtime and other injectors at 1007-hours runtime and performed an injector analysis on their performance. 2 Biodiesel Handling and Use Guidelines, 2006 (U.S. Department of Energy, Energy Efficiency and Renewable Energy). (Publication with no author given) McMinnville Electric System Page 11 of 182 Injector Analysis after 500 Hours McMinnville Electric System Page 12 of 182 McMinnville Electric System Page 13 of 182 McMinnville Electric System Page 14 of 182 McMinnville Electric System Page 15 of 182 Injector Analysis after 1007 Hours of Operations McMinnville Electric System Page 16 of 182 McMinnville Electric System Page 17 of 182 McMinnville Electric System Page 18 of 182 McMinnville Electric System Page 19 of 182 Budget: Description Preliminary Budget Generator - Caterpillar $ 505,850 Motor Testing $ 110,900 Fuel Tank $ 35,000 Fuel - Biodiesel $ 187,500 Transportation Cost $ 10,704 Catalyst $ 324,209 Payroll $ 100,921 Fringe Benefits $ 44,110 Attorney $ 16,000 Electrical Engineering $ 10,000 Energy Consultant & Misc. $ 10,000 Transformer /equip $ 21,000 Environmental Consultant $ 4,300 Crane Work $ 2,000 Testing of Emissions $ 45,000 Travel $ 3,000 TOTALS $ 1,430,494 Additional Notes: On November 28, 2006, SECOR International Inc. prepared an Emissions measurements Report for McMinnville Electric System as detailed in Construction Permit No. 957279F, issued by the Tennessee Department of Environment and Conservation (TDEC), Division of Air Pollution Control (APC). The Permit specified, in Condition 15 of the construction permit, that McMinnville Electric System must conduct an emissions performance test to demonstrate compliance with the NOx emissions limit. Testing was performed in accordance with test methods and procedures detailed in 40 CFR 60, Appendix A. The report from SECOR in its entirety is hereby incorporated into this Report as Attachment “F”. References: Mitchell, L. and DeCicco, S., 2007, EMx Prototype Testing on a Caterpillar 3616 TA Stationary Internal Combustion Engine Running on Various Blends of Petroleum Diesel and Soybean-Based BioDiesel. Working paper, EmeraChem, LLC., Research Notes (Working paper) McMinnville Electric System Page 20 of 182 McMinnville BioDiesel Test, 2007 (Publication with no author given) Czernichowski, A., Czernichowski, M., Wesolowska, K., 2006, Generation of 1 kg/h of Hydrogen from Soybean BioDiesel (White paper prepared for the American Chemical Society 232nd National Meeting and Exposition) Czernichowski, A., Czernichowski, M., 2006, Further development of Plasma sources: the GlidArc-III. (White paper prepared by ECP – GlidArc Technologies, La Ferté St Aubin, France) Czernichowski, A., Czernichowski, M., Czernichowski, P. Wesolowska, K., 2006, Hydrogen or Syngas Generation using Plasma Technology. (White paper from the Topsoe Catalysis Forum 2006, Future Hydrogen Generation and Application) Rawat, J., Wehri, S., Report No: 075-010807, Fuel System Test (Caterpillar Fuel System). (Internal Report) Rawat, J., Wehri, S., Report No: 075-011107, Fuel System Test (Caterpillar Fuel System). (Internal Report) Ewing, G., 2006, Emissions Measurement Report for McMinnville Electric System. (Compliance Report prepared by SECOR International Inc.) McMinnville Electric System Page 21 of 182 Attachment “A” (The following report, which pertains solely to BioDiesel fuel test results, was commissioned by the Tennessee Valley Authority for McMinnville Electric System and is therefore incorporated and is made part of this Final Report as submitted. The author is Jim Hedman with the Minnesota Department of Commerce, Weights and Measures Division, Petroleum Laboratory Services) McMinnville Electric System Page 22 of 182 McMinnville Electric System Page 23 of 182 Results for TVA samples shipped 11/17/05 Client Peak Wave % Blend Baseline Adjusted % Blend Trace File Name Sample I.D. I.D. Absorb. Number Predict Absorb. Absorb. Predict tva-unk aa_pm_1.spc tva-unk a A 0.355465 1745.0 36.37 0.005998 0.349467 36.82 tva-unk aa_pm_2.spc tva-unk a A 0.356975 1745.0 36.56 0.006157 0.350817 37.00 tva-unk aa_pm_3.spc tva-unk a A 0.356563 1745.0 36.50 0.005157 0.351406 37.07 Sample "A" Average 36.47 Average 36.96 Std. Dev. 0.10 Std. Dev. 0.13 Uncertainty 0.27 Uncertainty 0.35 tva-unk ba_pm_1.spc tva-unk b B 0.027395 1747.4 1.95 0.005124 0.022272 1.81 tva-unk ba_pm_2.spc tva-unk b B 0.027862 1747.2 1.99 0.004331 0.023531 1.92 tva-unk ba_pm_3.spc tva-unk b B 0.027143 1747.3 1.93 0.003951 0.023191 1.89 Sample "B" Average 1.96 Average 1.87 Std. Dev. 0.03 Std. Dev. 0.06 Uncertainty 0.09 Uncertainty 0.16 tva-unk ca_pm_1.spc tva-unk c C 0.216798 1746.1 20.17 0.004015 0.212783 20.68 tva-unk ca_pm_2.spc tva-unk c C 0.218639 1746.1 20.37 0.004519 0.214120 20.83 tva-unk ca_pm_3.spc tva-unk c C 0.218284 1746.1 20.33 0.004780 0.213504 20.76 Sample "C" Average 20.29 Average 20.76 Std. Dev. 0.11 Std. Dev. 0.07 Uncertainty 0.29 Uncertainty 0.14 tva-unk da_pm_1.spc tva-unk d D 0.058266 1746.9 4.61 0.001905 0.056362 4.87 tva-unk da_pm_2.spc tva-unk d D 0.058221 1746.9 4.61 0.002322 0.055899 4.83 tva-unk da_pm_3.spc tva-unk d D 0.059020 1746.9 4.68 0.002947 0.056073 4.85 Sample "D" Average 4.64 Average 4.85 Std. Dev. 0.04 Std. Dev. 0.02 Uncertainty 0.11 Uncertainty 0.06 tva-unk ea_pm_1.spc tva-unk e E 0.754085 1741.8 96.35 0.007425 0.746660 96.10 tva-unk ea_pm_2.spc tva-unk e E 0.760908 1741.8 97.55 0.007996 0.752911 97.18 tva-unk ea_pm_3.spc tva-unk e E 0.763620 1741.8 98.03 0.008349 0.755271 97.59 Sample "E" Average 97.31 Average 96.95 Std. Dev. 0.87 Std. Dev. 0.77 Uncertainty 2.40 Uncertainty 2.13 McMinnville Electric System Page 24 of 182 McMinnville Electric System Page 25 of 182 McMinnville Electric System Page 26 of 182 McMinnville Electric System Page 27 of 182 y = 0.999896x + 0.002818 R² = 0.999896 -5 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100Predicted % BiodieselActual % Biodiesel Validation: Calibration Using Non-Adjusted Peak Absorbance y = 0.999894x + 0.002875 R² = 0.999894 -5 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100Predicted % BiodieselActual % Biodiesel Validation: Calibration Using Baseline Adjusted Peak Absorbance McMinnville Electric System Page 28 of 182 December 20, 2005 Tennessee Valley Authority Attn: Ralph Boroughs 1110 Market Street, SP-5D Chattanooga, TN 37402 Mr. Boroughs: My apologies for the delay in getting results back to you on the samples you shipped November 17th for biodiesel blend determinations. However, from late fall through this early winter we have be swamped with priority investigative and oversight work on a number of pressing issues such as water in gasoline, gasoline octane misrepresentation, methanol in biodiesel, and a number of varied filter clogging issues involving biodiesel blended diesel fuel. As luck would have it, the delay was somewhat fortuitous because of a contamination problem I found with the B100 biodiesel sample you submitted, that probably would not have been recognized had I not had to deal with a biodiesel f lashpoint issue that materialized some weeks ago. I will deal with the contamination problem shortly, but for now I turn to the biodiesel blend determinations you requested. Results were as follows: Full validation of the quadratic regression calibration protocol employed has yet to be accomplished, but past work indicates uncertainties should range from about 0.2 vol.% up to B5, under 0.8 vol. % through B20, under 1.5 vol% through B40 and under 3 vol. % at B100. Details of the calibration procedure can be reviewed by viewing the companion Excel workbook file: TVA Biodiesel Blend Calibration.xls. Please note that data for a B50 standard was discarded from the calibration set as an “outlier”. I now return to the unanticipated problem with the biodiesel blend-stock submitted for use as a standard: There appeared to be something a bit unusual about the B100 trace file so I compared it, using an overlay plot, with a pair of soy methyl ester sample retains from different plants. It became immediately obvious that there was unusually high absorbance in the spectral trace for your  Sample A:  Sample B:  Sample C:  Sample D:  Sample E: 36.96 Volume % 1.87 Volume % 20.76 Volume % 4.85 Volume % 96.95 Volume % McMinnville Electric System Page 29 of 182 B100 at about 1030 cm-1. This would correlate with methanol contamination. I realized this because, as luck would have it, I have recently undertaken a methanol-in-biodiesel quantification project, based on FTIR spectroscopic analysis. Locally, some biodiesel flash point issues have materialized. As you are probably aware, the ASTM D6851 biodiesel specification sets a very high flash point criterion (130° C), as the test is employed as a surrogate for a direct methanol test. However, there are reproducibility problems with flash point determination at such high temperatures and also there are issues with the appropriateness of the test as a methanol screen when sampling is undertaken down -stream from a production plant. Therefore, I have recently initiated a project to develop a direct test for methanol contamination of biodiesel by FTIR. The project is still in the pilot phase, but I have acquired a few scans of methanol contamination standards for demonstration purposes. I compared your TVA submitted B100 sample, using an overlay plot, with B100 samples spiked with 5% and 1% methanol by volume. Also included was the stock sample used for the methanol spikes. This overlay is depicted in Figure 1. Figure 1. Overlay plot of spectral traces for TVA submitted sample compared with 0%, 1% and 5% methanol spiked B100 “standards”. The “signature” peak for methanol appears at about 1030 cm-1 and is very evident. McMinnville Electric System Page 30 of 182 Next, in Figure 2, the region about wavenumber 1030 cm-1 is shown in a zoomed view to facilitate detailed comparison of the spectra. Clearly, the peak absorbance for the TVA submitted sample lies about midway between the peak absorbencies for the 1% and 5% methanol spiked B100 standards. Therefore, it can be concluded that the TVA submitted sample contains roughly 3% methanol. A more precise estimate could be obtained by using additional interlaying standards, preferably employing them in a calibration protocol using regression or multivariate analysis. One potential problem in developing a low-level methanol in biodiesel calibration is the “native” biodiesel absorbance increase at about 1017 cm-1 wavenumber. The interference of the biodiesel peak is already evident in the 1% methanol trace. However, the methanol contamination in the TVA submitted B100 is far above the trace amounts anticipated in a method that will hopefully be sensitive to concentrations down to 0.2% or lower. Figure 2. Zoomed “methanol signature region” overlay plot of spectral traces for TVA submitted sample compared with 0%, 1% and 5% methanol spiked B100 “standards”. The contamination in the TVA sample appears to be about 3 vol. %. McMinnville Electric System Page 31 of 182 I was rather surprised, if not shocked, to see such a high level of methanol contamination in the B100 you supplied. The anticipated methanol level associated with the ASTM D685 1 flash point specification of 130° C is only 0.2% - with a view to harmonization with the European biodiesel specification. It should also be noted that a “methanol signature” was clearly evident in the TVA samples identified as A, C and E. These were th e three highest determined biodiesel blend levels. Assuming that the B100 indeed contains 3% methanol, at B5 the methanol concentration would be reduced to 3 x .05 = 0.15 vol.%, which is probably near the limit of detection for this sort of analysis (time and effort will tell). Again my apologies for the delay in submitting this report, but had I made the blend determinations in a very timely manner, the methanol contamination issue would most likely have been overlooked. See companion paper (in zipped file) Biodiesel - MeOH Detection (051129a).pdf for more information about the methanol detection method project that led to the discovery of the methanol issue with the samples you submitted. Please feel free to write or call me at 651-296-2990 with any questions, concerns or commentary. Sincerely, Jim Hedman, Metrologist McMinnville Electric System Page 32 of 182 Attachment “B” (The following report, which pertains solely to the Prototype SCONOX NOx emissions reduction system which as designed and built for this Project by EmeraChem, LLC., was commissioned by McMinnville Electric System and is therefore incorporated and is made part of this Final Report as submitted. The authors are Steve DeCicco and Lisa Mitchell, EmeraChem, LLC) McMinnville Electric System Page 33 of 182 Introduction Demonstration Goals The ultimate goal of MES’ demonstration program is to demonstrate the ability to generate electricity on 100% renewable fuel source, while maintaining emissions at or lower than non- renewable sources. This would allow McMinnville Electric System to qualify for TVA’s “Green Power Switch” program. EmeraChem’s goal is to demonstrate clean emissions on a large stationary IC engine across all liquid fuel blends. Furthermore, EmeraChem hopes to develop a commercial product for stationary diesel engine applications. Background EmeraChem’s NOx adsorber catalysts have been demonstrated on bench scale reactors and on diesel engine exhaust. Long term data on EMxTM performance exists from commercial applications on 5 – 50 MW natural-gas and dual fuel fired turbines demonstrating NOx emissions less than 1 ppm and virtually undetectable CO and HC levels. Long term data on EMxTM on diesel engines is limited to studies on small, 5 – 50 kW, diesel engine generator-set test platforms, demonstrating more than 90% removal of NOx, CO, and HC. This test represents the verification of the performance of the catalyst on the exhaust from a larger scale, 2 MW, stationary diesel engine. Experiment Design Engine Test Platform The engine generator-set is Caterpillar 3516B TA diesel engine (see Figure II-1) coupled to a Caterpillar SR4B generator and has a 4.46 g/bhp-hr NOx emission rating. The exhaust flow is assumed to be 5,636 scfm (wet, at 32˚F and 29.98” Hg) as predicted by Caterpillar. The gen-set is attached to the electric grid, and is at full load at all times. The engine was run for approximately 16 hours prior to the start of testing to break in the engine and ensure sealing of all exhaust system components. Figure II-1 McMinnville Electric System Page 34 of 182 Fuel and Lubricants Various blends of ultra-low sulfur diesel (ULSD) and 100% soybean based biodiesel (B100) fuels were used to fuel both the engine and the plasma reformer for regen gas generation. The fuels were mixed manually in the fuel day tank. The ratios of ULSD to B100 fuel were chosen to represent various fuel blends commercially available. The sulfur content of the ULSD was 15 ppm, and 0 ppm for the biodiesel. Testing Protocol Prior to the emissions testing, the plasma reformer operation was optimized to develop a start- up procedure and confirm sufficient hydrogen production. The catalyst modules were not installed prior to the initial day of ULSD fuel testing. The catalyst was operated with a cycle time of 4 minutes for early testing, and revised to 3.5 minutes before B20 testing began. After the 4-hour tests on each of the various fuel blends, the engine was run on B100 for the remainder of the testing. The catalyst was washed after each day of testing to eliminate the effects of sulfur accumulation from test to test. Catalyst System The catalyst system is a dual chamber reactor with inlet exhaust isolation valves controlling the direction of exhaust flow (see Figure II-2). The exhaust isolation valves used a common actuator and were linked together to assure that the engine exhaust would always have an open flow path and never be restricted. Regeneration was accomplished by generating a regen gas from the parent fuel using a plasma reformer. The regen gas was injected in a “forward-flow” geometry (the regen gas flowed in the exhaust flow direction). The catalyst system was mounted 20 ft from the gen-set exhaust as shown in Figure II-3; thus, allowing enough thermal loss in the un- insulated exhaust pipe to bring the catalyst operating temperature down to an operating temperature of 750˚F. The gen-set was enclosed in a sound-reducing housing and catalyst system was installed outside with no Figure II-2 McMinnville Electric System Page 35 of 182 additional rain protection. Catalysts The catalyst (see Figure II-4) used for the test was standard EMx, utilizing K2CO3 sorber on a barium-alumina washcoat. The catalysts were on 200 cpsi cordierite substrates. Four rows of EMx catalyst were installed in each chamber, with two 18” x 42” x 6” modules per row. The total catalyst volume for each catalyst chamber was 21 ft3. Sulfur management was handled by frequent washing of the catalyst, no ESx catalyst was installed. The EMx catalyst works by simultaneously oxidizing CO to CO2, VOCs to CO2 and H2O, NO to NO2, and then absorbing NO2 onto its surface through the use of an alkaline metal solution absorber coating such as potassium carbonate. These reactions are shown below, and are referred to as the “Oxidation/Absorption Cycle”. CO + ½O2  CO2 C H2O + O2  CO2 + H2O NO + ½O2  NO2 2NO2 + K2CO3  CO2 + KNO2 + KNO3 2NO2 + Ba2CO3  CO2 + BaNO2 + BaNO3 The small quantity of CO2 in the above reactions exhausts up the stack. Note that during this cycle, the potassium carbonate coating quantitatively and chemically bonds and traps the nitrogen oxides to form potassium nitrites and nitrates, which are then present on the surface of the catalyst. Before the potassium Figure II-4 Figure II-3 McMinnville Electric System Page 36 of 182 carbonate on the surface of the catalyst becomes saturated with nitrogen oxides, the catalyst enters the regeneration cycle. The regeneration of the EMx catalyst, one of the features that makes the system so unique, is accomplished by passing a reducing gas across the surface of the catalyst in the absence of oxygen. The reductants in the regeneration gas (hydrogen and carbon monoxide) react with nitrites and nitrates to form water, elemental nitrogen, and potassium hydroxide. Carbon dioxide in the engine exhaust reacts with potassium hydroxide to form potassium carbonate, which is the absorber coating that was on the surface of the catalyst before the oxidation/absorption cycle began. This cycle is referred to as the “Regeneration Cycle”, and the relevant reaction is shown below. KNO2 + KNO3 + 4H2 + CO2  K2CO3 + 4H2O + N2 Water (as steam) and elemental nitrogen are exhausted up the stack instead of NOx, and potassium carbonate is once again present on the surface of the catalyst, allowing the oxidation/absorption cycle to begin again. There is not a net gain or net loss of potassium carbonate after the oxidation/absorption and regeneration cycle. Regen System To qualify for the “Green Power Switch”, all energy sources utilized must be 100% renewable. To satisfy this requirement, a GlidArc plasma reformer fabricated by ECP of France was chosen. This unit had been demonstrated on many fuels including diesel, gasoline, natural gas, propane, canola, glycerol, and sugar water. The reformer was sized to mix approximately 3.5 gal/hr of biodiesel with 25 scfm of air and spray the mixture into an electric arc. The fuel penalty was 2.7%. When properly adjusted, the resulting regen gas is 18% H2, 18% CO, N2, H2O, and CO2. (For more information on GlidArc technology, see Attachment “C” and “D”, specifically the sections referring to McMinnville Electric System) The GlidArc reformer (see Figure II-5 and Figure II-6) is a cold plasma- assisted reformer that produces hydrogen and carbon monoxide through partial oxidation of liquid or gaseous fuels. A rich mixture of fuel, air, and water are sprayed through a low current, high-voltage gliding arc. The fuel is ignited by the arc and hot internals of the reformer. Excess fuel is converted to hydrogen and carbon monoxide in the in the activated refractory of the post-plasma zone. The entire reformer is surrounded by an annular pre-heat zone to pre-heat the air supply. The outside of the reformer is wrapped in ceramic wool insulation for safety. Figure II-5 McMinnville Electric System Page 37 of 182 Figure II-6 Number Description 1 Thermowell to measure internal temps 2 Thermowell to measure outlet temp 3 Lid insulation 4 Lid heat shield 5 Electrodes 6 External insulation 7 Stainless steel outer wrap 8 Outlet screen to contain refractory 9 Internal insulation 10 Activated refractory – post-plasma zone 11 Plasma zone McMinnville Electric System Page 38 of 182 For partial oxidation, the reforming temperature must be maintained between 800ºC and 1000ºC as shown in Figure II-7. This temperature is maintained through the addition of air and/or steam. Adding air alone increases the temperature of the reaction, and can lead to damaging the reformer internals. Adding steam alone decreases the temperature and can lead to soot formation. By adding the appropriate mixture of air and steam, a stable reformation temperature can be achieved, leading to the formation of maximum hydrogen. Soot Stable region Air/Fuel Ratio 1000 2000 1 0 0 Adding steam cools reaction Stay below 1000 C or you’ll damage reactor components Reforming window Stay above 800 C to remove soot CH2 + ½ O2 CO + H2 Adding air + steam reaches optimum reforming T, H2 Adding air gives too high temperature Figure II-7 McMinnville Electric System Page 39 of 182 Data Acquisition The data acquisition system for the test included a gas chromatograph, a portable combustion analyzer, and a portable hydrogen analyzer as shown in Figure II-8. Exhaust was sampled from four positions: 1. “engine out” - upstream of the catalyst system 2. “catalyst out” – downstream of the catalyst system (inside the catalyst chamber, below the last module) 3. “system out) – downstream of the union joining the exhaust from both reactor chambers, as in Figure II-9 4. “regen inlet” – in the regen gas supply pipe Dry gas from the chiller was analyzed with a gas chromatograph and a H2 analyzer. The portable combustion analyzer was used to analyze the exhaust gas directly at the sample ports. Temperature measurements were made upstream of the catalyst system. In addition, the catalyst temperature in each chamber was measured at one location downstream of the catalyst in each chamber. A manometer was used to sample the exhaust pressure upstream and downstream of the catalyst system. Original Test Plan The installation and commissioning of the components and the system was planned for 6/28/05 through 7/3/05. During this time, the following steps were planned: 1. Install plasma reformer and connect engine exhaust to the EMx emission control system Figure II-8 Figure II-9 McMinnville Electric System Page 40 of 182 2. Start the plasma reformer and optimize operation 3. Run the plasma reformer for 24 hour to verify long-term stability 4. Install the EMx catalyst 5. Start the engine and optimize the function of complete system 6. Verify oxygen depletion in the chamber during regen, as an indication that regen gas was reaching all of the modules 7. Verify sufficient mass flow rate of hydrogen content in the regen gas for complete regeneration of the modules After commissioning the system, we would begin a series of short term tests. Each test would last 4 hours, with sampling beginning 2 hours into the test. After each test, the catalyst would be washed to re-establish the baseline. The fuels tested would be ULSD, 2% biodiesel, 5% biodiesel, 20% biodiesel, 50% biodiesel, and 100% biodiesel. Long-term test would follow, using only 100% biodiesel. The test would last 1500 hours, with stops at 500, 1000, and 1500 hours for Caterpillar to inspect the engine for wear and inspect the fuel filters and injectors for blockage. Installation and Commissioning Table III-1 summarizes the plasma reformer runs during installation and commissioning. The engine was not operating and the EMx catalyst was not installed until just before run 9. Start-up issues such as overheating, soot built-up and insufficient supplies of ULSD prevented a 24 hour run of the plasma reformer prior to EMx system start-up. Run Date Fuel Duration Reason Stopped Comments 1 6/29/05 ULSD 0:24 End of day 2 6/30/05 ULSD 1:29 Flare impinging on isolation valve actuator 3 6/30/05 ULSD 1:33 End of day 13% H2, no soot, full air and fuel flow for 30 minutes 4 7/1/05 ULSD 1:09 Storm/lightening 18% H2, no soot, invisible flame H2 flow – 0.924 kg/hr, Total regen gas flow – 35.1 scfm 5 7/1/05 ULSD 0:25 Reformer thermowell breached Fuel valve closed, mixture went lean, max temp ~ 1271˚C 6 7/2/05 ULSD 1:51 End of day New thermowell, H2 – 15.2% - 18.5%, CO ~ 18% 7 7/3/05 B100 0:55 Storm/lightening McMinnville Electric System Page 41 of 182 8 7/3/05 B100 2:25 Test complete Changed to larger fuel and water lines, H2 – 21% - 23.5%, CO ~ 21% 9 7/3/05 USLFO 1:00 End of day Initial loading of catalyst. Closed vent - Cycling regen gas through EMx (engine off) 10 7/4/05 ULSD 4:33 Decreasing air flow Electrodes, Ni balls & thermowell in good condition 11 7/5/05 ULSD 7:17 Decreasing air flow, concerned about soot formation Post-run inspection revealed heavy carbon buildup, some fusing of catalyst media, some Ni balls fused Table III-1 Figure III-1 shows the increase in the chamber temperature (with catalyst installed) during startup as well as the untreated engine CO and NOx emissions during start-up. Both the temperature and emissions are stabilized in less than 1 hour. Figure III-1 System Temperatures During Startup 0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 0:00:00 0:14:24 0:28:48 0:43:12 0:57:36 1:12:00 1:26:24 1:40:48 1:55:12 Elapased Time From Engine Startup (h:m:s)Temperature (F)0 100 200 300 400 500 600 700 800 Engine NOx (ppm) and CO (ppm)Engine Exhaust RX100 RX200 Regen Gas NOx CO Startup Conditions: ambient 75F and raining. Long-term steady state temperatures, sunny 85F ambient temp: 750F engine exhaust and 730-740F EMx exhaust. Note that actual regen gas temp at reactor inlet is approximately 750F. McMinnville Electric System Page 42 of 182 Figure III-2 shows the oxygen depletion in both chambers measured approximately 1 foot below (downstream) of the last catalyst module. Chamber RX100 is slower to respond and is not reaching as low of a level of oxygen concentration as chamber RX200. This could indicate a small leak in the isolation valve (allowing engine exhaust to leak into the chamber during regen) or a leak in the regen gas supply line (lowering the flow rate of regen gas into the chamber). Figure III-2 EMx Outlet - Oxygen Depletion ULSD - 7/5/05 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0:00:00 0:01:26 0:02:53 0:04:19 0:05:46 0:07:12 0:08:38 0:10:05 0:11:31 0:12:58 0:14:24 0:15:50 0:17:17 0:18:43 0:20:10 Elapsed TimeOxygen Concentration (%)Oxygen RX 100 Oxygen RX 200 Sorption Mode Starts - 8:37 - 4 minute cycle Regen Mode Starts - 4:25 - 4 minute cycle McMinnville Electric System Page 43 of 182 Figure III-3 shows the NOx concentration in each chamber below the last EMx module throughout the cycle on one of the commissioning runs using ULSD. These measurements were not simultaneous, but were overlaid to represent the same point in the sorption/regen cycle. These values indicate the concentration in ppm of NOx in the reactor chamber downstream of the catalyst and have not been normalized for flow rates. (During regen, the concentration of NOx was high, but the flow rate was very low – approximately 0.7% of engine flow, so the effect that it had at the stack was very low.) Chamber RX200 is capturing more NOx than Chamber RX 100. This could be caused by lower NOx sorption capacity due to insufficient regeneration or by untreated exhaust leakage past the catalyst modules in this chamber. Figure III-3 EMx Outlet - NOx Levels ULSD - 7/5/05 0 20 40 60 80 100 120 140 160 180 200 0:00:00 0:01:26 0:02:53 0:04:19 0:05:46 0:07:12 0:08:38 0:10:05 0:11:31 0:12:58 0:14:24 0:15:50 0:17:17 0:18:43 0:20:10 Elapsed TimeNOx Concentration (ppmvd - uncorrected)NOX RX100 NOX RX200 Engine Exhaust: 750F 619 ppmv Nox 601 ppmv NO 18 ppmv NO2 116 ppmv CO 10.8% oxygen Sorption (High flow) Regen (Low flow) Sorption (High flow) Regen (Low flow) 4 min cycle time McMinnville Electric System Page 44 of 182 Figure III-4 shows the CO concentration in each chamber below (downstream of) the last EMx module throughout the cycle. These measurements were not simultaneous, but were overlaid to represent the same point in the sorption/regen cycle. These values indicate the concentration in ppm of CO in the reactor chamber downstream of the catalyst and have not been normalized for flow rates. (During regen, the concentration of CO was high, but the flow rate was very low – approximately 0.7% of engine flow, so the effect that it had at the stack was very low.) The slow drop in CO concentration during later sorption cycles represents the slow recovery of the analyzer after exposure to very high concentrations of CO, and does not represent actual CO outlet concentrations during sorption. The first sorption cycle is indicative of actual CO outlet concentrations. Chamber RX100 is releasing much less CO than chamber RX200 during regeneration. This seems to indicate the chamber RX100 is not getting sufficient regeneration, therefore more of the CO is getting used as the reducing agent to regenerate the catalyst. Figure III-4 EMx Outlet - CO Levels ULSD - 7/5/05 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 0:00:00 0:01:26 0:02:53 0:04:19 0:05:46 0:07:12 0:08:38 0:10:05 0:11:31 0:12:58 0:14:24 0:15:50 0:17:17 0:18:43 0:20:10 Elapsed TimeCO Concentration (ppmvd - uncorrected)CO RX100 CO RX200 Engine Exhaust: 750F 619 ppmv Nox 601 ppmv NO 18 ppmv NO2 116 ppmv CO 10.8% oxygen Regen gas composition is ~18% CO. During regeneration, this value is reduced by the EMx catalyst to <1%. 4 min cycle time Sorption (High flow) Regen (Low flow) Sorption (High flow) Regen (Low flow) Slow analyzer recovery McMinnville Electric System Page 45 of 182 Figure III-5 shows the results of the NOx capacity test run on chamber RX100. Earlier, observations were made that indicate that chamber RX100 is not achieving full regeneration: 1. Figure III-2 shows that the concentration of oxygen in chamber RX100 is getting reduced slower than in chamber RX200 2. Figure III-2 also shows that the minimum concentration of oxygen achieved in chamber RX100 is 1-2%, which is higher than chamber RX200. 3. Figure III-3 shows that chamber RX100 has a higher NOx outlet concentration during sorption than chamber RX200. 4. Figure III-4 shows that chamber RX100 has a lower CO outlet concentration during regeneration than chamber RX200. 5. We later learned that a significant amount of untreated engine exhaust was leaking around the catalyst and influencing the outlet NOx values. All of these observations indicate that the catalyst in chamber RX100 is not performing as well as it could be. The regen gas is displacing the exhaust gas at a slower rate and does not appear to be fully displacing the exhaust gas during the regen cycle. This chamber is relying on the CO in the regen gas to regenerate the catalyst. The catalyst in chamber RX100 is not performing as well as the catalyst in RX200. The actual sorption capacity of EMx catalyst under these conditions is therefore higher than this graph would indicate. This data SHOULD NOT be used as a representation of the maximum capacity of EMx catalyst at this temperature and space velocity. Figure III-5 EMx NOx Capacity Test RX100 Outlet 7hr ULSD - 7/5/05 0 100 200 300 400 500 600 700 0:00:000:00:430:01:260:02:100:02:530:03:360:04:190:05:020:05:460:06:290:07:120:07:550:08:380:09:220:10:050:10:480:11:310:12:140:12:580:13:410:14:240:15:070:15:500:16:340:17:170:18:000:18:430:19:260:20:100:20:530:21:360:22:190:23:020:23:460:24:29Elapsed TimeNOx Concentration (ppmvd - uncorrected)Catalyst Outlet Engine Out Engine Exhaust: 750F 619 ppmv Nox 601 ppmv NO 18 ppmv NO2 116 ppmv CO 10.8% oxygen tau-10 @5.5 min 0.98 ft3 NOx/ft3 catalyst tau-20 @8.5 min 1.48 ft3 NOx/ft3 tau-30 @12 min 1.99 ft3 NOx/ft3 tau-40 @15 min 2.4 ft3 NOx/ft3 tau-50@20 min 2.86 ft3 NOx/ft3 CAUTION: SIGNIFICANT EXHAUST BYPASS AROUND CATALYST DURING THIS MEASUREMENT. McMinnville Electric System Page 46 of 182 Figure III-6 shows the long term CO emissions during the NOx capacity test. The CO outlet slightly increases over the duration of the test. This confirms that the oxidation performance of the catalyst is virtually unaffected by the level of saturation of the NOx sorbent. Figure III-6 CO Performance Test RX100 Outlet 7hr ULSD - 7/5/05 0 20 40 60 80 100 120 0:00:000:00:430:01:260:02:100:02:530:03:360:04:190:05:020:05:460:06:290:07:120:07:550:08:380:09:220:10:050:10:480:11:310:12:140:12:580:13:410:14:240:15:070:15:500:16:340:17:170:18:000:18:430:19:260:20:100:20:530:21:360:22:190:23:020:23:460:24:29Elapsed TimeCO Concentration (ppmvd - uncorrected)CO Catalyst Outlet Engine Out Engine Exhaust: 750F 619 ppmv Nox 601 ppmv NO 18 ppmv NO2 116 ppmv CO 10.8% oxygen CAUTION: SIGNIFICANT EXHAUST BYPASS AROUND CATALYST DURING THIS MEASUREMENT. McMinnville Electric System Page 47 of 182 Short Term Fuel Blend Testing ULSD Test Figure IV-1 shows the engine emissions of NOx and CO when the engine is running ultra-low sulfur diesel as fuel. Figure IV-2 shows the catalyst outlet emissions measured in the common exhaust stack for both reactor chambers. The measurements combine the treated engine exhaust from the chamber in absorption mode with the spent regen gas leaving the chamber in regeneration mode. The NOx removal for both chambers averages 80%. We later learned that a significant amount of untreated engine exhaust was leaking around the catalyst and influencing the outlet NOx values. The NOx emissions for RX 100 are higher than those for RX 200 as observed during installation and commissioning. The CO spikes are due to the high CO concentration regeneration gas being swept in a plug flow from the chamber that was in regeneration. The higher CO spike is from regeneration on chamber RX 200, as observed during installation and commissioning. Figure IV-3 shows the CO removal if we remove the effects of the CO in the regen gas. This would indicate that a CO removal efficiency of 95.5% would have been measured if a shift reactor were added to remove the CO from the regen gas. This CO removal efficiency is low for an oxidation catalyst with this level of precious metal loading when operated at 750˚F at this gas hourly space velocity. This low oxidation performance indicates exhaust gas bypass around the catalyst. The pressure drop from the engine exhaust pipe ahead of the reactor to just below the catalyst (during sorption) was measured to be 16.2” H2O. This includes pressure drop through the elbows, the open isolation valve, the perforated diffuser plate, the catalyst, the 180˚ turn in the reactor, and the union joining the two reactor exhaust stacks. For this test, the plasma reformer was run for a total of 10 hours, 20 minutes. The air flow remained steady for the run, ranging from 20.7 scfm at 13:20 to 19.8 at 19:50. The regen gas was composed of 19.6% hydrogen, 18.7% carbon monoxide, 54.5% nitrogen, 1.6% methane, and unmeasured concentrations of carbon dioxide and water vapor. After the ULSD fuel blend test, the catalyst was chemically washed to remove any sulfur and any other masking agents that may have accumulated on the catalyst. The modules were still slightly damp when they were re-installed into the chamber. McMinnville Electric System Page 48 of 182 Figure IV-1 Figure IV-2 Engine Exhaust Characteristics ULSD - 7/6/05 0 100 200 300 400 500 600 700 3:02:53 PM 3:04:19 PM 3:05:46 PM 3:07:12 PM 3:08:38 PM 3:10:05 PM 3:11:31 PM 3:12:58 PM 3:14:24 PM 3:15:50 PM 3:17:17 PM 3:18:43 PM 3:20:10 PM 3:21:36 PM 3:23:02 PM 3:24:29 PM 3:25:55 PM 3:27:22 PM 3:28:48 PM 3:30:14 PM TimeNOx & CO Concentration (ppmvd - uncorrected)NOX CO Average NOx - 612 ppm Average CO - 90 ppm 10.8% O2 EMx Stack Measurment from Both Reactor Chambers ULSD - 7/6/05 0 100 200 300 400 500 600 700 2:13:55 PM 2:15:22 PM 2:16:48 PM 2:18:14 PM 2:19:41 PM 2:21:07 PM 2:22:34 PM 2:24:00 PM 2:25:26 PM 2:26:53 PM 2:28:19 PM 2:29:46 PM 2:31:12 PM 2:32:38 PM 2:34:05 PM 2:35:31 PM 2:36:58 PM 2:38:24 PM TimeNOx & CO Concentration (ppmvd - uncorrected)NOX CO 4 min cycle time Engine Exhaust: 730F 612 ppmv Nox 579 ppmv NO 33 ppmv NO2 90 ppmv CO 10.8% oxygen RX 100 - 68 to 85% NOx DRERX 200 - 73 to 87% NOx DRE CO spike from residual regen gas swept out of chamber at start of sorption cycle. CAUTION: SIGNIFICANT EXHAUST BYPASS AROUND CATALYST DURING THIS MEASUREMENT. McMinnville Electric System Page 49 of 182 Figure IV-3 B2 Test Figure IV-4 shows the engine emissions of NOx and CO when the engine is running 98% ultra- low sulfur diesel mixed with 2% biodiesel as fuel. Figure IV-5 shows the catalyst outlet emissions. The NOx removal for both chambers averages 88%. The NOx emissions for both chambers are lower than in the previous test, but the discrepancy between the two is greater. The CO spikes are much lower than in the previous test. It was later discovered that exhaust bypass around the catalyst was occurring. The pressure drop from the engine exhaust pipe ahead of the reactor to just below the catalyst (during sorption) was measured to be 15.1” H2O on the RX100 chamber and 16.7” H2O on the RX200 chamber side. For this test, the plasma reformer was run for a total of 8 hours, 5 minutes. The regen gas was composed of 17.2% hydrogen, 17.1% carbon monoxide, 54.9% nitrogen, and 1.6% methane, and unmeasured concentrations of carbon dioxide and water vapor. The regen gas flow rate was calculated to be 31.5 scfm (76 acfm) versus design specification of 35 scfm. The hydrogen flow rate was calculated to be 0.76 kg/hr versus design specification of 1 kg/hr. The plasma Average CO Destruction Efficiency ULSD Fuel - 7/6/05 0 20 40 60 80 100 120 3:01:26 PM 3:02:53 PM 3:04:19 PM 3:05:46 PM 3:07:12 PM 3:08:38 PM 3:10:05 PM 3:11:31 PM 3:12:58 PM 3:14:24 PM 3:15:50 PM 3:17:17 PM 3:18:43 PM 3:20:10 PM 3:21:36 PM 3:23:02 PM 3:24:29 PM 3:25:55 PM 3:27:22 PM 3:28:48 PM 3:30:14 PM TimeCO Concentration (ppmvd - uncorrected)CO from Engine CO out of EMx CO from Engine Averaged 90 ppmv CO from EMx (Measured Downstream of Catalyst on RX100) Averaged 4 ppmv During Sorption Average Destruction Efficiency of 95.5% Engine Exhaust: 730F 612 ppmv Nox 579 ppmv NO 33 ppmv NO2 90 ppmv CO 10.8% oxygen 4 min cycle time CAUTION: SIGNIFICANT EXHAUST BYPASS AROUND CATALYST DURING THIS MEASUREMENT. McMinnville Electric System Page 50 of 182 reformer’s flow rate and hydrogen concentration were below the design specification, which adversely affected the EMx catalyst performance. Figure IV-4 Figure IV-5 Engine Exhaust Characteristics B2 - 7/7/05 0 100 200 300 400 500 600 700 800 1:00:29 PM 1:01:55 PM 1:03:22 PM 1:04:48 PM 1:06:14 PM 1:07:41 PM 1:09:07 PM 1:10:34 PM 1:12:00 PM 1:13:26 PM 1:14:53 PM 1:16:19 PM 1:17:46 PM 1:19:12 PM 1:20:38 PM 1:22:05 PM 1:23:31 PM TimeNOx & CO Concentration (ppmvd - uncorrected)NOX CO Average NOx - 658 ppm Average CO - 86 ppm 10.9% O2 EMx Stack Measurment from Both Reactor Chambers B2 - 7/7/05 0 100 200 300 400 500 600 700 2:22:34 PM 2:24:00 PM 2:25:26 PM 2:26:53 PM 2:28:19 PM 2:29:46 PM 2:31:12 PM 2:32:38 PM 2:34:05 PM 2:35:31 PM 2:36:58 PM 2:38:24 PM 2:39:50 PM 2:41:17 PM 2:42:43 PM TimeNOx & CO Concentration (ppmvd - uncorrected)NOX CO 4 min cycle time Engine Exhaust: 750F 650 ppmv Nox 625 ppmv NO 25 ppmv NO2 85 ppmv CO 10.9% oxygen RX100 93.8 to 96.3% RX200 80 to 90% CAUTION: SIGNIFICANT EXHAUST BYPASS AROUND CATALYST DURING THIS MEASUREMENT. McMinnville Electric System Page 51 of 182 B5 Test Figure IV-6 shows the engine emissions of NOx and CO when the engine is running 95% ultra- low sulfur diesel mixed with 5% biodiesel as fuel. Figure IV-7 shows the catalyst outlet emissions. The NOx removal for both chambers averages 72%. The discrepancy in NOx emissions for the two chambers continues to grow. The CO spikes are higher than the previous test, but lower than the ULSD test. It was later discovered that exhaust bypass around the catalyst was occurring. For this test, the plasma reformer was changed from B2 to B5 while running and continued to run for another 2 hours, 13 minutes. The regen gas was composed of 17.2% hydrogen, 17.0% carbon monoxide, 52.6% nitrogen, and 1.5% methane, and unmeasured concentrations of carbon dioxide and water vapor. Figure IV-6 Engine Exhaust Characteristics B5 - 7/7/05 0 100 200 300 400 500 600 700 800 4:46:34 PM 4:48:00 PM 4:49:26 PM 4:50:53 PM 4:52:19 PM 4:53:46 PM 4:55:12 PM 4:56:38 PM 4:58:05 PM 4:59:31 PM 5:00:58 PM 5:02:24 PM 5:03:50 PM 5:05:17 PM 5:06:43 PM 5:08:10 PM 5:09:36 PM 5:11:02 PM TimeNOx & CO Concentration (ppmvd - uncorrected)NOX CO Average NOx - 724 ppm Average CO - 87 ppm 10.7% O2 McMinnville Electric System Page 52 of 182 Figure IV-7 Observations on ULSD, B2 and B5 Short Term Fuel Blend Tests The incoming NOx levels were much higher than were expected. At an inlet concentration of nearly 700ppm, the flow rate of NOx is approximately 4.2 scfm or 16.8 ft3 in a 4 minute cycle. This consumes 84% of the measured Tau 10 capacity. This minimizes the safety factor to compensate leakage or inefficient regeneration. Reducing the cycle time to 3 minutes would reduce the loading to 63% of the Tau 10 capacity. More hydrogen is needed for regeneration to compensate for the higher incoming NOx loading. The primary reaction taking place is: KNO3 + 3H2 = 3H2O + ½ N2 cycle lbH moleH lbH moleNO molesH ft moleNONOxft 2 2 2 2 2 3 2 3 266.2*3*379 1*min4 8.16  hour H Kg 1.8 hour H lb 4 hour cycles 15 *cycle H lb 0.266 222 EMx Stack Measurement from Both Reactor Chambers B5 - 7/7/05 0 100 200 300 400 500 600 700 3:57:36 PM 3:59:02 PM 4:00:29 PM 4:01:55 PM 4:03:22 PM 4:04:48 PM 4:06:14 PM 4:07:41 PM 4:09:07 PM 4:10:34 PM 4:12:00 PM 4:13:26 PM 4:14:53 PM 4:16:19 PM 4:17:46 PM 4:19:12 PM 4:20:38 PM 4:22:05 PM 4:23:31 PM 4:24:58 PM 4:26:24 PM TimeNOx & CO Concentration (ppmvd - uncorrected)NOX CO 4 min cycle time Engine Exhaust: 750F 724 ppmv Nox 691 ppmv NO 32 ppmv NO2 87 ppmv CO 10.7% oxygen RX100 92.3 to 84.5% RX200 52 to 74% CAUTION: SIGNIFICANT EXHAUST BYPASS AROUND CATALYST DURING THIS MEASUREMENT. McMinnville Electric System Page 53 of 182 As noted during the B2 fuel blend test, the actual hydrogen production was measured at 0.76 kg/hr. This is 42% of the minimum theoretical requirement. Fortunately, CO is contributing to regen. Regen gas containing approximately 18% CO is reduced to approximately 2% after catalyst during regen. This suggests that the CO is a strong reductant, contributing to the regeneration of the catalyst. It should be noted that the actual time available for regeneration was less than 4 minutes. During the oxygen depletion study, it was shown that it took approximately 1 ½ minutes for the oxygen to be depleted. Depending on the mixing in the chamber, some of the hydrogen is reacting with the oxygen present. As the chamber fills with regen gas, it begins to move in a plug flow through the catalyst modules and displaces the oxygen. Regeneration will not begin until regen gas containing hydrogen and carbon monoxide reaches the first row of catalyst, most likely after plug flow is established. At the end of the cycle, regen gas that is introduced in approximately the last 30 seconds does not reach the last row of catalyst before the sorption cycle begins and the regen gas is swept out of the chamber by the engine exhaust. The performance of chamber RX 200 has degraded significantly throughout the course of the tests. The performance of chamber RX 100 has fluctuated, but remained close to 90% at its peak value. It appears that chamber RX 200 is not getting sufficient regeneration, and the working capacity is steadily decreasing as a result. Bypass could also be worsening as the wall separating the two chambers warps. Furthermore, leaks have been witnessed around the doors of the chambers, which is not only dangerous for the on-site observers, but can reduce the affectivity of the regeneration. When the catalyst chamber was opened for inspection and to wash the catalyst, black leak paths were observed at the corners of the catalyst as shown in Figure IV-8 indicating that exhaust and regen gas were leaking out of the doors and around the catalyst modules. A leak path was also identified behind the catalyst. The chambers had a common wall on the back that was a flat plate of carbon steel. During the initial heat-up of the system, the wall was subjected to uneven heating as first one side of the wall was heated and then the other. As a result, the wall warped, leaving a leak path behind the modules as shown in Figure IV-9. When the catalyst modules were chemically washed after the B2 and B5 fuel blend runs, the wash solution was unusually dirty. There was a high amount of floating, black, oily matter that would rise out of the blocks. The surface of the dry blocks, even the 4th downstream row, had oil that could be wiped off with a finger. After 5 modules had been washed, tank 1 looked like dense brown, turbid muddy water. Sludge had already begun to accumulate on the bottom of the tank. Tank 1 needed original charge plus 2 additional changes to wash 16 modules. We concluded that some of the washes were not effective, and may have re-contaminated the catalyst with oil. McMinnville Electric System Page 54 of 182 Before the start of the B20 testing, some modifications were made to improve the results. The sorption/regen cycle time was reduced to 3.5 minutes. Gaskets were added to the door to improve sealing on the front side of the catalyst as seen in Figure IV-10. Bars were added to the back wall of the catalyst chamber to straighten the wall and to act as a labyrinth seal as seen in Figure IV-11. On the plasma reformer, the catalyst media was re-activated using Ni salts, a short section of small diameter pipe at the reformer inlet was removed to increase the regen gas flow and hydrogen flow rate, and a small supply of water was added downstream of the reformer to add steam to the regen gas and increase the regen gas flow rate. Figure IV-11 Figure IV-10 Figure IV-9 Figure IV-8 McMinnville Electric System Page 55 of 182 B20 Test Figure IV-12 shows the engine emissions of NOx and CO when the engine is running 80% ultra- low sulfur diesel mixed with 20% biodiesel as fuel. Compared to Figure IV-6, note that NOx emissions decreased from an average of 724 ppm to an average of 619 ppm. Figure IV-13 and Figure IV-14 show the catalyst outlet emissions. The NOx removal for both chambers averages 97%. The revisions made to the system seem to have removed the discrepancy in NOx removal between the two chambers observed during installation and commissioning and earlier testing. The CO spike from regeneration is still higher for one chamber than the other, as observed during earlier testing, indicating that there is still a discrepancy in regen efficiency between the two chambers. The CO spike is significantly higher for this test than in previous testing. This may indicate that the catalyst is fully regenerated by the hydrogen in the regen gas and therefore consuming less CO or that there is less oxygen present in the chamber, which would allow the CO to oxidize to CO2. Figure IV-15 shows the oxygen depletion study as repeated on chamber RX100. The increased flow rate of regen gas has decreased the amount of time required for the chamber to be purged. In supplementary testing, the oxygen concentration in chamber RX100 was down to 0.5% in 1:50 minutes, and down to 0.0% in 2:20 minutes. For chamber RX200, the oxygen concentration was down to 0.5% in 3:02 minutes and down to 0.0% in 3:40 minutes. The minimum oxygen concentration achieved was 0%, which was not achieved in the earlier oxygen depletion testing. The oxygen concentration in chamber RX100 was below 2% by 54 seconds into the regeneration cycle. For this test, the plasma reformer was run for a total of 7 hours, 38 minutes total, with a switch to B50 about 5 hours into the test. Approximately 23 liters per hour of water was added as a secondary supply of steam to the regen gas. This rate of water injection far exceeded the amount of heat available to vaporize it (even if it had been finely atomized). As a result, water accumulated in the bottom of the regen outlet piping, increased the resistance of the system and eventually reduced the air flow rate. Small droplets of grey/black water were showering from the vent stack. During the B20 portion of the test, the regen gas was composed of 19.2% hydrogen, 19.8% carbon monoxide, 52% nitrogen, and 0.86% methane, and unmeasured concentrations of carbon dioxide and water vapor. McMinnville Electric System Page 56 of 182 Figure IV-12 Figure IV-13 Engine Exhaust Characteristics B20 - 7/9/05 0 100 200 300 400 500 600 700 800 1:45:07 PM 1:46:34 PM 1:48:00 PM 1:49:26 PM 1:50:53 PM 1:52:19 PM 1:53:46 PM 1:55:12 PM 1:56:38 PM 1:58:05 PM 1:59:31 PM 2:00:58 PM 2:02:24 PM 2:03:50 PM 2:05:17 PM 2:06:43 PM 2:08:10 PM 2:09:36 PM 2:11:02 PM TimeNOx & CO Concentration (ppmvd - uncorrected)NOX CO Average NOx - 619 ppm Average CO - 85 ppm 10.7% O2 EMx Stack Measurement from Both Reactor Chambers B20 - 7/9/05 0 100 200 300 400 500 600 700 2:32:38 PM 2:34:05 PM 2:35:31 PM 2:36:58 PM 2:38:24 PM 2:39:50 PM 2:41:17 PM 2:42:43 PM 2:44:10 PM 2:45:36 PM 2:47:02 PM 2:48:29 PM 2:49:55 PM 2:51:22 PM 2:52:48 PM 2:54:14 PM 2:55:41 PM 2:57:07 PM 2:58:34 PM 3:00:00 PM 3:01:26 PM TimeNOx & CO Concentration (ppmvd - uncorrected)NOX CO 3.5 min cycle time Engine Exhaust: 750F 619 ppmv NOx 592 ppmv NO 27 ppmv NO2 85 ppmv CO 10.7% oxygen McMinnville Electric System Page 57 of 182 Figure IV-14 Figure IV-15 EMx Stack Measurement from Both Reactor Chambers B20 - 7/9/05 0 100 200 300 400 500 600 700 2:32:38 PM 2:34:05 PM 2:35:31 PM 2:36:58 PM 2:38:24 PM 2:39:50 PM 2:41:17 PM 2:42:43 PM 2:44:10 PM 2:45:36 PM 2:47:02 PM 2:48:29 PM 2:49:55 PM 2:51:22 PM 2:52:48 PM 2:54:14 PM 2:55:41 PM 2:57:07 PM 2:58:34 PM 3:00:00 PM 3:01:26 PM TimeNOx Concentration (ppmvd - uncorrected)Stack NOX Engine NOx Engine Exhaust: 750F 619 ppmv NOx 592 ppmv NO 27 ppmv NO2 85 ppmv CO 10.7% oxygen 3.5 min cycle time 94.9 to 98.9% NOx Removal Oxygen Depletion RX100 0:03:23 0:03:08 0:04:11 0:03:55 0 1 2 3 4 5 6 7 8 9 10 11 12 00:00 00:15 00:29 00:44 00:59 01:13 01:28 01:43 01:58 02:12 02:27 02:42 02:56 03:11 03:26 03:40 03:55 04:10 04:24 Elapsed Time from Opening of Regen Supply Valve (mm:ss)Oxygen (vol%)7/9/2005 7/5/2005 Engine Exhaust: 725F 10.7% oxygen 3.5 min cycle time McMinnville Electric System Page 58 of 182 B50 Test Figure IV-16 shows the engine emissions of NOx and CO when the engine is running 50% ultra- low sulfur diesel mixed with 50% biodiesel as fuel. Figure IV-17 and Figure IV-18 show the catalyst outlet emissions. The NOx removal for both chambers averages 98%. The NOx removal is nearly identical between the two. The pressure drop from the engine exhaust pipe ahead of the reactor to just below the catalyst (during sorption) was measured to be 13.5” H2O on the RX100 chamber and 14.9” H2O on the RX200 chamber side. For this test, the plasma reformer was run for a total of 7 hours, 38 minutes total, with a switch to B50 about 5 hours into the test. During the B50 portion of the test, the regen gas was composed of 19.3% hydrogen, 18.2% carbon monoxide, 50.8% nitrogen, and 1.5% methane, and unmeasured concentrations of carbon dioxide and water vapor. The regen gas flow rate was calculated to be 39.9 scfm (91.3 acfm). The hydrogen flow rate was calculated to be 1.24 kg/hr. This flow rate and hydrogen production rate exceed the design specification. Figure IV-16 Engine Exhaust Characteristics B50 - 7/9/05 0 100 200 300 400 500 600 700 800 4:35:02 PM 4:36:29 PM 4:37:55 PM 4:39:22 PM 4:40:48 PM 4:42:14 PM 4:43:41 PM 4:45:07 PM 4:46:34 PM 4:48:00 PM 4:49:26 PM 4:50:53 PM 4:52:19 PM 4:53:46 PM 4:55:12 PM 4:56:38 PM 4:58:05 PM 4:59:31 PM TimeNOx & CO Concentration (ppmvd - uncorrected)NOX CO Average NOx - 649 ppm Average CO - 65 ppm 10.9% O2 McMinnville Electric System Page 59 of 182 Figure IV-17 Figure IV-18 EMx Stack Measurement for Both Reactor Chambers B50 - 7/9/05 0 100 200 300 400 500 600 700 3:50:24 PM 3:51:50 PM 3:53:17 PM 3:54:43 PM 3:56:10 PM 3:57:36 PM 3:59:02 PM 4:00:29 PM 4:01:55 PM 4:03:22 PM 4:04:48 PM 4:06:14 PM 4:07:41 PM 4:09:07 PM 4:10:34 PM 4:12:00 PM 4:13:26 PM 4:14:53 PM 4:16:19 PM 4:17:46 PM 4:19:12 PM TimeNOx & CO Concentration (ppmvd - uncorrected)NOX CO 3.5 min cycle time Engine Exhaust: 750F 644 ppmv NOx 625 ppmv NO 19 ppmv NO2 64 ppmv CO 10.9% oxygen EMx Stack Measurement for Both Reactor Chambers B50 - 7/9/05 0 100 200 300 400 500 600 700 3:50:24 PM 3:51:50 PM 3:53:17 PM 3:54:43 PM 3:56:10 PM 3:57:36 PM 3:59:02 PM 4:00:29 PM 4:01:55 PM 4:03:22 PM 4:04:48 PM 4:06:14 PM 4:07:41 PM 4:09:07 PM 4:10:34 PM 4:12:00 PM 4:13:26 PM 4:14:53 PM 4:16:19 PM 4:17:46 PM 4:19:12 PM TimeNOx Concentration (ppmvd - uncorrected)Stack NOX Engine NOx 3.5 min cycle time Engine Exhaust: 750F 644 ppmv NOx 625 ppmv NO 19 ppmv NO2 64 ppmv CO 10.9% oxygen 96.3 to 99.4% NOx Removal McMinnville Electric System Page 60 of 182 B100 Test Figure IV-19 shows the engine emissions of NOx and CO when the engine is running 100% biodiesel as fuel. The NOx concentration of 786 ppm is an 18% increase over the B50 NOx emissions of 649 ppm and a 25% increase over the ULSD emissions of 612 ppm. Figure IV-20 and Figure IV-21 show the catalyst outlet emissions. The NOx removal for both chambers averages 96.6%. The NOx emissions for one chamber are slightly higher than those for the other chamber. The CO spikes slightly higher than the previous test, and displays a less consistent pattern. For this test, the plasma reformer was run for a total of 5 hours, 4 minutes. The regen gas was composed of 19% hydrogen, 18.4% carbon monoxide, 54% nitrogen, and 0.83% methane, and unmeasured concentrations of carbon dioxide and water vapor. The hydrogen concentration remained steady throughout the test, varying from 17.5% to 20%. TVA had defined a target NOx emission level of 0.5 g/bhp-hr for the system, when running on B100. For this test, we calculated a NOx emission level of 0l21 g/BHP-hr, far exceeding the required reduction in NOx emissions. Figure IV-19 Engine Exhaust Characteristics B100 - 7/10/05 0 100 200 300 400 500 600 700 800 12:12:58 PM 12:14:24 PM 12:15:50 PM 12:17:17 PM 12:18:43 PM 12:20:10 PM 12:21:36 PM 12:23:02 PM 12:24:29 PM 12:25:55 PM 12:27:22 PM 12:28:48 PM 12:30:14 PM 12:31:41 PM 12:33:07 PM 12:34:34 PM 12:36:00 PM TimeNOx & CO Concentration (ppmvd - uncorrected)NOX CO Average NOx - 786 ppm Average CO - 41 ppm 10.9% O2 McMinnville Electric System Page 61 of 182 Figure IV-20 Figure IV-21 EMx Stack Measurement for Both Reactor Chambers B100 - 7/10/05 0 100 200 300 400 500 600 700 11:28:1 9 AM 11:29:4 6 AM 11:31:1 2 AM 11:32:3 8 AM 11:34:0 5 AM 11:35:3 1 AM 11:36:5 8 AM 11:38:2 4 AM 11:39:5 0 AM 11:41:1 7 AM 11:42:4 3 AM 11:44:1 0 AM 11:45:3 6 AM 11:47:0 2 AM 11:48:2 9 AM 11:49:5 5 AM 11:51:2 2 AM 11:52:4 8 AM 11:54:1 4 AM 11:55:4 1 AM 11:57:0 7 AM TimeNOx & CO Concentration (ppmvd - uncorrected)NOX CO 3.5 min cycle time Engine Exhaust: 750F 644 ppmv NOx 625 ppmv NO 19 ppmv NO2 64 ppmv CO 10.9% oxygen EMx Stack Measurement from Both Reactor Chambers B100 - 7/10/05 0 100 200 300 400 500 600 700 800 11:28:1 9 AM 11:29:4 6 AM 11:31:1 2 AM 11:32:3 8 AM 11:34:0 5 AM 11:35:3 1 AM 11:36:5 8 AM 11:38:2 4 AM 11:39:5 0 AM 11:41:1 7 AM 11:42:4 3 AM 11:44:1 0 AM 11:45:3 6 AM 11:47:0 2 AM 11:48:2 9 AM 11:49:5 5 AM 11:51:2 2 AM 11:52:4 8 AM 11:54:1 4 AM 11:55:4 1 AM 11:57:0 7 AM TimeNOx Concentration (ppmvd - uncorrected)Stack NOX Engine NOx 3.5 min cycle time Engine Exhaust: 750F 781 ppmv NOx 765 ppmv NO 16 ppmv NO2 42 ppmv CO 11.0% oxygen Average NOx 26.8 ppmv 96.6% NOx Removal Efficiency Target: Actual: 1.57 lb/MW-h .688 lb/MW-h .5 g/BHP-h .21 g/BHP-h McMinnville Electric System Page 62 of 182 Fuel Affects Figure IV-22 and Figure IV-23 show the affect on engine emissions of NOx and CO when the engine is running various blends of ULSD and Biodiesel. In Figure IV-22, the effect is shown as a % change from pure ULSD fuel. In Figure IV-23, the NOx emissions (in ppmvd) are shown as measured before and after the catalyst. Figure IV-24 shows the affect on regen gas composition when the plasma reformer is fueled by various blends of ULSD and Biodiesel. Figure IV-22 Effect of BioDiesel on Emissions Untreated Engine Exhaust 0% 20% 40% 60% 80% 100% 120% 140% 0%10%20%30%40%50%60%70%80%90%100% % BioDiesel Blended with ULSD% Change from ULSDCO NOx McMinnville Electric System Page 63 of 182 Figure IV-23 Effect of BioDiesel on Engine NOx Emissions 0 100 200 300 400 500 600 700 800 0 10 20 30 40 50 60 70 80 90 100 % BioDieselNOx Emissions (ppmvd - uncorrected)Engine After EMx NOTE: Other variables were changed in addition to fuel type such as sorption cycle time, regen gas flow rate, etc. McMinnville Electric System Page 64 of 182 Figure IV-24 Effect of Biodiesel on Regen Gas 0 10 20 30 40 50 60 0%10%20%30%40%50%60%70%80%90%100% % BiodieselConcentration (%)Hydrogen Carbon Monoxide Nitrogen Oxygen & Argon Methane NOTE: Other variables were changed in addition to fuel type such as operating temperature, water addition, air to fuel ratio, etc. McMinnville Electric System Page 65 of 182 Long Term Test Limiting Factors Throughout the short term testing, the plasma reformer was not run for a full 24 hours. To achieve a 1500 hour run, this hurdle must first be crossed. After run 28, the EMx catalyst was removed from the system and never tested again. The entire demo program shifted to developing the plasma reformer. Table V-1 and Table V-2 summarize the plasma reformer runs completed in an attempt to accomplish this goal. Reliable, stable operation of the plasma reformer was not accomplished for periods longer than 30 hours. Run Date Fuel Duration (H:MM) Reason Stopped Comments 17 7/21/05 B100 0:58 Head leaking, replace gasket & re-tap bolts Air in water line – short excursion to 1200˚C. Post-run inspection revealed no damage to Ni balls, electrodes, or thermowell. Applied high temp caulk, new grade 5 zinc coated steel bolts, trimmed damaged wires, re-connected ceramic connectors, trimmed damaged section of air line flex connection and re-attached, re- routed wires to avoid hot surfaces, bled air out of fuel and water lines. 18 7/22/05 B100 0:43 Air line leaking Pre-heat flex hose leaking 19a 7/25/05 B100 2:49 Engine tripped Engine fuel supply empty 19b 7/25/05 B100 5:17 Engine tripped 21 7/26/05 B100 14:54 Low air flow Air flow decreased from 21.6 scfm at 51 Hz at 7:04 to 18 scfm at 60 Hz at 21:21. Post run inspection revealed reformer full of carbon. Cleaned Ni balls, replaced granules 22 7/30/05 B100 0:33 Air line leaking Pre-heat flex line leaking 23 8/4/05 B100 1:30 Temps too high Temps too high for entire run. Water dripping at air inlet on lower side of reformer. 24 8/8/05 B100 3:58 Engine tripped Installed new thermowell 2” deep in Ni balls. Right electrode wire burned through, arcing to center pipe. 25 8/9/05 B100 7:43 Low air flow Air flow rates dropped from 21.6 scfm at 50 Hz at 10:45 to 18.2 scfm at 60 Hz at 5:40. McMinnville Electric System Page 66 of 182 Run Date Fuel Duration (H:MM) Reason Stopped Comments 26 9/21/05 B100 0:18 Liquids leaking around spark plugs Tightened plugs but couldn’t restart – both porcelain insulators cracked inside reformer, arcing to lid instead of electrodes. Replaced plugs, connectors, and sealed plugs with ceramic caulk. 27 9/22/05 B100 19:54 Longest run to date. Low air flow Targeting 4.9 air/fuel ratio and 900˚ - 1000˚ C per Dr. Tom Reed. Air flow rates dropped from 21.6 scfm at 50 Hz at 13:17 to 18.5 scfm at 60 Hz at 8:24. Engine shut down and went to flare at 9:50 to reduce backpressure. During this run, we used the temperature 1” above the reformer outlet as the control point, maintaining temperatures above 800˚ C minimum temperature to avoid the soot formation region described by Dr. Reed. This was the longest run to date, confirming the proper air/fuel/water ratios for soot-free operation. White/grey powder on Ni balls analyzed and determined to be due to minerals in water used for reformer. Demineralizer was installed before next run. 28 11/6/05 – 11/7/05 B100 30 hr Longest run to date. Fuel filter plugged As with run 27, we used the temperature 1” above the reformer outlet as control point. This was our most successful run, with extremely stable temperature and H2 production. Air flow very slowly decreasing from 22.5 scfm at 50 Hz at 16:28 on 11/6 to 21.6 scfm at 52 Hz at 19:17 on 11/7. Fuel flow decreasing from 2.81 ml/s at 58% full stroke at 17:01 on 11/6 to 2.47 ml/s at 75% full stroke at 22:40 on 11/7. System left unattended after above adjustment. Overnight, fuel filter plugged, mixture went lean and overheated. By 6:30am, system was cool, air and water were still on at full flow. Table V-1 At this point, the plasma reformer had suffered high temperature thermal damage. The reformer was completely redesigned and rebuilt with heavier gauge metals, more readily available components (e.g. spark plugs), larger internal capacities, and better fuel, air, and water mixing and atomization. McMinnville Electric System Page 67 of 182 During run 28, measurements were taken to determine if the EMx system is still performing as it did in the early days of testing. Figure V-1 shows the temperature profile inside the plasma reformer during runs number 27 and 28, the two most stable runs that were achieved. Figure V-1 850 900 950 1000 1050 1100 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Temperature (C) Plasma Reformer Temperature Profile Temperature Profile 85090095010001050110001234567891011121314151617181920212223242526Position in Reformer (inches)Temperature (C)Run 27 9/23/05 10:34 Run 27 9/22/05 16:10 Run 28 11/9/05 06:05 Run 28 11/9/05 12:00 Temperature Profile85090095010001050110001234567891011121314151617181920212223242526Position in Reformer (inches)Temperature (C)Run 27 9/23/05 10:34Run 27 9/22/05 16:10Run 28 11/9/05 06:05Run 28 11/9/05 12:00Granules Ni Balls McMinnville Electric System Page 68 of 182 Figure V-2 shows the oxygen depletion during run 28. The low minimum value indicates that there is little or no bypass around the catalyst modules. The amount of time that elapses before the minimum oxygen concentration is achieved is longer than in the tests run just after the catalyst and door seals were repaired. It is possible that the total regen gas flow rate is lower than during the earlier test. Figure V-2 After the catastrophic failure of the catalyst at the end of run 28, a new plasma reformer was designed as shown in Figure V-3 and Figure V-4. The design included a thicker walled vessel, with off-the-shelf components wherever possible. All components were specified to be available in the United States, in contrast with the original design, which contained electrode connectors only available in Europe. Oxygen Depletion During Regen - RX 100 11/9/05 0 2 4 6 8 10 12 0:00 0:14 0:28 0:43 0:57 1:12 1:26 1:40 1:55 2:09 2:24 2:38 2:52 3:07 3:21 3:36 3:50 4:04 4:19 4:33 Elapsed Time of Regen Cycle (m:ss)Oxygen Content %Regen Valve XV-102 Opens 0:05Isolation Valves in Position 0:10Regen Valve XV-202 Closed 0:15Isolation Valves Begin to Move 3:30Sampling / Instrument Lag Time Isolation Valves in Position 3:40 McMinnville Electric System Page 69 of 182 Figure V-3 McMinnville Electric System Page 70 of 182 Item Qty Description 1 1 316 SS Female Tee ¼” NPT x ¼” NPT (McMaster 4464K48) 2 2 316 SS Adapter ¼” NPT Male x ¼” YOR-LOK (McMaster 5182K111) 3 1 316 SS ¼” Pipe Nipple, Threaded 3 ½ “ Long (McMaster 4548K144) 4 1 Static Mixer, 3 ½” Long 5 2 316 SS Adapter ¼” NPT Female x ¼” Tube Socket Weld (McMaster 51255K302) 6 1 316 SS ¼” Tube, Thick Walled 7 1 Thermowell 8 1 316 SS Adapter 1/8” NPT Male x 1/8” YOR-LOK (McMaster 5182K804) 9 1 316 SS Forged Type Blind Flange 3” (McMaster 44695K118) 10 1 316 SS Forged Type Slip-On Flange 3” (McMaster 44695K38) 11 1 316 SS 1 ½” SCH 40 Pipe 12 1 Bete 1/4P28@5 303 SS Fogging Nozzle 13 1 316 SS 3” Pipe Nipple, Threaded 1 End, 6” Long (McMaster 9110T79) 14 1 316 SS Reducing Coupling, 3” x 1” Butt-Weld x Butt-Weld 15 1 316 SS 1” Pipe Union, Socket-Weld x Socket-Weld Figure V-4 McMinnville Electric System Page 71 of 182 Run Date Fuel Duration (H:MM) Reason Stopped Comments 29 10/9/06 B100 2:43 Unstable, not making hydrogen First restart after complete rebuild of system. Larger vessel, larger diameter fuel and water lines, additional filter on fuel, additional filter on water, larger pre-heat air line, fuel/water mixing chamber installed with atomizer, no Ni balls – only untreated alumina balls 30 10/10/06 B100 Unknown Temp spiked - 1343˚C Very unstable. Max fuel and water reached, could not push one without restricting other. Temp spike was 10” into balls. Melted thermowell. Lid wedged onto studs, had to be pried off. 31 10/26/06 B100 0:59 Leaking fuel caught on fire Oil leaking from mixing chamber onto lid. Max H2 – 14%, flare unstable. During cool down (very low air flow) H2 reached 22%. Temp higher in the middle of the reformer. Suspect air/water mixture was hitting walls and running down, resulting in worse atomization than original design 32 10/27/06 B100 1:23 Temperature spike to 1250˚C Added ¾” layer of Ni balls to top of reformer and moved spray nozzle to mixer throat. Reformer lid gasket leaking, small leak between mixing chamber and reformer. H2 never above 8%, flare unstable, pulsating sound noted. Temperature climbed suddenly and rapidly to 1250˚C at 3” into the balls. Thermowell melted at 18” from bottom, Ni balls meted at top, ceramic balls slumped 1” below bridge of Ni balls. Ceramic balls free flowing. Nozzle black w/baked on carbon and tar. McMinnville Electric System Page 72 of 182 Run Date Fuel Duration (H:MM) Reason Stopped Comments 33 10/28/06 B100 ~ 1 hr Too hot - 1264˚C at 18” from top of reformer. Polished mating surface between mixing chamber and reformer to eliminate leaks. Replaced Ni thermowell and cut new gasket for reformer lid. Replaced spray nozzle and coated alumina beads with Ni salt prior to run. Very unstable hydrogen concentration – 4% to 18%. When more water is added, the fuel pressure goes up, resulting in lower fuel flow. Also, increasing water flow rate reduced temps at top of bed but greatly increased them in the middle of the bed. Suspect exothermic water/gas shift reaction is occurring inside the media. Table V-2 After run 33, it was concluded that the plasma reformer instability had increased. Further progress will require extensive development efforts on the plasma reformer or an alternative source for regen gas. This may be due to the substitution of alumina balls for the fragile, but catalytically activated pumice stones, or the coating of the alumina balls with a nickel salt catalyst solution. Conclusions The EMx catalyst system performs at greater than 90% NOx removal, even with very high inlet NOx concentrations and operating temperatures of 750˚F. The EMx system eliminates the visible plume and significantly silences the engine exhaust. Summary of Results 1. NOx performance was high across all blends of biodiesel. 2. Exhaust bypass around the catalyst was discovered midway through the testing. This compromised several of the measurements made during commissioning and on the short- term fuel blend tests. 3. Unfortunately, most of the demo project focused on developing the plasma reformer, e.g., learning how to start up the system, how to operate it in a stable manner, how to avoid soot formation and high temperature extremes, how to retrofit it with safety interlocks, etc. Several design/rebuild retrofits were undertaken to improve the reliability of the reformer system and the durability of the hardware components. Ultimately, the plasma reformer proved to be unstable and unreliable and curtailed the studies of EMx performance. McMinnville Electric System Page 73 of 182 Future Work 1. Future studies should include long-term studies of the effects of biodiesel on the catalyst. 2. All future studies hinge on the availability of reliable technologies to produce regeneration gas. McMinnville Electric System Page 74 of 182 Attachment “C” (The following report, which pertains solely to the Caterpillar 3516B genset, was commissioned by McMinnville Electric System as part of this Project and is therefore incorporated and made part of this Final Report as submitted. The author is unknown) McMinnville Electric System Page 75 of 182 McMinnville Electric System Page 76 of 182 McMinnville Electric System Page 77 of 182 McMinnville Electric System Page 78 of 182 McMinnville Electric System Page 79 of 182 McMinnville Electric System Page 80 of 182 McMinnville Electric System Page 81 of 182 McMinnville Electric System Page 82 of 182 McMinnville Electric System Page 83 of 182 McMinnville Electric System Page 84 of 182 McMinnville Electric System Page 85 of 182 McMinnville Electric System Page 86 of 182 McMinnville Electric System Page 87 of 182 McMinnville Electric System Page 88 of 182 Attachment “D” McMinnville Electric System Page 89 of 182 McMinnville Electric System Page 90 of 182 McMinnville Electric System Page 91 of 182 Attachment “E” McMinnville Electric System Page 92 of 182 McMinnville Electric System Page 93 of 182 McMinnville Electric System Page 94 of 182 McMinnville Electric System Page 95 of 182 McMinnville Electric System Page 96 of 182 Attachment “F” (The following report, which pertains solely to emissions measurement testing, was commissioned by McMinnville Electric System as part of this Project and is therefore incorporated and made part of this Final Report as submitted.) McMinnville Electric System Page 97 of 182 McMinnville Electric System Page 98 of 182 McMinnville Electric System Page 99 of 182 McMinnville Electric System Page 100 of 182 McMinnville Electric System Page 101 of 182 McMinnville Electric System Page 102 of 182 McMinnville Electric System Page 103 of 182 McMinnville Electric System Page 104 of 182 McMinnville Electric System Page 105 of 182 McMinnville Electric System Page 106 of 182 McMinnville Electric System Page 107 of 182 McMinnville Electric System Page 108 of 182 McMinnville Electric System Page 109 of 182 McMinnville Electric System Page 110 of 182 McMinnville Electric System Page 111 of 182 McMinnville Electric System Page 112 of 182 McMinnville Electric System Page 113 of 182 McMinnville Electric System Page 114 of 182 McMinnville Electric System Page 115 of 182 McMinnville Electric System Page 116 of 182 McMinnville Electric System Page 117 of 182 McMinnville Electric System Page 118 of 182 McMinnville Electric System Page 119 of 182 McMinnville Electric System Page 120 of 182 McMinnville Electric System Page 121 of 182 McMinnville Electric System Page 122 of 182 McMinnville Electric System Page 123 of 182 McMinnville Electric System Page 124 of 182 McMinnville Electric System Page 125 of 182 McMinnville Electric System Page 126 of 182 McMinnville Electric System Page 127 of 182 McMinnville Electric System Page 128 of 182 McMinnville Electric System Page 129 of 182 McMinnville Electric System Page 130 of 182 McMinnville Electric System Page 131 of 182 McMinnville Electric System Page 132 of 182 McMinnville Electric System Page 133 of 182 McMinnville Electric System Page 134 of 182 McMinnville Electric System Page 135 of 182 McMinnville Electric System Page 136 of 182 McMinnville Electric System Page 137 of 182 McMinnville Electric System Page 138 of 182 McMinnville Electric System Page 139 of 182 McMinnville Electric System Page 140 of 182 McMinnville Electric System Page 141 of 182 McMinnville Electric System Page 142 of 182 McMinnville Electric System Page 143 of 182 McMinnville Electric System Page 144 of 182 McMinnville Electric System Page 145 of 182 McMinnville Electric System Page 146 of 182 McMinnville Electric System Page 147 of 182 McMinnville Electric System Page 148 of 182 McMinnville Electric System Page 149 of 182 McMinnville Electric System Page 150 of 182 McMinnville Electric System Page 151 of 182 McMinnville Electric System Page 152 of 182 McMinnville Electric System Page 153 of 182 McMinnville Electric System Page 154 of 182 McMinnville Electric System Page 155 of 182 McMinnville Electric System Page 156 of 182 McMinnville Electric System Page 157 of 182 McMinnville Electric System Page 158 of 182 McMinnville Electric System Page 159 of 182 McMinnville Electric System Page 160 of 182 McMinnville Electric System Page 161 of 182 McMinnville Electric System Page 162 of 182 McMinnville Electric System Page 163 of 182 McMinnville Electric System Page 164 of 182 McMinnville Electric System Page 165 of 182 McMinnville Electric System Page 166 of 182 McMinnville Electric System Page 167 of 182 Attachment “G” (The following report, which pertains solely to emissions measurement testing and fuel quality was submitted by Ralph Boroughs on behalf of the Tennessee Valley Authority as part of this Project and is therefore incorporated and made part of this Final Report as submitted.) McMinnville Electric System Page 168 of 182 TVA Activities in Support of the McMinnville BioDiesel / SCONOX Project Introduction TVA’s interest in this project stemmed from a desire to develop clean, renewable generation, and to be responsive to customer proposals; in this case, a proposal from McMinnville Electric System (MES) to install a new generator, dedicated to biodiesel fuel, and equipped with an advanced NOx removal system. As a ‘power production’ project, TVA was not interested in any new diesel generation, primarily because of the high NOx emissions of diesels. Biodiesel, while cleaner burning in many respects, has not been shown to reduce NOx, and in some cases even increase NOx emissions. (TVA’s policy “not to pursue or consider new contracts for use or installation of additional diesel generation”, was established February 22, 2001, and was re-affirmed in July 2002 after reviewing the MES proposal.) Nevertheless, TVA did approve our participation in a ‘research and development’ project to test advanced emissions controls, using biodiesel in a diesel-generator set. The TVA role in the overall project was focused on primarily on emission monitoring, to verify that the NOx removal system could meet TVA emissions targets. A target NOx limit was set at 0.16 pounds per million BTU, 1.57 pounds per MWh, or 0.5 grams per brake horsepower-hour, based on TVA’s internal environmental review in 2002. This limit corresponds to emissions achievable by a well-controlled coal- fired plant, but is much stricter than any then-existing emissions standards for diesels. TVA’s NOx target has no direct linkage to EPA or Tennessee regulations for diesels, but is instead a goal, which, if met, might motivate and inform a reconsideration of TVA’s policy on diesel generation. [TVA did not participate in emissions testing for compliance with state air permits, because state- approved testing of diesels requires specialized equipment as well as detailed knowledge of state regulations and how they are applied.] Although TVA was primarily focused on emissions monitoring, TVA also contributed to the project planning process and fuel testing. Test Plan The actual testing schedule did not follow the planned schedule. Phases 1 through 3 were compressed to compensate for time and cost overruns earlier in the project. Phase 4 was compressed further, because problems with the plasma reformer (which we had hoped to resolve in phase 1) kept recurring, and attempts to resolve these problems added additional costs. The table below compares an early test plan to the actual implementation. McMinnville Electric System Page 169 of 182 Test Phase Description Planned Duration Actual Duration 1 Commissioning, Start up, & Regeneration System Tests 10 days ~4.5 days 6/29-7/3 2 Initial Reactor Testing & Characterization 5 days ~1.5 days 7/3-7/5 3 Fuel Blend Tests, ULSO, B2, B5, B20, B50, B100 8 Days 5 days 7/6-7/10 4 B100 Long-term Tests 1500 hours 1000 hours Note that optimization of the EMx reactor (phase 2) continued into the blend testing (phase 3). Emissions Test Equipment TVA used a PACE 400 electrochemical analyzer, by Ametek, which simultaneously measures NO, NO2, SOx, O2, CO, CO2 and hydrocarbons. The PACE 400 determines the contents of a flue gas stream by continually extracting samples that are routed into a peltier-effect (thermo-electric) cooling system, which removes water vapor from the sample before the sample is passed over electrochemical gas sensors for analysis. The probe and hose assembly contain an integrated thermocouple and water trap, to measure the stack temperature and pressure. A laptop computer was used to capture data from the data analyzer. Before use of the analyzer, TVA Environmental Technologies Group3 performed characterization tests with synthetic stack gas to determine the linearity, precision, stability, and response time for NO, NO2, CO, SO2 and O2. The study found that at a 15-second sample scan the linearity and precision were very good with values of less than 1 percent of the test concentration levels. This was better than the factory specifications. As for response time, 90 percent of a step change was registered for NO, NO2, CO, SO2 and O2 within15 to 30 seconds. 3 “Report on laboratory testing of THERMOX PACE2 400 analyzer with synthetic stack gas mixtures”, Ralph J. Valente and Vince Van Pelt, TVA, Environmental Technologies, April 6, 2005 McMinnville Electric System Page 170 of 182 The system was calibrated by utilizing standardized calibration gases. The accuracy of the sensor readings are displayed below: O2 ± 0.3% of gas concentration CO2 ±5% of reading, 0-2000 ppm ± 10% of reading, 2001-40,000 ppm NO ±5% or ± 5 ppm of reading, 0-2000 ppm NO2 ±5% or ± 5 ppm of reading, 0-500 ppm SO2 ±5% or ± 10 ppm of reading, 0-2000 ppm Combustibles ±5% of full scale Pressure .±2% of reading of or ±0.05 millibar whichever is greater Stack Temperature ±4 oF between 32 and 255 oF ±6 oF between 256 and 480 oF ±8 oF between 481 and 752 oF Calibrations were also done regularly during field tests. Emissions were monitored before and after the EmeraChem EMx Catalyst system. Two ports, at right angles were available for use at each test plane. In early runs, traverses were conducted to look for signs of flow stratification. Baseline emissions tests were conducted on Ultra-Low Sulfur Fuel Oil (ULSFO) and then emissions tests were conducted with nominal blends of 2%, 5%, 20% and 50 % biodiesel, or B2, B5, B20 and B50. Finally, emissions tests were done with 100% biodiesel or B100. In presenting test results, the raw measurements (here, in ppm or parts per million) may be thought to be the most reliable, because these can be tied directly to the instrument calibration. However, any air in-leakage into the sampling train can dilute the concentration, and engine controls may change the air McMinnville Electric System Page 171 of 182 to fuel ratio. Furthermore, the concentration has no direct relevance to the environmental impact, which is more a function of the pollutant mass flow. The best compromise between usefulness and accuracy is usually found by expressing the emissions results as mass per unit of fuel-energy-input. Thus, power plant measurements are usually expressed as pounds per million BTU. The pollution load can also be expressed as mass per unit power output. This has the virtue of encouraging efficient use of fuel, but introduces some additional uncertainty, due to the need to accurately measure efficiency, which depends on both fuel input flows and energy output. To ensure these additional measurements are accurate, one must usually operate at a fixed condition over a long period of time, which proved impractical for our tests. Units for these measurements are typically expressed as pounds per MWh for large plants, and grams per brake horsepower hour for engine manufacturers. The information is presented below in both pounds per million BTU and grams per brake-horsepower- hour, but is always calculated from the fuel-input. When showing output-based numbers, a fixed efficiency is assumed for convenience, although we recognize that actual efficiency will vary, depending primarily on ambient air temperature and pressure. Results Baseline Tests Testing began on July 6, 2005, with baseline tests, using ultra-low sulfur (<15 ppm) petroleum diesel. (This was before the <15 ppm Sulfur fuel was widely available, starting in September of 2006.) Untreated engine emissions were about 3.1 g/hp-h, well within the manufacturer’s specification for NOx emissions of <4.45 g/hp-h. As discussed by EmeraChem (see Attachment “B”, pages 47 and 53) the NOx removal in this test was lower than expected, due to significant leakage of untreated exhaust around the catalyst. This leakage was subsequently discovered and corrected prior to the B20 tests. Despite these leaks, the time averaged NOx emissions (0.131 lbs/million BTU) met TVA’s goal (<0.16 lbs/million BTU), although peak concentration (0.266 lbs/million BTU) did not. Results are shown on the next page. The saw-toothed pattern reflects the regeneration cycle, while the overall upward trend reflects incomplete regeneration, especially in chamber Rx 200. Causes for this incomplete regeneration include the bypass leakage flow and the lower than expected hydrogen output from the reformer (for more details see Attachment “B”, page 53). The alternating pattern of high and low NOx peaks reflects the greater bypass flow in chamber Rx200, compared to Rx100. McMinnville Electric System Page 172 of 182 Again as grams per brake horsepower-hour: NOx vs Time Ultra Low Sulfur Petrol. Fuel, July 6, 2005 0 0.2 0.4 0.6 0.8 1 1.2 0:00 0:10 0:20 0:30 0:40 0:50 1:00 1:10 1:20 1:30 1:40 1:50 2:00 2:10 2:20 2:30 Time, hr:minNOx, pounds per million BTUUpstream NOx Goal, 0.16 Final NOx NOx vs Time Ultra Low Sulfur Petrol. Fuel, July 6, 2005 0 0.5 1 1.5 2 2.5 3 3.5 4 0:00 0:10 0:20 0:30 0:40 0:50 1:00 1:10 1:20 1:30 1:40 1:50 2:00 2:10 2:20 2:30 Time, hr:minNOx, grams/(bhp-hr)Upstream NOx Goal, 0.5 Final NOx McMinnville Electric System Page 173 of 182 B2 BioDiesel Blend Testing continued on July 6, with a nominal 2% blend of BioDiesel in Ultra-Low Sulfur Petroleum Diesel. The final NOx averages about 0.24 pounds per million BTU, well above the TVA target. Reasons for this poor performance include lower than expected hydrogen output from the reformer, leakage flow bypassing the catalyst, and incomplete washing of the catalyst. More details are given in EmeraChem’s report (see Attachment “B”, page 53). Based on a 2002 study by EPA (EPA420-P- 02-001), we had expected the untreated NOx emissions to increase by about 10% for B100, and a proportionately smaller increase (<1%) for B2. In contrast, our B2 test showed a 6.87% increase. While still well below the manufacturer’s specification, untreated NOx emissions increased from 3.07 g/hp-hr for petroleum diesel to 3.28 g/hp- hr. for B2. NOx vs Time B2 Tests, July 7, 2005 0 0.2 0.4 0.6 0.8 1 1.2 0:00 0:10 0:20 0:30 0:40 0:50 1:00 1:10 1:20 1:30 1:40 1:50 2:00 Time, hr:minNOx, pounds /million BTUUpstream NOx Goal, 0.16 Final NOx McMinnville Electric System Page 174 of 182 B5 BioDiesel Blend Later the same day, July 6, testing began on a nominal 5% BioDiesel blend. Results are shown below. Again, it is clear that performance is deteriorating, and the same causes are suspected, namely: lower than expected hydrogen output from the reformer, leakage flow bypassing the catalyst, and incomplete washing of the catalyst. Obviously, the NOx removal goal was generally not met, although it was met periodically for a few minutes. After completion of the B5 test, EmeraChem re-washed the catalyst and reworked the catalyst chamber. Bypass flow was restricted by the use of bars to straighten and stiffen the back wall of the catalyst chamber and act as a labyrinth seal. Gaskets were added to the door to improve sealing on the front side. The reformer was also modified and refurbished. Details are given in Attachment “B”, page 54. NOx vs Time B5 Tests, July 7, 2005 0 0.1 0.2 0.3 0.4 0.5 0.6 0:00 0:10 0:20 0:30 Time, hr:minNOx, pounds per million BTUGoal, 0.16 Final NOx McMinnville Electric System Page 175 of 182 B20 BioDiesel Blend Testing Testing resumed on July 9, using a nominal 20% BioDiesel blend. NOx removal results were excellent, and far exceeded our goal NOx vs Time B20 Tests, July 9, 2005 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0:00 0:10 0:20 Time, hr:minNOx, pounds /million BTUGoal, 0.16 Final NOx McMinnville Electric System Page 176 of 182 B50 BioDiesel Blend Testing Testing continued on July 9, using a nominal 50 % BioDiesel blend. Again, NOx removal results were excellent, and far exceeded our goals. NOx vs Time B50 Tests, July 9, 2005 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0:00 0:10 0:20 Time, hr:minNOx, pounds /million BTUGoal, 0.16 Final NOx McMinnville Electric System Page 177 of 182 B100 BioDiesel Testing Testing continued on July 9, using 100 % BioDiesel. Again, results were excellent, and the TVA target was met. The results are displayed on the in two forms: pounds/million BTU, and grams per brake- horsepower-hour. The untreated NOx emissions average 4.23 g/hp-hr, using B100. This compares to 3.075 g/hp-hr using Ultra-Low Sulfur Petroleum Diesel, and 3.07 g/hp-hr using B2. NOx emissions are apparently a non- linear function of the percentage biodiesel, and using B100 increases the NOx emissions by about 38% compared to petroleum diesel. (Recent studies have shown that biodiesel does not increase NOx emissions in some engines, and this may depend on the engine’s timing adjustment.) NOx vs Time B100 Tests, July 10, 2005 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0:00 0:10 0:20 0:30 0:40 0:50 1:00 1:10 1:20 Time, hr:minNOx, pounds per million BTUUpstream NOx Goal Final NOx McMinnville Electric System Page 178 of 182 TVA had planned to follow-up with two additional emissions test campaigns during the B100 endurance runs, but these tests were postponed, and then cancelled because of the reformer failure. Nevertheless, we conclude that if a reliable reformer or other source of hydrogen can be developed, a system similar to that demonstrated here should be capable of meeting TVA goals for NOx emissions control, when fueled with any blend of biodiesel and low sulfur petroleum diesel. NOx vs Time B100 Tests, July 10, 2005 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0:00 0:10 0:20 0:30 0:40 0:50 1:00 1:10 1:20 Time, hr:minNOx, grams/(bhp-hr)Upstream NOx Goal, 0.5 Final NOx McMinnville Electric System Page 179 of 182 Fuel Characterization Tests In order to make the stoichiometric calculations required for emissions analysis, one must first analyze the fuel to determine an empirical formula. For the fuels used in this work, the major constituents are carbon, hydrogen and oxygen. This analysis was done by TVA’s central laboratory, using standard methods. The results are tabulated below: Sample Description ULSFO B02 B05 B20 B50 B100 Carbon wt.% 85.2 85.06 84.57 83.08 82.65 76.23 Hydrogen wt.% 13.83 13.75 13.55 13.26 13.1 11.85 Nitrogen wt.% <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 BTU/lb 19801 19674 19649 19212 18679 16725 Ash <0.001 <0.001 <0.001 0.001 0.003 -- Density g/ml 0.8195 0.8205 0.8219 0.8305 0.8398 0.8785 Oxygen by dif, wt % 0.97 1.19 1.88 3.659 4.247 11.92 Oxygen content is plotted below vs. the nominal percentage BioDiesel. It appears that the B5 and B20 blends were higher than nominal BioDiesel, while the B50 was lower. McMinnville Electric System Page 180 of 182 Samples were sent to two labs to help understand this situation. Findings are tabulated below: Sample Name Nominal Blend Magellan Minnesota % Methanol MG-B 2% 1.40% 1.87% MG-D 5% 11.00% 4.85% MG-C 20% 25.10% 20.76% MG-A 50% 32.60% 36.96% MG-E 100% 93.20% 96.95% ~3% Jim Hedman of the State of Minnesota’s Department of Commerce, Weights and Measures Division, was able to identify a problem with the B100 blend; it contained about 3% Methanol! (His analysis is detailed in Attachment “A”) With levels this high, the samples would not have passed the flash point Oxygen vs BioDiesel % 0 2 4 6 8 10 12 14 0 20 40 60 80 100 BioDiesel %OxygenOxygen Linear (Oxygen) McMinnville Electric System Page 181 of 182 test. Taking into account for the methanol effect, the blend ratios appear to have been reasonably accurate, except for the B50, which was clearly an outlier. Because the methanol issue heightened our sensitivity to the issue, we began routinely gathering biodiesel samples and having them analyzed for conformity to a subset ASTM standard D-6751. As discussed earlier (Report, page 8), samples were frequently found to have excess glycerin. Excess methanol was not a problem in later samples. McMinnville Electric System Page 182 of 182 This Page Left Intentionally Blank