HomeMy WebLinkAboutIgiugig Wind Resource Feasibility Study and Conceptual Design Report - Mar 2020 - REF Grant 7071072INTERGRID
164 Hill Road, PO Box 48, Temple, NH 03084 ~ (603)801-4749 ~ www.Intergrid.US
Josi Hartley
Assistant Project Manager
Alaska Energy Authority
813 West Northern Lights Blvd.
Anchorage, AK 99503 March 11, 2020
Igiugig Wind Resource Feasibility Conceptual Design
Grant Agreement Number 7071072 * Project 410098 * Proposal RE71072
Dear Josi, I am submitting this final report on behalf of the Igiugig Village Council (IVC).
This report summarizes work done to determine the feasibility of wind power for the village of Igiugig.
It includes an analysis of turbine and energy storage options and an initial conceptual design.
In response to your comments and questions on the draft report, I have responded line by line to the
AEA wind study guidelines, and attached same as an appendix to this report.
Please call or email if you have further questions -
Sincerely
Dr. Robert Wills, P.E.
President
Intergrid, LLC
rwills@intergrid.us
(603)801-4749 (mobile)
y
Assistant Project Manager
Page | 1
Table of Contents
1. Background...................................................................................................................................... 2
2. Data Collection................................................................................................................................ 3
3. Igiugig Wind Resource.................................................................................................................... 3
Annual Variation................................................................................................................................. 6
Wind Shear.......................................................................................................................................... 8
Wind Direction.................................................................................................................................... 9
4. Village Electrical Demand............................................................................................................... 9
5. Generator Dispatch and Operating Costs....................................................................................... 12
6. Wind Turbine Options ................................................................................................................... 15
Bergey 15.......................................................................................................................................... 16
QED 20 ............................................................................................................................................. 16
Eocycle EO25 ................................................................................................................................... 16
North Wind 100................................................................................................................................ 17
Xant M.............................................................................................................................................. 17
7. Wind Turbine Performance............................................................................................................ 17
8. Wind Turbine Economics.............................................................................................................. 18
9. Wind Turbine Location.................................................................................................................. 19
10. Geotechnical Studies & Foundation Cost.................................................................................. 21
Turbine Tower Foundations.............................................................................................................. 22
11. Multiple Turbines, MHK & Microgrid Control......................................................................... 23
12. Energy Storage System .............................................................................................................. 26
13. Dispatchable Thermal Loads......................................................................................................27
14. Power Quality Issues.................................................................................................................. 28
15. Siting and Permitting Issues....................................................................................................... 28
16. Eocycles Installed Fleet and Commercial Outlook.................................................................... 29
17. Ballasted Foundation for the EO25............................................................................................ 30
18. Reduced costs of QED turbines ................................................................................................. 30
19. Recommendations...................................................................................................................... 30
Best and Worst Case Scenarios......................................................................................................... 31
Appendix A - Response to Alaska Wind Program Guidelines for Conceptual Design Reports .............. 33
A1 - Wind Resource Study....................................................................................................................... 33
A2 - Existing Electrical System Overview............................................................................................... 35
A3 - Heat Loads Overview....................................................................................................................... 41
A4 - Compiling the Final Conceptual Design Report...............................................................................42
Page | 2
Igiugig Wind Resource Feasibility Conceptual Design
Grant Agreement Number 7071072 * Project 410098 * Proposal RE71072
Final Report (March 11, 2020, Rev B)
1. Background
This project was initiated under Grant Agreement # 7071072 in July, 2014 to the Igiugig Village
Council (IVC). Initial work on wind resource site assessment was done by Knight-Piesold engineering
in 2012 in conjunction with the Lake and Peninsula Borough. Direction for this current grant work was
provided by Rich Stromberg of AEA. He stated in his summary:
Although not exceptionally high, the wind resource is generally good near Igiugig and certainly better
than the wind resource model predicts – by almost two wind classes. If the energy demand were higher
in the community, there might be benefit in searching for a stronger wind turbine site nearby. Instead,
the low average electrical load and small heating oil loads create the major challenges of a successful
integrated wind energy project that creates a long term economic benefit for the community.
The conceptual design phase should focus on studying varying configurations (turbine models,
quantities and sizes coupled with secondary loads, storage or power set point control) that make best
use of the excess energy available from variable wind power with a high percentage contribution to an
isolated diesel microgrid. Additional heat loads need to be identified in the community and they must be
served in the most economic manner that meets the technical requirements of the local energy system.
Varying amounts of energy storage (batteries, ultra-capacitors, other) should be considered, but these
may not be economically feasible options until storage costs come down sometime in the future.
A suitable wind turbine site should be identified early. The met tower site would have been acceptable
prior to the community expansion out near the landfill. Now, occupied homes and workplaces near the
landfill are within the hazard and annoyance zone surrounding a community-scale wind turbine. It is
recommended the turbine sites further east-southeast along the higher terrain be studied so long as they
are located inside village corporation land and avoid Native allotments.
Copies of referenced documents are available in the following Google Drive folder:
https://drive.google.com/open?id=1mOcAOCDithUPBJtETF8T1rr1blLpiGLv
Intergrid, LLC, a consulting engineering company, has been involved with power system research and
development at Igiugig since 2016. We were first engaged to support the installation of a marine-hydro-
kinetic (MHK) water turbine at the village with Ocean Renewable Power Company (ORPC). This
project, using an ABB variable speed drive to control the Rivgen turbine, was successful. Later we were
asked to design power electronics and integration strategies for a vertical wind turbine (VAWT) array in
the village. A two turbine system was demonstrated in 2017, but the machines themselves proved to be
unreliable. We also were asked to help with repairs to the 4 out of 6 Skystream wind turbines at the site.
Intergrid developed the in-nacelle inverter for the 2.5 kW Skystream turbine for Southwest Windpower.
This work was completed in November, 2019 and yielded useful information on power quality at the
village.
In 2018, Intergrid joined with ORPC and was awarded an EETF grant to develop a microgrid power
system for Igiugig incorporating a MHK turbine and battery energy storage.
fpgp g
Additional heat loads need to be identified in the community and they must be gfyy
served in the most economic manner that meets the technical requirements of the local energy system.
f ygy
recommended the turbine sites further east-southeast along the higher terrain be studied so long as they f gg
are located inside village corporation land and avoid Native allotments.
gp g g ( ) y
A two turbine system was demonstrated in 2017, but the machines themselves proved to berg
unreliable. e 4 out of 6 Skystream wind turbines at the site.
,p
Page | 3
Also in 2018, Intergrid received the grand prize laboratory testing award in the Alaska Center for
Microgrid Technologies (at UA-Fairbanks) commercialization competition. Research was completed in
their simulation lab to fully model the Igiugig power system including the MHK turbine, battery energy
storage and diesel generators (Figure 1).
Figure 1 – ACEP Modeling Block Diagram
2. Data Collection
Information collected for the wind study has included:
- Powerhouse data (including village electric load) from the village on three occasions
- Village electric and fuel records
- Skystream wind turbine production data from the Alvirez residence
- Studies of power quality problems
- Reviews of potential dispatchable heat loads
- A visit with National Renewable Energy Lab (NREL) staff in January, 2019 to map out a long
term energy strategy for the village.
- Discussions with village administration regarding turbine siting and dispatchable thermal loads
- Preliminary wind turbine energy production and cost analysis
3. Igiugig Wind Resource
The Knight-Piesold (K-P) wind study performed in 2012 provides a solid basis for understanding the
Igiugig wind resource. A 33 meter meterological tower was erected and a year of high quality wind data
was collected. A Computational Fluid Dynamics (CFD) model was used to determine wind variation at
potential sites. The main conclusion was that the majority of the village is a class 3 wind site (6.4 – 7.0
m/s at 50 m height)1. The optimal site for turbine power production is to the east of the village on the
lake shore, however this is not viable due to native allotment land ownership, planned development and
FAA permitting issues (as it is directly in line with the runway).
1 https://rredc.nrel.gov/wind/pubs/atlas/tables/1-1T.html
gg()p p
their simulation lab to fully model the Igiugig power system including the MHK turbine, battery energy yg
storage and diesel generators (Figure 1).
jy g (
The optimal site for turbine power production is to the east of the village on the g)p p p g
lake shore, however this is not viable due to native allotment land ownership, planned development and ,
FAA permitting issues (as it is directly in line with the runway).
Page | 4
The wind site is an IEC wind turbulence class IV 2 (average wind speed <= 6 m/s and max 50 year peak
< 50 m/s). The K-P study saw a maximum wind speed of 28 m/s.
Figure 2 – Knight Piesold Wind Speed Modeling Output
The met tower location for the K-P study was at the village dump, south of the landing strip and adjacent
to the greenhouse (near “Site 1” above).. The results from the met tower gave an annual average wind
speed of 5.87 m/s and a wind distribution shown in Figure 3.
2 https://en.wikipedia.org/wiki/IEC_61400
–Knight Piesold Wind Speed Modeling Output
Page | 5
Figure 3 – K-P Study Wind Speed Distribution
Figure 3 shows both the measured frequency, and a calculated probability curve with the same average
wind speed (A Weibull distribution with k=2, which is a Raleigh distribution). The peak is higher for the
measured data, and also there is less time in the “zero” wind speed bin. A model using the measured
distribution will produce about 30% more energy than one using a Raleigh distribution estimate.
This wind speed distribution can be used to estimate wind turbine output using the “method of bins”3:
Power curves can be used to estimate the annual energy production using the “Method of Bins.”
The Method of Bins takes power production at each wind speed and multiplies it by the hours per
year the wind blows at that wind speed; this results in an energy “bin” for each different wind
speed. The total energy output is calculated by adding the energy production in all bins.
3 https://www.e3a4u.info/energy-technologies/small-wind/estimating-energy-production/
0.00%
2.00%
4.00%
6.00%
8.00%
10.00%
12.00%
14.00%
16.00%
0 5 10 15 20 25 30Wind Speed Probability (%)Wind Speed (m/s)
Met Tower
Weibull
Page | 6
Bin m/s Start End Sum Percent Weibull Weighted
0 0 0.49 436 0.83% 4.60% 0.00
1 0.5 1.49 2582 4.91% 8.57% 0.05
2 1.5 2.49 4393 8.36% 11.43% 0.17
3 2.5 3.49 6042 11.49% 12.92% 0.34
4 3.5 4.49 7055 13.42% 13.07% 0.54
5 4.5 5.49 7624 14.50% 12.11% 0.73
6 5.5 6.49 6291 11.97% 10.40% 0.72
7 6.5 7.49 4917 9.35% 8.35% 0.65
8 7.5 8.49 3512 6.68% 6.29% 0.53
9 8.5 9.49 2563 4.88% 4.47% 0.44
10 9.5 10.49 1928 3.67% 3.00% 0.37
11 10.5 11.49 1299 2.47% 1.90% 0.27
12 11.5 12.49 1064 2.02% 1.14% 0.24
13 12.5 13.49 795 1.51% 0.65% 0.20
14 13.5 14.49 658 1.25% 0.35% 0.18
15 14.5 15.49 498 0.95% 0.18% 0.14
16 15.5 16.49 311 0.59% 0.09% 0.09
17 16.5 17.49 225 0.43% 0.04% 0.07
18 17.5 18.49 160 0.30% 0.02% 0.05
19 18.5 19.49 86 0.16% 0.01% 0.03
20 19.5 20.49 67 0.13% 0.00% 0.03
21 20.5 21.49 29 0.06% 0.00% 0.01
22 21.5 22.49 15 0.03% 0.00% 0.01
23 22.5 23.49 8 0.02% 0.00% 0.00
24 23.5 24.49 4 0.01% 0.00% 0.00
25 24.5 25.49 4 0.01% 0.00% 0.00
26 25.5 26.49 1 0.00% 0.00% 0.00
27 26.5 27.49 2 0.00% 0.00% 0.00
28 27.5 28.49 1 0.00% 0.00% 0.00
29 28.5 29.49 0 0.00% 0.00% 0.00
30 29.5 30.49 0 0.00% 0.00% m/s 0.00
Total 52570 100.00% Avg 5.87
Table 1 – Igiugig Met Tower Wind Speed Distribution
Annual Variation
The annual variation of wind speed is important for economic analysis as the village electric and heat
loads vary dramatically by season, and generation needs to be sized according to load.
The K-P report, using one year of data, showed the following variation:
Page | 7
Figure 4 – K-P Average Wind Speed by Month
There is another source of long-term wind data at the village. The Alvarez family installed a Skystream
wind turbine in late 2010 and have recorded daily production through 2017. This turbine is installed on a
23 m (75 ft) tower and so is more representative of available wind than the 10 m anemometer at the
airport (which is another source of wind data).
It is difficult to see a pattern in the Skystream daily data (Figure 5):
Figure 5 – Alvarez Five-Year Average and Maximum Wind Speed by Day
Month averages present a clearer picture (Figure 6). The wind speeds shown are calculated using the
cube-root of energy production, normalized to produce an average annual wind speed of 5.87 m/s (Table
2).
The Alvarez family installed a Skystreamgg
wind turbine in late 2010 and have recorded daily production through 2017.
Page | 8
Figure 6 – Month Average Wind Speed and Production
Month kWh/Day WS Estimate
Jan 13.7 6.53
Feb 13.3 6.47
Mar 8.5 5.57
Apr 10.8 6.03
May 9.9 5.86
Jun 8.4 5.54
Jul 6.5 5.10
Aug 6.5 5.09
Sep 9.2 5.72
Oct 10.7 6.01
Nov 11.2 6.11
Dec 12.8 6.39
Average 10.1 5.87
Table 2 – Seasonal Wind Variation from Skystream Data
Wind Shear
The K-P report showed a wind shear exponent of 0.237 which is typical for low, sparse forest.
The wind shear exponent Į extrapolates the wind speed (V1) at a measured height (H1) to a different
height (V2 atH2):
ܸଶ
ܸଵ
=൬ܪଶ
ܪଵ
൰
ן
For instance, a wind turbine with a hub height of 80 ft would experience 5% less average wind speed
than a turbine at 100 ft (the met tower height). This would result in about 15% less annual power
production due to the cubic relationship of wind power to velocity.
Figure 6 –Month Average Wind Speed and Production
Page | 9
Wind Direction
Knight-Piersold reported the following in their wind study:
Figure 7 – Mean Wind Speed Figure 8 – Wind Direction Frequency
The prevailing wind direction is from the east, with the highest mean wind speed from that direction
(Figure 7), and a predominance of measurements also from the east (Figure 8);
Figure 9 – Wind Energy Rose
The most important result is however the wind energy rose shown in Figure 9. It is calculated from wind
direction weighted by the cube of the wind speed. It shows that for power generation, the easterly winds
are all that matter. This impacts turbine siting, as ideally multiple turbines should be positioned in a line
perpendicular to the prevailing wind (i.e., north-south) and spaced at least 2 rotor diameters apart.
Turbines should be spaced 10 rotor diameters from upstream obstacles and other turbines.
4. Village Electrical Demand
The village has maintained accurate electricity and fuel records from 2011 to the present. Figure 10
shows village electric demand (kWh/month) for the 7 year period from 2011 to 2018.
Page | 10
Figure 10 – Village Electric Demand (kWh/Month)
Figure 11 shows annual demand overlaid by year:
Figure 11 – Annual Electric Demand
Village records also record peak electric demand by month from the station production meter (Figure
12).
Figure 10 –Village Electric Demand (kWh/Month)
Page | 11
Figure 12 – Peak Electric Demand
The overlaid version of this graph shows the seasonal nature of peak load.
Figure 13 – Seasonal Peak Load by Month
The implication of peak load reaching 100 kW is that either a 100 kW or more inverter will be required
for “diesels off” operation, or else some load control measures will need to be implemented. Intergrid
installed a “soft start” VFD on the main 2 HP school ventilation fan in 2018 and another on the 3 HP fan
for the gym in 2019. This may significantly reduce peak load as these two motors are the largest in the
village, and they previously started direct on line. We also have 10 minute average load data from the
50
55
60
65
70
75
80
85
90
95
100
2012 2013 2014 2016 2017 2018Peak Demand (kW)Figure 13 –Seasonal Peak Load by Month
q
Intergrid p, p g
installed a “soft start” VFD on the main 2 HP school ventilation fan in 2018 and another on the 3 HP fan
for the gym in 2019.
Page | 12
powerhouse SCADA system. Unfortunately recent data is not complete, but four months of data is
available for the beginning of 2017 (Figure 14).
Figure 14 – 2017 10 Minute Average Load, and Generator Run Hours
Despite a recorded peak demand (from the powerhouse Shark meter), of 97 kW for January, 2017, the
10 minute average is only 60 kW. This indicates that the peak loads are likely relatively short-lived,
such as motor starts. If this is the case, it should be reflected in the peak and average load seen after the
soft starts were installed in the school.
The monthly average kWh demand over the years 2012 to 2018 is shown in Table 3.
kWh
/ Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Mon
32,319
28,868
30,238
27,588
26,701
24,684
26,047
27,072
27,159
29,173
30,721
30,864
Table 3 – Monthly Energy Demand
The average demand over this 6 year period is 935 kWh per day, 28,435 kWh per month and 341.4
MWh per year. It has remained about the same (340 MWh) from year to year – there is no growth or
decline evident.
5. Generator Dispatch and Operating Costs
Figure 14above also shows the operating hours per day for each diesel generator. The three generators
at Igiugig have John Deere PowerTech 4.5 liter 4045TFM75 marine engines rated at 65 kW for prime
power use. They were installed in 2012.
Page | 13
Figure 15 – Diesel Generator Fuel Consumption
Figure 15 shows the diesel fuel consumption chart for these engines. Assuming a 90% alternator
efficiency, the engine brake power at 50 kW electric load is 55 kW with a corresponding fuel
consumption of 4 gallons, or 12.5 kWh/gallon. Powerhouse records agree with this estimate – the 6 year
average from 2012 – 2018 is 13 kWh/gallon.
The fuel consumption curve for these generators intercepts zero load at about 0.25 gallons per hour, or
6% of full load. Cooling fans and generator losses generally add 2-3 kW to the shaft load, or another
0.25 gallons, so the cost of operating a generator can be modeled as:
Cost per hour = Cost/Gallon x [ (kW load / 13 kWh/gal) + 0.5 gallons],
Or equivalently, $0.50/kWh plus $3.27 per generator hour. For 340 MWh and 8760 hours per year (one
generator), this comes to about $200K per year.
The actual cost of generating electricity includes capital depreciation, maintenance, supplies such as oil
and filters, and staff salary. The overall cost of generation is reflected in the price charged to residential
customers of $0.909/kWh. A typical cost for generator O&M is about $0.02/kWh or $1/hr for a 50 kW
generator. So the total operating cost of a generator is about $0.50/kWh plus $4/hour.
The present price of fuel delivered to the village is $6.55/gallon. This will of course fluctuate with
changing US oil prices. The delivered price includes a round trip for a DC3 aircraft from Anchorage
(250 miles each way).
Page | 14
Figure 16 – Historical US Oil Price (Federal Reserve Economic Data)
If US oil prices increase to 2011 – 2015 levels (Figure 16), the cost of oil at the village could exceed
$10/gallon.
The village has three 65 kW generators but only needs to run one at time for most of the year. In 2014
(the most complete powerhouse data year), the three generators ran for 38%, 23% and 54%,
respectively, of the 8760 hour year, with 84% of that time single generator and 16% two generators.
The need to run two generators is based on the need for “spinning reserve” – the ability to carry step
load increases such as the peak loads shown in Figure 12. Presently, when the village load exceeds 55
kW, a second generator is brought on-line. The system reverts to a single generator when the load drops
below 50 kW for 5 minutes. This is shown in Figure 17. As the daily average load increases, the need to
operate two generators increases.
Figure 17 – Generator Run Time vs Average Load
If US oil prices increase to 2011 –2015 levels (Figure 16),the cost of oil at the village could exceedp
$10/gallon.
Page | 15
A battery energy storage system (ESS) can be used to provide spinning reserve and so reduce generator
operating time. Replacing dual generator operation (which is about 1400 hours per year) with an ESS
could save $5,500 per year in fuel and maintenance costs.
6. Wind Turbine Options
There are unfortunately fewer options available now for village-scale wind turbines than just two years
ago. Companies such as Northern Wind Power (100 kW) and Endurance (50 and 85 kW) are no longer
in business. While options exist for 15, 20, 25 and 100 kW turbines, most are relatively new and have
had little operating experience in Alaska. The exception is Eocycle as they have fielded more than 50 of
their second generation 25 kW machines including some for arctic locations.
A key question to be answered is what are the tradeoffs between few large turbines and “many” smaller
ones. The laws of physics show that larger turbines are more cost effective than smaller ones as energy
production is proportional to rotor area (i.e., blade length-squared) whereas cost is roughly proportional
to blade length. This has to be balanced against the cost of foundations, the need for cranes, and the
increased resiliency offered by multiple turbines.
It should be noted that Igiugig already has six 2.5 kW Skystream turbines installed. Two of these have
run reliably since 2012. They generate 10 kWh per day or 3.6 MWh per year which is worth $3,300 at
the Igiugig cost per kWh. Intergrid repaired the remaining four in November, 2019 with some success –
we understand that five are now running. We learned from these repairs that the reliability problems of
Skystream turbines on the Igiugig grid are mainly due to power quality (see following section); rapid
increases in voltage can cause an uncontrolled inrush in the inverter which causes failure. This
realization has led to the addition of protection components to Skystream inverters for future repairs. It
has also led to the addition of similar components in Intergrid’s larger inverters.
Of the four turbines that were not operating, two had grid-induced inverter failures and two were
installation errors (wrong setpoints and a mi-seated plug).
The Skystream turbine is not considered for this wind study as:
1/ It cannot compete economically with larger turbines
2/ The company last manufacturing the Skystream (Xzeres) ceased operation in 2019.
The following turbines are considered (Table 4):
Name
Rated
kW
Rotor
Diam m Area m2
AEP
MWh
6ms
Windcad
MWh
Simulator
MWh Efficiency
Bergey15 15 9.6 290 45 36 44 0.33
QED20 20 12.5 491 60 47 55 0.28
EOCycle25 25 15.8 784 99 83 94 0.32
Northern100 100 24.4 1870 284 257 298 0.43
Xant100 100 24 1810 252 241 283 0.39
Table 4 – Turbine Properties
Here AEP column is the manufacturer-certified Annual Energy Production (MWh) with a 6 m/s annual
wind speed. The Windcad column is the energy production predicted using Igiugig wind data and the
“bin method”. The Simulator column is the prediction using Intergrid’s 10 minute time-series microgrid
A battery energy storage system (ESS) can be used to provide spinning reserve and so reduce generator ygy gy () p p g g
operating time. Replacing dual generator operation (which is about 1400 hours per year) with an ESSpg pgg p (
could save $5,500 per year in fuel and maintenance costs.
Of the four turbines that were not operating, two had grid-induced inverter failures and two werepg,g
installation errors (wrong setpoints and a mi-seated plug).
Page | 16
simulator and the Knight-Piesold wind data from 2012. The simulator results in particular agree well
with the manufacturer’s AEP numbers. The bin method tends to under-estimate production.
The last column –Efficiency – is the turbine efficiency calculated from:
ܲ =ଵ
ଶ ߩܣݒଷ ܥ
Where P = turbine power, ߩ = air density (~1.2 kg/m3), A=rotor area and Cp = efficiency.
The maximum theoretical limit on efficiency is 59.3% (the Betz limit) which is related to the need for
air to continue to move as it exits the turbine (and so the exit pressure cannot be zero). Practical wind
turbines have efficiencies of 30 - 40%. This number was calculated from the 6 m/s power value in the
turbine power curve to ensure that the published power curves were reasonable.
Bergey 15
The new Bergey Excel 15
kW turbine is “a
breakthrough in cost and
ease of installation”. It is
the product of five years
of research and
development, and lowers
the payback period for
small (10 – 15 kW) wind
turbines by 50%.
QED 20
The Phoenix 20 turbine is from QED (Tucson AZ). They have approx. 50 units in the field and are a
strong player in this power range. They were recently awarded an NREL grant for US turbine
certification.
Eocycle EO25
The EO25 is Canadian turbine rated 25 kW. This is a second-generation machine from Eocycle. They
have about 60 units currently in the field and are fully certified. They also have the advantage of
considerable arctic experience.
Figure 18 – Bergey 15
Figure 19 – QED 20
Page | 17
North Wind 100
The NW100C/28 turbine is provided as a
reference as Northern Power is no longer
in business in the USA. Knight Piesold
based their report on the smaller NP 100
Arctic turbine with 21m diameter blades.
They estimated production to be about 230
MWh/year. They stated the installed cost
of the NP100B/21 at $1.38 million which
makes it the least economic of the turbines
considered.
Xant M
This 100 kW
turbine is
manufactured in
Belgium. Two
units have been installed at Pilot Point, AK. The
installation cost $600K each, a significant
improvement over the NP-100. These are in early
production and so must be considered unproven at
this time. Also a turbine of this size requires
approximately 200 cubic yards of concrete for a
conventional foundation which is prohibitively
expensive at Igiugig.
Information sheets for each of these turbines are available at
https://drive.google.com/open?id=1mOcAOCDithUPBJtETF8T1rr1blLpiGLv
7. Wind Turbine Performance
There are three estimates of energy production given in Table 4 above:
- The “Windcad” value that uses the method of bins
- The manufacturer’s Annual Energy Production number at 6 m/s, and
- Intergrid’s 10 minute time series simulator (“Simulation”)
All produced similar results. The time series simulator also factors in ambient temperature which has a
significant impact in cold climates due to higher air density at low temperatures. It also allows the time
match between the wind resource and the village electrical load to be evaluated, so that a measure of
wind energy used directly vs. energy curtailed or diverted to heat loads.
The time series model is based on:
- 2014 load data reduced to 10 minute averages and scaled by 1.2.
- 2012 wind data (10 minute averages)
- Turbine power curves, interpolated and compensated for air density at ambient temperature.
Figure 21 – Eocycle 25
Figure 22 – Northern 100
Figure 20 – Xant 100
d scaled by 1.2.
Page | 18
8. Wind Turbine Economics
Because there are many details of an actual wind turbine installation that will be determined during
engineering design, this initial economic analysis is an estimate of total installed cost only.
Turbine kW
Simulation
MWh/yr
BaseCost
($000s)
Ship
ANC
Ship
IGI
Found-
ation
Install-
ation
Total
Cost MWh/$
Bergey15 15 44.1 55 4 3 1 7 70 0.63
QED20 20 55.4 102* 5 4 8 8 127 0.44
EOCycle25 25 94.2 87 13 5 10 10 125 0.75
Northern100 100 298 1,300 20 10 50 20 1,400 0.21
Xant100 100 283 600 20 10 50 20 700 0.40
Table 5 – Wind Turbine Economics
* QED have updated their costs – see Section 18 - Reduced costs of QED turbines.
The base cost is derived from discussions with manufacturers apart from the Northern 100 where the
Knight-Piesold value is used.
The shipping cost to Anchorage AK from the manufacturer is also an estimate. The cost for the Eocycle
25 includes the cost of a container which would be left on site for storage and housing the turbine
controls.
The cost of shipping to Igiugig from Anchorage is an estimate based on the ability to ship by barge from
Homer AK to the other side of Cook Inlet, then by road to Pedro Bay on Lake Iliamna, and then by
barge again to Igiugig.
The foundation cost is based on helical piles for the Bergey turbine, an ARE ballasted foundation for the
QED and Eocycle machines, and conventional concrete foundations for the Northern and Xant
machines. The foundation cost is discussed further in the geotechnical section.
The cost benefit of using wind turbines in Igiugig is primarily related to reductions in diesel fuel
consumption and diesel engine run time.
Turbine kW Simulation
MWh/yr Curtailed Fuel
Gallons Fuel
Reduction
Annual
Value
($000s)
Installed
Cost
($000s)
Simple
Payback
Years Base 407 31,325
Bergey15 15 44.1 0.1 27,934 3,391 $ 22.2 $ 70 3.1
QED20 20 55.4 0.2 27,074 4,251 $ 27.8 $ 127 4.6
EOCycle25 25 94.2 2.6 24,281 7,044 $ 46.1 $ 125 2.7
Northern100 100 298 137 18,978 12,347 $ 80.9 $ 1,400 17.3
Xant100 100 283 127 19,323 12,002 $ 78.6 $ 700 8.9
Table 6 – Fuel Saving and Payback for installing a Single Turbine
Table 6 shows the calculated fuel savings from installing a single turbine at the village. These values
were determined using the Intergrid time series simulator. The Fuel Gallons column is the projected
annual fuel usage. The base case, with no wind turbines, uses 31,325 gallons per year. The Fuel
Reduction column shows the annual fuel savings for a single turbine of each type. The Annual Value
column is the resulting savings at $6.55/gallon. This divided by the estimated installed cost yields an
Page | 19
estimate of simple payback time. The result is clear – while the economics for the Bergey turbine are
compelling, and a breakthrough for small wind turbines, the greater swept area of the Eocycle turbine
makes it the economic winner.
Using a total installed cost of $150K and an annual maintenance cost of $2,500, and a system lifetime of
20 years, the Levelized Cost of Electricity (LCOE 4) for the EO25 is 10.5 cents per kWh. While
administrative costs must be added, the cost to village consumers will be very much less than the present
rate of 90.9 cents per kWh.
The Curtailed column shows the amount of energy that would either be lost or diverted to heat loads for
a single turbine operating in parallel with a diesel generator with a minimum loading of 15 kW, and no
energy storage in the system. This shows that 100 kW rated turbines do not work well in this
configuration; about half the generated energy must either curtailed or diverted.
9. Wind Turbine Location
The considerations for wind turbine location are:
1/ A location of high wind speed. The best location from this viewpoint is on the lake shore, east of the
village (shown as “original proposal” in Figure 23). This area, however, is slated for a community
heritage center.
2/ Away from residences and occupied buildings. This is to protect from hazards such as ice thrown
from blades, and to minimize noise pollution.
3/ Near the three phase power distribution system. The Knight-Piesold study estimates that line
extension of the existing 12.5 kV underground distribution system would cost $300K/mile or about
$60/ft. This is high as the village has its own construction company and heavy equipment so the line
extension cost to the village would be for cable only – about $20/ft.
4/ Outside of aircraft approaches and below glide path, including that for the future float plane site.
5/ In a north-south line perpendicular to the prevailing east wind and spaced at least 2, but preferably 4
rotor diameters apart.
6/ Located on village or corporation-owned land, or on native allotment that can be purchased by the
village.
4 Derived using NREL LCOE calculator: https://www.nrel.gov/analysis/tech-lcoe.html
Page | 20
Figure 23 – Igiugig Village Map
The proposed location is to the southwest of the village and to the west of the man-camp (Figure 24). It
is about 5000 ft from the airport and 10,000 ft from the proposed float plane harbor. It is about 1500 ft
and downwind from the nearest residence.
The proposed location is to the southwest of the village and to the west of the man-camp
Figure 23 –Igiugig Village Map
(g )
It is about 1500 ft
p
and downwind from the nearest residence.
t 5000 ft from the airport and 10,000 ft from the proposed float plane harbor. ff
p g
p,
Page | 21
Figure 24 – Proposed Location for Wind Turbines
The location is a native allotment which the village plans to purchase. If this is not possible, wind
turbines could be located to the south-east, near the gravel pit on village-owned land. In either case,
three-phase power is available within 1000 ft as the 12 kV lines are buried beside the road.
10.Geotechnical Studies & Foundation Cost
Intergrid commissioned a geotechnical study for the location shown in Figure 24 with Northern
Geotechnical Engineering (NGE - Anchorage, AK). The University of Alaska, Anchorage also
commissioned a study for vertical wind turbine installations adjacent to the greenhouse about 1500 ft
east of the proposed site. Both reports are available in the Google Drive folder:
https://drive.google.com/open?id=1mOcAOCDithUPBJtETF8T1rr1blLpiGLv
The NGE report showed about 5 ft of sand over gravel and sand, and a silt layer at about 25 ft below
grade (TP2 was taken from the lake bed about 25 ft below grade from TP1). Figure 25 shows the
layering and transition from sand to sandy gravel in TP1. Figure 26 shows the collection of a silt sample
at TP2.
The report concluded that the allowable soil bearing capacity for a “mass concrete” foundation would be
in the order of 2,500 pounds per square foot. Lateral loading and steel pile foundation estimates are also
provided.
The 2014 report by Hattenburg, Dilley and Linnell had similar conclusions:
- A bearing capacity of 2,500 psf
- Little settlement of the underlying clay
- Similar lateral earth pressures.
If this is not possible,
,gpg
three-phase power is available within 1000 ft as the 12 kV lines are buried beside the road.
The location is a native allotment
g
The University of Alaska, Anchorage alsogg(g, ) y ,g
commissioned a study for vertical wind turbine installations adjacent to the greenhouse about 1500 ft y
east of the proposed site. B
Page | 22
Figure 25 – Test Pit at TP1 Figure 26 – Sampling at TP2
Turbine Tower Foundations
The Bergey 15 system uses helical piles and so is ideal for this soil type and location.
The QED turbine uses a 15 x 15 x 3 ft conventional concrete pad buried five feet in the ground for it’s
80 ft Ambor monopole tower (Figure 27). This foundation is sufficient for soils with a bearing capacity
of at least 1,500 psf.
Figure 27 – QED Pad Foundation
This foundation requires about 30 yards of concrete which in normal construction locations would cost
$120/yard or $4000. At Igiugig, however, the cost of bringing in concrete and rebar by barge, and the
lack of a local concrete plant puts the cost of concrete at $1000/yard 5, adding $30K to the project cost
per turbine. Igiugig does have a small portable mixer capable of making about 3 yards per batch.
The standard Eocycle foundation uses an innovative cross-shaped concrete pad that requires a similar
amount of concrete. In both cases, the cost and logistics involved in concrete pad foundations are
untenable.
Ten years ago, Southwest Windpower developed a ballasted foundation system for its Skystream
turbines. This concept has been expanded to larger wind turbines and telecom towers by ARE Telecom 6
in Minnesota and Ambor Structures
7 in China.
5 Quote from Karl Hill, Illiamna Construction Company, Igiugig: $1044/cu yard.
6 https://aretelecom.com/ballasted-foundations/
7 https://www.amborstructures.com/
Page | 23
Figure 28 – ARE AFS-1100 Foundation
Eocycle has been in communication with ARE in the past, but has not developed a suitable ballasted
foundation for the EO25 yet, as concrete foundations are less expensive in more accessible locations.
We propose a collaboration with Eocycle and ARE or Ambor to design and certify a ballasted
foundation for the EO25. We have estimated ballasted foundation costs at $10,000 per EO25 turbine in
Section 8,Wind Turbine Economics.
11.Multiple Turbines, MHK & Microgrid Control
Intergrid developed a time-series simulation model for examining the value of multiple wind turbines at
the village, and also how they would work in conjunction with MHK water turbines and battery energy
storage (ESS). The simulator uses 10 minute wind data from the 2012 Knight-Piersold study and the full
year of load and diesel loading information from the powerhouse SCADA system in 2014. The 2014
load was scaled by 20% to match present annual village load.
ygy
We have estimated ballasted foundation costs at $10,000 per EO25 turbine in
Section 8,Wind Turbine Economics.
y
The 2014ygp
load was scaled by 20% to match present annual village load.
Page | 24
Figure 29 – One Eocycle 25 Wind Turbine and Rivgen
Figure 29 shows the simulation output for one EO25 wind turbine and a Rivgen water turbine delivering
30 kW. The ESS is sized at 200 kW. The various system values are shown in the legend. What can be
seen here is that diesel operation (brown trace) is cut to less than 50%, with operation mainly in times of
low wind (green trace). Very little energy needs to be curtailed or diverted (yellow trace).
Figure 30 – Two Rivgens and no Wind Turbines
Page | 25
Figure 30 shows operation with two Rivgens and zero wind turbines. This configuration allows “diesels
off” operation for nearly the whole year with just a 100 kWh battery. Water turbines have the advantage
of much higher capacity factors and predictable output, but will have to prove reliability with river ice in
order to be viable for the village in the long term. ORPC has a Rivgen operating now in Igiugig and is
monitoring performance and ice in the river.
Considering just Windpower, the simulator was used to calculate fuel savings vs. multiple EO25
turbines and increasing amounts of energy storage (Figure 31).
Figure 31 – Fuel Saved vs # Turbines and ESS Size
This shows that there is little value in the ESS alone (the model does not presently represent reduction in
spinning reserve which could be considerable). The main benefit is achieved with a single turbine and a
small ESS – almost 4000 gallons per year. The value of 2 – 5 wind turbines diminishes as the turbine
count is increased and more energy is curtailed or diverted. The value of storage however increases with
more turbines as the extra storage allows more wind power to be fed to loads.
Figure 32 – Cost of EO25 + Storage Systems
Figure 32 shows the cost of wind energy systems with varying amounts of turbines and storage. It is
based on $150K per turbine and $20K per 100kWh of storage.
Page | 26
Figure 33 – Contour Plot of Gallons/Year Saved Per Dollar Invested
Figure 33 shows the result of combining the two previous graphs. The X-axis represents ESS size
ranging from 0 – 1.1 MWh. The Y-axis shows the number of EO25 wind turbines. The most cost-
effective benefit is for one EO25 turbine and 100 kWh of storage. Two EO25s and 300 - 500 kW of
storage have about the same benefit per dollar invested.
12.Energy Storage System
In early 2019, Intergrid entered into a program with Bergey Windpower and the Nissan USA motor
company to develop electronics for use with “second life” Nissan Leaf electric vehicle (EV) batteries.
This is a potential breakthrough for stationary energy storage applications as second life batteries are
available now at a cost of $100/kWh in low quantities and likely $50/kWh as the used battery market
grows. This can be contrasted with the cost of CALB cells (previously proposed for Igiugig) which is
$350/kWh. In addition, the Nissan Leaf batteries come complete with BMS, internal fusing and
contactors, and a fireproof steel enclosure. The added value of these components is $100-$200/kWh, so
the comparison is a six-to-one ratio of cost. There are presently 140,000 Nissan Leaf vehicles on USA
roads, and so the availability of used battery packs is assured. In time, the batteries from other
manufacturers will become available, but the Leaf pack is the best choice at present as it does not use
liquid cooling, and has a very simple electrical and control interface.
The Nissan battery has an operating voltage range of 336 to 393 Vdc. The dc bus voltage for inverter
needs to be 400 – 700 Vdc and so a dc-dc converter is required to boost the battery voltage and control
power flow. Intergrid has designed a converter for this purpose.
The Nissan battery has 182 cells and a nominal energy rating is 65 x 370V = 24 kWh. Second life
batteries retain 70-80% of this capacity, or about 16 kWh per battery module. The Nissan cells have
lithium-manganese-oxide with nickel oxide (LiMn2O4 with LiNiO2) cathodes which is inherently safer
than lithium-cobalt-oxide cathode materials used in many electric vehicle batteries. The battery is
packaged in a steel case that is designed to mount under the car (Figure 29 ()). In stationary applications,
it provides excellent mechanical and fire protection.
. The most cost-gg
effective benefit is for one EO25 turbine and 100 kWh of storage. Two EO25s and 300 -500 kW of
storage have about the same benefit per dollar invested.
q
Intergrid has designed a converter for this purpose.
Page | 27
Figure 34 – Nissan Leaf Battery Pack - Outside
Figure 35 – Nissan Leaf Battery Pack - Inside
The orange connector is for the power conductors. The black one is for control.
The white connector on top is a segmentation disconnect that divides the battery into
two strings.
Inside (Figure 30), the battery has 48 battery modules, a Battery Management System
(BMS) and main and pre-charge contactors. The batteries can be mounted in any
orientation but “upside down” (segment disconnect down). For the Iguigig project,
we will install “on end” with a dolly adapter rail and wheels that allows the battery to
be rolled into place inside a 20 ft container (Figure 31).
This storage project is part of ORPC’s EETF work with AEA and Intergrid. We are
finishing the logistics for shipping six batteries (100 kWh) to Igiugig now and beginning testing of the
dc-dc converter at a single battery system in our lab in New Hampshire in January. We expect to install
and test the 100 kW ESS in Igiugig in the spring of 2020.
13.Dispatchable Thermal Loads
As the amount of renewable energy increases in the village, the need to either curtail renewable
production, or to divert it for other uses increases. This is only significant if more than 2 EO25s are
installed or a Rivgen is operating. A well-controlled system with two EO25s and 200 kWh of storage
would provide half of the village electric demand from wind and require essentially no curtailement or
diversion.
With 4 turbines, 70% of the village load is carried by wind and diesel run time is about 4000 hours per
year (50% of the 8760 annual value), but 117 MWh of energy needs to be curtailed (Case 1).
With a Rivgen MHK turbine operating at 30 kW, two EO25s and 400 kWh of storage, diesel fuel
consumption and generator run time is cut by 70% and 44 MWh needs to be curtailed (Case 2).
The heating value of diesel fuel is about 38 kWh per gallon
8 for a 100% efficient, non-condensing
furnace, so every diverted MWh of electricity can displace about 25 gallons of heating fuel. If ground-
or air-source heat pumps are used, this value increases to 75 – 100 gallons per MWh.
8 https://en.wikipedia.org/wiki/Gasoline_gallon_equivalent
Figure 36 - Dolly
yg
A well-controlled system with two EO25s and 200 kWh of storage gpg y g
would provide half of the village electric demand from wind and require essentially no curtailement or p
diversion.
Page | 28
The main heating fuel use in the village is for the school (5000 gallons/year) and the airport building
(1300 gallons/year) but the latter is offset by the use of a waste-oil furnace. The obvious initial target for
a thermal diversion load is the school building. The full amount of excess power generation could be
used with an electric boiler in both cases. A heat pump and a hot water storage tank would greatly
increase the value of such a system. As the school has hydronic heating, installation of a boiler or heat
pump would be relatively straightforward. The use of heatpumps vs. boilers needs to be balanced against
reliability and maintenance costs, but 5000 gallons of heating oil costs $30K at todays prices. This can
cover an adequate maintenance budget.
Other locations where diverted electricity could be used for heating are:
- The diesel power house, where engines need to be kept at operating temperature
- The adjoining buildings to the powerhouse that are heated from the heat recovery loop
- The water plant storage tank
- The airport building
- The village greenhouse which has in-slab hydronic heating.
In all cases, air to water or ground to water heat pumps would provide greater value that electric boilers.
In the Northeastern USA, air-source heatpumps are proving reliable and cost-effective in residential and
small commercial applications, event at temperatures as low as -20C. In Canada, Maritime Geothermal
has introduced a line of air to water heat pumps that could work well in Igiugig 9.
Backup or diversion heat for the power house, heat recovery loop and water tank must be provided in
any scenario that results in extended periods of diesels-off operation. This will have to be a part of the
design for any significant renewable installation at the village.
14.Power Quality Issues
Power quality issues have been observed in the village since the installation of test vertical axis wind
turbines in 2016. These are seen as periodic dips in grid voltage and frequency.
15.Siting and Permitting Issues
The FAA requires a “determination of no hazard” for a turbine greater than 200 ft hight, or at a location
“within 10,000 feet of an airport or seaplane base that does not have a runway more than 3,200 feet in
length and the object would exceed a 50:1 horizontal slope (50 feet horizontally for each 1 foot
vertically) from the nearest point of the nearest runway”10. As the proposed wind turbine location is
about 5000 ft from the airport and 10,000 ft from the proposed float plane base, and the maximum
height of the obstruction 125 ft (100 ft tower plus 25 ft blade length), an FAA determination will be
required. As a 125 ft obstruction falls under this criteria, the FAA may require marking or obstruction
lighting on the tower.
There are other permitting issues that should be investigated during the wind turbine project design
phase. Some of these are listed in the 2012 Knight-Piesold report. The most important is a review by
Alaska fish and game for bird habitat near the turbine site. This is unlikely to be a problem as the
proposed site is sparsely vegetated with low trees and probably has little or no bird nesting population.
9 https://www.nordicghp.com/2016/09/high-temperature-hot-water-heat-pump/
10 FAA AC No: 70/7460.2K
Qy
Power quality issues have been observed in the village since the installation of test vertical axis wind qy g
turbines in 2016. These are seen as periodic dips in grid voltage and frequency.
Page | 29
16.Eocycles Installed Fleet and Commercial Outlook
We contacted Paul Dawson, VP of Eocycle, and requested information on the current operating fleet of
their second generation machines. We also asked for clarification of the status of the first generation
machine at Kotzebue, AK.
Paul supplied the attached document “Eocycle project history Jan 2020.pdf” which states that there are 8
operational turbines at this time all EO25-Gen 2.
He further stated that the Gen 2 units have accumulated 10 equivalent run years of operation.
We plan to interview some of the owners of operating machines as references for an installation in
Igiugig.
Regarding the generation 1 machine at Kotzebue, there has been concern for several years that the
turbine has not been operating. Paul assured me that the machine is ready to run once the customer
(Kotzebue Electric Association - KEA) replaces a part that was supplied “long ago”. He also provided
contact information for Matt Bergen who is the project engineer there.
A conversation with Matt led to an understanding that the main problems with this “Generation 1” EO25
are hydraulic leaks, complex controls (two PLCs) and two rotor to generator couplings that have failed.
Kotzebue electric coop would like Eocycle to install the new provided parts and re-commission, but it
appears funding and time for this have not been available. This does not necessarily impact a viability
review for the “Generation 2” wind turbines, as many manufacturers have trouble in supporting legacy
machines.
More concerning has been learning of emergency brake failures on two Gen-2 machines located in
upstate New York. This came out at the Distributed Wind Energy Association meeting in Washington
DC in early March, 2020. We learned from the turbine installer that the emergency disk brakes on two
machines failed to operate resulting in turbine overspeed conditions, but no thrown blades (which is a
good sign that the fundamental design is sound). The machines were brought under control by manually
steering the turbine yaw perpendicular to the wind. The installers expressed concerns that:
- There was no sensor to detect worn disk brake pads, or a startup sequence that proved the
mechanical braking capability
- The turbine controller appeared to be responding improperly to minor failures (grid voltage
imbalance) by going into mechanical brake mode
- Controller interfaces relied on internet connections rather than having local control panels
available.
- Sharing hydraulic rams for tower lowering is time consuming and somewhat dangerous.
Recommend that each tower have dedicated rams.
I spoke to Paul Dawson of Eocycle who was at the conference and he said that the root cause of the
problem was an out of specification disk material from an alternate supplier. They are presently
retrofitting all operating turbines with brake rotors from the original supplier. Paul was with Northern
Windpower in the past and brings a lot of experience now to Eocycle.
I also spoke with Trevor Atkinson, a past Northern employee who also now works for Eocycle. He is a
competent engineering who has been in the wind business for many years and is the main reason now
that I believe that Eocycle will overcome this current set of field problems and produce a reliable
machine.
Page | 30
Still, we need to move ahead with caution as field service at Alaskan villages is orders of magnitude
more difficult and costly that working with turbines in New York.
17.Ballasted Foundation for the EO25
As the cost of concrete at Igiugig is very high (~$1000/yard) we have recommended using ballasted
foundations instead of Eocycle’s innovative cross-shaped concrete foundation.
The concrete foundation needs about 30 yards of concrete per turbine.
We contacted Dion Johnson of ARE Telecom (St Paul MN) who are the main supplier of ballasted
towers for distributed wind and telecom worldwide. (There is a competing Chinese company – Ambor –
who could supply similar product, but we prefer a US company. In addition, ARE Telecom has already
supplied ballasted tower bases to Igiugig for vertical wind turbine experimental systems.
Dion gave a verbal quote as follows:
- ARE AFS400 ballasted foundation and 80 ft tower $30,000
- Freight from Chinese manufacturing facility to Anchorage $5000
- 40 ft container$2200
The previous quotation from Eocycle included an 80 ft tower, which is an add-cost of $14,500.
The cost of the ballasted foundation then is about $15,500 plus shipping.
The ballasted foundation then has a net cost advantage over a concrete foundation of about $15,000 or
50%. In addition, on-site labor is greatly reduced as gravel is abundant, while concrete making using
Igiugig’s 6 yard mixer is labor and time intensive.
18.Reduced costs of QED turbines
QED is presently undergoing a design review with NREL and entering into a certification agreement
with Windward Engineering (Spanish Fork, UT) and the Small Wind Certification Council. This is good
news as a lack of certification is a reason to keep a turbine manufacturer from consideration for Alaskan
villages. The NREL review questioned the need for a safety clutch on the yaw drive. It’s removal
resulted in a $6,000 saving in the base turbine cost. That increased generation to 46.2 MWh/$ and
reduced the simple payback time from 4.6 to 4.4 years. They still do not match the performance of the
Bergey and Eocycle (3.1 and 2.7 years respectively).
A discussion with QED regarding their slightly lower aerodynamic efficiency that the Eocycle and
Bergey revealed that to some extent, this is an intentional decision to trade low acoustic noise for lower
energy performance. It will be interesting to see how the acoustic testing results for the turbine bear that
out.
19.Recommendations
In response to the draft version of this report, we have answered most of the questions posed in the AEA
Wind Program Guidelines for Conceptual Design Reports and added this as Appendix A, following.
The village of Igiugig can benefit significantly from the installation of wind turbines, primarily due to
the high landed cost of diesel fuel now at the village ($6/gallon), but also due to the impact that an oil
price spike could have on village finances.
pp y p , p p y ,
supplied ballasted tower bases to Igiugig for vertical wind turbine experimental systems.
Page | 31
Our recommendations at this time are to:
1/ Plan on an initial installation of two Eocycle 25 wind turbines. The shipping and installation costs for
two turbines are significantly lower than for one. For example, two turbines can be commissioned in one
trip, and a single container can be used for housing the turbine electronics and controls. The cost of
installing two turbines will be about $350K including engineering design and permitting.
2/ Use the time needed to write grants and otherwise raise funds for wind turbine installations to
establish the reliability of this chosen turbine product. Keep monitoring field progress of Bergey and
QED machines as backup possibilities. Due to present Eocycle field problems, turbine should not be
purchased until 6-12 months of reliable operation is demonstrated at multiple sites.
2/ Complete the purchase by the village of the proposed wind site.
3/ Perform a wildlife survey with Alaska fish and game.
4/ Work with Eocycle and ARE or Ambor to design ballasted foundations for the Eocycle towers.
5/ Work with an Alaskan engineering company such as Gray Staysel (who have done much of the
electrical planning for the village distribution system and ORPC’s MHK generator) to develop final
plans for the wind turbine installation.
5/ Explore funding and grant opportunities for the system.
6/ Continue to work with Intergrid on diesel control upgrades for the powerhouse and incorporating
energy storage into the system.
7/ Collaborate with the Alaska Center for Energy and Power on microgrid control systems.
Best and Worst Case Scenarios
The best case scenario for the village is that all of these efforts will come together to create a working
microgrid for the village, and a major reduction in diesel fuel use.
This would include:
- 80 kW / 100 kWh of energy storage at the power house
- One or two operating 30-40 kW Rivgen hydrokinetic turbines
- Two Eocycle 25 wind turbines on the ridge behind the man-camp
- Acquisition of the land for the turbine site
- 80 kW / 100 kWh of energy storage at the wind site
- The introduction of electric heat at the school, town office/hangar, health center and
greenhouses. Ideally this will be heat pump rather that resistance heat, but both have a place
considering reliability, peak wind season vs. building load requirements, and cost of
maintenance.
- Addition of heat pumps or boilers to supply heat to the powerhouse, water treatment facility,
storage tank and buildings now supplied by heat recovery from the diesels. This to support
diesels-off operation.
- Implementation of viable local control based on local grid voltage and frequency, plus
communications from a master controller and SCADA system via WiFi and long-range radio.
- Implementation of a master controller that can make all the components work together.
In total, this is a big list that will take a lot of engineering and design effort to implement.
The good news is that the items above are not interdependent and can be implemented individually as an
organic growth path. For instance, energy storage that will be demonstrated under an EETF grant can be
installed without a wind system and provide benefit in the form of spinning reserve and time-shifting of
Rivgen energy. A wind turbine installation can offset diesel consumption without energy storage and
still have financial benefits.
Addition of heat pumps or boilers to supply heat to the powerhouse, water treatment facility,p p pp y p ,y
storage tank and buildings now supplied by heat recovery from the diesels. This to supportg
diesels-off operation.
Page | 32
The worst case scenario is that turbines are installed and the manufacturer goes out of business. This is
certainly a concern as most major players in the US and Canada have failed in the past few years:
Northern (100 kW), Xzeres (10 & 2.5 kW), Xant (100 kW), Endurance (50 kW), AOC (50 kW)…
The counter argument is that a simple payback of less than 3 years means that the turbine company (or
its insurance) would only have to survive for 3 years for break-even. It’s worth the bet, but still with
appropriate caution.
Page | 33
Appendix A - Response to Alaska Wind Program Guidelines for Conceptual
Design Reports
These are the bare essential aspects that should be addressed when developing wind feasibility studies.
Wind turbines are not a stand-alone component, but rather an energy source that must be integrated into an
over-arching power generation and distribution system for the community. Conversely, a CDR with an overly
broad scope wastes time and money and can make it more difficult to recommend next steps.
These are a guideline rather than a mandate. The situation at a particular village may require more
emphasis on some aspects and less on others. Also, the state of the US wind turbine industry means
that a simple choice of a particular turbine is clear, leading to all necessary engineering.
A1 - Wind Resource Study
Most of these questions are answered in the original Knight-Piesold (K-P) wind study performed in
2012, and the additional analysis provided in the Intergrid report.
¾How reliable is the overall data? Are there gaps? Did any sensors or datalogger fail? Was a
log sheet filled out during tower erection?
TheK-P report shows no gaps, sensor failures or datalogger failures for the wind speed and direction
measurements.
¾How fast is the wind? Average speed, maximum, std. dev.?
The K-P report (Page A-1) shows an average wind speed of 5.8 m/s at 33 meters height, and a max
wind speed of 27.6 m/s (62 mph). The site then is Wind Power Class 2
(https://en.wikipedia.org/wiki/Wind_profile_power_law).
¾How does the wind speed vary throughout the day? Month tomonth?
This is covered adequately in the Intergrid report
¾What does the wind speed distribution look like? Weibull K? Is it bi-modal with periods of
calm then severe storms? Is the distribution more continuous?
The distribution is continuous and not bimodal. The Weibull K value is 1.72 for the 33m anemometer
(K-P report Page A-2).
This indicates that the distribution is skewed more to lower wind speeds than a K=2 (Raleigh
distribution) or higher K values which approach normal distributions.
This graph shows how Weibull K changes the distribution.
Source: https://www.homerenergy.com/products/pro/docs/latest/weibull_k_value.html
Page | 34
¾How does the wind shear change with elevation (power law exponent)? How turbulent is
the wind? What are the predicted maximum speeds over 20 and 50 years?
The shear exponent was calculated as 0.237 (K-P report page A-1). This is typical for level terrain with
low bushes and trees. This value can be used to estimate the value of higher towers. For instance, a
100 ft hub height could be expected to have (100/80)^0.237 = 5% higher wind speed than an 80 ft
tower resulting in 15% more annual energy capture due to the cubic law of wind energy to wind
speed.
Predicting maximum wind speed over 20 and 50 years is a complex and largely academic matter.
A reasonable estimate can be made by comparison to other Alaskan sites – e.g. the “mile 217” site in
the report at https://www.arlis.org/docs/vol1/Susitna/SUS/0/SUS48.pdf
With a measured maximum gust of 79 mph yielded estimated extreme wind speeds of 130 mph (25
year) and 137 mph (50 year), a ratio of 165% and 173% respectively.
This results in a 25 year estimate of 45 m/s (102 mph) and a 50 year estimate of 49 m/s (108 mph).
The statistical approaches for estimating extreme wind rely on past maximum wind events and do not
factor in major climate change transitions that appear to be impacting world weather now, so any
estimate is just that – we cannot fully predict the weather.
¾How much icing is experienced at the site? How thick is the icing and how long does itlast?
There was no mention of icing in the 2012 Knight Piesold wind study. We re-analyzed the raw data for
the three tower anemometers with the following criteria:
Sum of 3 anemometers > 3 m/s
Any anemometer < 0.5 m/s
Any other anemometer > 2 m/s
Because of calibration offset, the minimum recorded wind speed shows as 0.4 m/s in all channels.
57,000 records were analyzed showing just one possible significant icing event:
12/13/2011 11:10 – 13:00
Ch3 anemometer was at 1.9 m/s while Ch1 & Ch2 were at 0.4 m/s. At 13:00, Ch3 recorded a max wind
speed of 5.3 m/s with Ch1 & Ch2 were at 0.4 m/s.
After that time, all channels were tracking
Data from the FAA weather station at Igiugig does not appear to have icing information (or
precipitation measurements) although it does have wet and dry bulb temperatures which allow
humidity to be calculated.
We asked for local knowledge from villagers to obtain a better understanding of icing frequency.
We were told that icing events, though frequent (1-2 per month), seldom lasted longer than a day, and
temperatures were sufficiently high for icing on structures to dissipate rapidly.
Anecdotally, there has been less snow cover in the village in recent years, to the extent that there is
little opportunity for cross-country skiing and dog sledding. This indicates that while icing events can
occur, they will not result in sustained periods of turbine down time.
Eocycle includes vibration sensors in their nacelle that detect imbalance and shut down the turbine if
ice buildup occurs on the blades. Even if the ice clears, a manual reset/restart is required, although this
can be done remotely.
Page | 35
¾What is the air temperature and density?
The annual mean air temperature in 2012 was -0.7 °C (-18 °F). The FAA station at the Igiugig airport is
“PAIG”.
The 2012 K-P data set has the following temperature statistics.
Temperature C Jan Feb Mar AprMay Jun Jul Aug Sep Oct Nov Dec Year
Max 2.4 5.6 6.3 14.3 17.8 24.2 22.7 20.4 15.8 11.2 5.8 6.9 24.2
Min -35.6 -30.8 -29.6 -15.1 -7.5 1.3 -12.2 1.4 -7.1 -10.3 -28.3 -33.9 -35.6
Average -21.9 -4.5 -11.9 1.2 4.6 9.7 10.8 11.3 7.4 2.8 -9.8 -7.9 -0.7
The time series model includes a temperature compensation for air density. This is relative to the
manufacturer’s power curve which uses a reference air density of 1.27 kg/m3.
¾How consistent is the wind data from one year to the next? How does it compare with
long- term trends?
This is covered in the body of the Intergrid report.
¾How was the met tower site chosen? Are there nearby obstructions?
This is covered in the K-P report.
¾How does the wind speed and wind rose compare with the statewide wind resource model
for that location?
The 2012 measured average wind speed of 5.7 m/w agrees with the AK wind map values of 5.5-6 m/s.
See https://windexchange.energy.gov/files/u/visualization/pdf/ak_80m.pdf
¾How closely will wind turbines be placed near the met towersite?
Within 1 km. The proposed site is across the road from the met tower site. The proposed site is slightly
higher and so should yield wind energy slightly greater than the met tower site.
¾How does the wind rose affect siting for multiple turbines?
The prevailing wind is from the east. The proposed wind turbine siting is in a north-south row,
perpendicular to the prevailing wind, thus avoiding wake effects on downwind turbines.
¾What issues were raised by the FAA and US Fish & Wildlife Service during the met
tower permitting process?
None.
¾What is the estimated net production for turbines being considered, assuming no
wasted/excess power? Windographer defaults to an 82% availability. This is a reasonable
estimate.
The time series model calculates wasted or excess power. The amount of curtailment or wasted power
is a function of available energy storage.
A2 - Existing Electrical System Overview
¾How does the community electrical load vary throughout the day? Month to month? What
is the average, peak and minimum?
This is covered in the main report.
¾Are there seasonal loads due to commercial or traditional activities? How do the residential
electrical loads compare with industrial and commercial loads throughout the day and month
to month? What are the station service loads?
Page | 36
There are no seasonal loads due to commercial or traditional activities. The largest non-residential
loads are the school and water pumping/purification system. The station service loads are minimal
(radiator fans and circulating pumps).
¾Are there existing diversion electrical loads in the community?
No.
¾Are there electrical loads that could be converted to dispatchable loads if needed?
None at present apart from school ventilation.
¾What is the make, model, kW rating and age of each diesel genset? What are the fuel curves
for each unit? What type of mechanical or electronic throttle controls exist? What are the
actual reported kWhrs per gallon of fuel for this facility?
This is covered in the Intergrid report.
¾What kind of switch gear exists – make, model, manual/automatic? What kind of
SCADA currently exists?
AEA standard switchboard installed in 2012. Fabricated by Controlled Power (Bothel, WA). Automatic
diesel control using Woodward GCP-30 controllers (which are now obsolete).
¾Are upgrades or replacements planned for any key system components?
Igiugig is “on the list” for a diesel control upgrade to Woodward easYgen in 2021.
¾Is there a heat recovery system? What loads does it feed? How are those heat loads
monitored/quantified? How much heat is lost in the system?
Heat recover is used to heat the village water tank, water processing building and old village office
(washeteria, ILC office, post office etc).
¾Are there additional potential electrical loads in the community that are not currently
being met? Are any new electrical loads being planned?
There is some consideration for providing electric (heat pump) heating for greenhouses, residences,
the village offices in the hangar and the school building. A large cultural center building is planned for
a location on the lake, near the present barge landing. This could easily contribute another 20 kW of
load. We will work with design architects towards as much energy efficiency and long-term
electrification of heat loads as possible.
¾Where are the major electrical loads located in the community from a geospatialperspective?
Mainly in the old village close to the powerhouse.
¾How well are the phases balanced in the distribution system? How are the transformers in
the community loaded or overloaded? Where is there phase or transformer capacity to add
additional loads?
The phases are well balanced. The main transformer is 112.5 kVA. It has adequate capacity for growth
(the present power peak is about 80 kW) and the addition of a wind + storage power system. The
addition of energy storage at the wind site (80 kW planned) would increase the capacity of the
distribution system as it would connect at the far end near the landfill.
¾What is the condition of the distribution lines, transformers and poles?
The distribution system was installed in stages from ~2001 to 2012. It is an underground 12.5 kV three
phase system. There are no poles.
¾Provide a map showing single versus three-phase power lines and varying voltagelevels?
Page | 37
The village power system consists of three diesel generators rated at 65 kW each feeding a 12.47 kV
(medium voltage – MV) underground distribution system.
Figure 37 – 65kV John Deere Generator
A one-line diagram of the powerhouse is shown in Figure 1.
Figure 38 – One Line Diagram of Powerhouse
The distribution system for the main village (by the river) is shown in Figure 2.
Page | 38
Figure 39 – Village Distribution System (1999 Upgrade)
Key elements in the distribution system are:
-A 112.5 kVA delta-Y step up transformer to 12.47 kV three phase power. It has no load losses of
620W and full load loss of 1970W. It has an impedance of 3.5%, a resistance of 1.21% and a
series inductive reactance of 3.29%. The X/R ratio is 2.73.
Figure 40 –112 kVA Transformer
-Three series line reactors rated 2A, 9100 mHȍZLWKDtest power loss of 94W each at 2A
rated current.
Page | 39
Figure 41 – Line Reactors
-For the main village area, the medium voltage (MV) distribution system is underground
Prysmian (now General Cable)301437A “DoubleSeal” 133% insulation level, 15 kV Ethylene
Propylene Rubber (EPR)-insulated cable with a #2AWG aluminum core and concentric neutral.
The insulation has a dielectric constant of 2.45. The coQGXFWRUUHVLVWDQFHLVȍNPDW&
and the cable capacitance is 0.157 uF/km. The three-phase positive sequence impedance of the
FDEOHLV ȍNPThe zero-sequence impedance (used for fault analysis) is approx. 2.5ȍ/km.
The propagation velocity is 147.14 m/us 11.
-
Figure 42 – 2 AWG 15 kV Distribution Cable
-Step down transformers such as the ones shown in Figure 7 & Figure 8.
11 General Cable 297388 data sheet.
Page | 40
Figure 43 – 15 kVA Single-Phase Distribution Transformer
These are typically Cooper/Eaton Catalog WB2A43072Y3.
Figure 44 – Three-Phase Distribution Transformer
-Three parallel reactors rated 10 kVA at the school crossroads.
In 2001, the MV distribution system was extended to the west, away from the river, from the school to
the new dump area. This consisted mainly of a three conductor extension using 1/0 AWG Prysmian
cable.
Page | 41
¾What are the parasitic and other system losses?
The main parasitic and other system losses are attributed to:
-Transformer losses
-Series inductor and shunt reactor losses.
-Distribution system resistive losses
-Power house radiator fans
The calculated line loss for the power system is high – ranging from 14 to 20% in 2018. It may be that
the long feeder to the man-camp, subdivision and dump is a major contributor. There is a 112.5 kVA
transformer at the dump that was installed for a planned Northern Power wind turbine in 2001 that
presently runs a street light, but it’s no-load loss is less than 100W, so while contributing, it is not the
cause of the line loss.
A3 - Heat Loads Overview
¾What is the heat recovery percentage of each diesel genset? What heat loads are tied into the
Page | 42
heat recovery system? How are those heat loads monitored/quantified? How much heat is lost
in the system? What additional capacity is available?
The heat-recovery load is not presently monitored.
¾What is the daily and month-to-month profile of each heatload?
¾Where are the major heat loads located in the community from a geospatialperspective?
The present heat recovery system feed only buildings in the old village area (water treatment plant,
washeteria, ILC office, post office and store). It may be practical to extend heat recovery to the hangar
and new village office building, but the future scenario is for less diesel time and available heat, and so
conversion of all buildings to heat-pump electric heat is a desirable goal.
¾Are there additional potential heat loads in the community that are not currently being
met? Are any new heat loads being planned? Where are they located relative to the
powerhouse?
Heat-pump heating for the school, hangar office space, health center and greenhouse is possible.
None of these is a candidate for heat recovery from the powerhouse due to distance.
¾What is the efficiency of current boilers? Where is space available to add electricboilers?
There is space available to add electric boilers in the powerhouse or in the water treatment plant.
There is a backup boiler in the water treatment plant. It could be used to supply heat during diesels-off
operation.
A4 - Compiling the Final Conceptual Design Report
In addition to answering all of the above questions, please provide the following materials in your
report.
¾Proposed electrical system line drawings showing turbines, transmission lines,
distribution system and powerhouse. Label voltage and phase of lines, plus conductor
type, size and resistance factor at 0 deg Celsius.
As location and turbine type are still not completely determined, this level of detail is not justified at
this time. It is sufficient to say that the existing three phase distribution system extends to within a few
hundred meters of planned turbine sites, and as small (25 kW or less) turbines are planned, their
individual impact is small on the whole system. In addition, the energy storage component of the
village microgrid would be located at least in part at the wind turbine site to minimize voltage
variation and flicker. The long term plan for energy storage for the village is to install an 80 kW energy
storage system (ESS) at the power house and another at the wind turbine site. This provides
redundancy and improves overall power quality as there will be “stiff” power sources at both ends of
the distribution system.
¾How will turbine type, quantity and location affect power quality issues such as reactive
power, power factor, voltage rise and other distributed generation issues? Does a basic
voltage drop/rise calculation indicate the need for additional analysis using the DG Toolbox
or running a load flow analysis? Is complex PSSE modeling required?
The days of direct-connected induction generators for wind turbines are over. The benefits of variable
RPM control and de-coupling of the rotor inertia from the power grid are so significant that no US
manufacturer is considering direct generator connection. Inverters for wind turbines and ESS generally
operate at close to unity power factor (zero reactive power), however reactive power control is
possible.
Page | 43
It would be normal on a village microgrid to enable power/frequency and reactive-power/voltage
control to maintain frequency control and voltage support.
The 1/0 Prysmian cable that extends the distribution system to the proposed wind turbine area is
approximately 2 miles long (3.3 km) and has a conductor ĂĐƌĞƐŝƐƚĂŶĐĞŽĨĂďŽƵƚϬ͘ϱɏͬkm. 50 kW of
wind generation on a 7200V (line to line) system creates a phase current of 4 Amps, and a line-neutral
voltage rise of 0.16%. This also the per-unit (p.u.) ratio of 50 kW impedance to the line impedance.
The other significant series impedance is the powerhouse step-up transformer. It has a rated
impedance of about 3% p.u., but is rated 112.5 kVA, and thus 50 kW of generation would result in a
voltage rise of 1.3% p.u. which is acceptable apart from flicker considerations. The new IEEE1547
standard for grid connection of renewable energy systems to the power grid requires that flicker be
managed to below the “Borderline of Irritation” in IEEE519-1992, Figure 10.3:
This can be interpreted for wind systems with a typical spectrum of:
https://www.hindawi.com/journals/jwe/2013/739162/fig10/
The majority of the spectral energy is in the 0.01 to 0.2 Hz range, and higher frequency variations will
be absorbed by rotor inertia, so the flicker limit will be about 1% p.u. A turbine without local ESS may
operate without causing flicker at nearby locations, but a small amount of short-term energy storage
would completely eliminate problems. Of note is that the IEEE standard was for the era of
Page | 44
incandescent lamps. While some LED lamps now have excellent current regulation and are immune to
line voltage variations, lower quality LED ballasts are particularly susceptible. So compliance with
these IEEE standards may not be sufficient and responsive local ESS systems are likely required.
¾Detailed line drawing showing how wind power connects to the powerhouse
through switchgear and how wind, diesel and diversion loads integrate with each
other.
At this stage, detailed drawings are not justified as the turbine type and location are not yet fixed.
Conceptually, the Igiugig microgrid power system could consist of:
x The existing diesels and distribution system at the powerhouse
x 80 kW / 100 kWh of ESS at the powerhouse for support in the old village
x One or two 30 kW Rivgen hydrokinetic turbines at the fish landing half way between the
powerhouse and dump.
x Two 25 kW wind turbines north of the man-camp, close to the dump location.
x An additional 80 kW / 100 kWh of ESS at the wind site for support in the subdivision/man-
camp area
x Relocation of the existing 112.5 kVA 7200/208V transformer at the landfill to the wind turbine
site.
¾Proposed and existing SCADA system drawing and description.
The existing SCADA system is a 2012-vintage AEA standard system using a Windows 2008 server in the
powerhouse switchboard. The proposed wind system will be a modification of the proposed 2021
upgrade (Woodward easYgen) to the powerhouse diesel controls. Intergrid has been working with
Woodward and the Alaska Center for Energy and Power (ACEP - UA-Fairbanks) to design such a
system. Diesel, ESS and wind dispatch will be handled by a first level of autonomous control, with a PC-
based system optimizer communication with the diesel plant, two ESS systems, Rivgens and wind
turbines. The main communications channel to the various sites will be WiFi using long-range
antennas etc (e.g. Ubiquiti systems). There will also be a low-data rate backup system using LoRa
radios. These operate at 800 MHz and have line of sight range of up to 50 miles.
¾Proposed physical layout at turbine site, powerhouse and transmission route.
At this stage, detailed drawings are not justified as the turbine type and location are not yet fixed.
¾Proposed and existing diversion load drawing and description.
Each wind turbine (e.g. EO25) will have a diversion load capable of carrying full turbine power. Wind
turbine and Rivgen power can be curtailed as needed. Further work is needed to specify alternate
electric heat loads, and ways of keeping the water tank and powerhouse heated when the diesels are
not running.
¾Wind turbine models, sizes and quantities considered. Power curves for each turbine.
Which qualified third-party test facility has certified the proposed turbines?
The power curve and certification for the EO25 turbine is attached. The Bergey 15 turbine is currently
completing certification. The QED turbine is just beginning certification.
¾Proposed budget and schedule based on current turbine pricing and constructionestimates.
Eocycle’s standard pricelist is attached. QED recently provided revised cost numbers based on design
Page | 45
improvements stemming from reviews by NREL under DOE Competitiveness Improvement Project
funding. This is discussed in the report final summary.
¾A list of what permits will be needed for the project.
Listed in report.
¾A copy of the geotechnical reconnaissance report.
Attached
¾HOMER model with accurate wind resource, electrical load, thermal load, wind turbine
power curves, turbine availability, diesel power curves and diversion loads. Pay special
attention to the excess power in the system and how that can be put to value-added use.
(Include the electronic HOMER file in your submission, but limit the printed report to
HOMER output from the proposed system.)
Our alternative approach using a custom time-series model is detailed in the report.
¾Show how the economies of scale are affected by using different types and quantities of
turbines. How do these options vary the overall system cost, the cost per installed kilowatt
and unusable excess power? This analysis should reflect that offsetting electrical load has
greater economic benefit than offsetting heat loads due to the varying efficiencies of diesel
generators versus oil-fired boilers.
This analysis is included in the report. A transition from conventional AEA philosophy is that diesels-
operation will be viable due to the very low cost of second-life (or in time new) batteries for energy
storage. A resilient system however should retain full diesel generation capability.
¾If the project involves, or could involve, the intertie of two or more communities, analysis
becomes more complex to determine where diesel and wind power generation are located
relative to community loads. Cost and efficiency of reliable communication between the wind
site and the powerhouse should be considered. Savings may be gained through consolidation
of bulk fuel facilities or idling of power plants. Further, the larger load of the combined
communities may allow for larger turbines with better economies of scale. These benefits
should be weighed against any loss of rural employment or higher heating oil delivery costs
for communities losing power plants.
No intertie is planned.