Microgrid Economics

To: "Wayne E. Copeland Jr." <wecjr@cox.net> Subject: capital cost of solar electricity

This is an excerpt of a piece by Willem Post in Citizenswindtaskforce.org blog this week

From: Willem Post [1]

ECONOMICS OF Green Mountain Power’S STAFFORD HILL “MICROGRID/ISLANDING” PROJECT

Microgrids and islanding are the wave of the future, as envisioned by GMP. Such setups include solar systems combined with battery systems, plus heat pumps for building heating and cooling, Powerwall 2.0 batteries at user premises, demand management by turning on/off user appliances, and, in the somewhat distant future, the management of EV batteries. 
Other Measures Should Come Before Microgrids and Islanding: Long before the GMP vision could possibly be realized at an acceptable cost, it would be much wiser, and much more economical, to first insulate and seal almost all buildings to be close to “zero-net-energy” or “energy-surplus”, and to have almost all light duty vehicles as high-mileage hybrids, such as the Toyota’s Prius hybrid, rated at 52 EPA Combined.
Those two measures would actually reduce the energy bills of already-struggling households and businesses, and would significantly reduce the CO2eq emissions associated with the building and transportation sectors, which emit about 65% of Vermont’s 8.3 million metric ton of CO2eq emissions.
Vermonters have absolutely no need for EVs (and associated charging stations all over the state), which perform poorly in winter, on muddy dirt roads, with snow on roads, and going uphill. How many all-wheel-drive EVs and 4-wheel drive EVs are being marketed? Regarding EVs, the wisest approach would be to wait at least 5 years.
The Need for Battery Systems: With many solar systems on a distribution grid, their output would become highly variable/irregular during variable cloudy weather, as already is the case in southern California and southern Germany. Battery systems are required to quickly “smooth” the variations/irregularities. The levelized cost of those battery systems is not charged to solar system owners, the disturbers.

http://www.windtaskforce.org/profiles/blogs/solar-energy-does-not-r... <http://www.windtaskforce.org/profiles/blogs/solar-energy-does-not-reduce-grid-investments>

Stafford Hill solar/battery system combo, owned by GMP

Battery system: 4 MW/3.4 MWh; turnkey cost $5.345 million; $1572/kWh Battery system operated to deliver 2 MW for 1.7 hour, for a total of 3.4 MWh Solar system: 2.5 MW DC/2.0 MW AC; turnkey cost $7.155 million Total = $12.5 million

NOTE:  

Sterling, MA, solar/battery system combo, owned by Sterling Municipal Light Department Battery system: 2 MW/3.9 MWh; turnkey cost $2.7 million; $692/kWh Battery system operated to deliver 2 MW for 1.95 hour, for a total of 3.9 MWh Solar System: 2.4 MW (existing); turnkey cost about $7.0 million Total = $9.7 million Stafford Project Economic Payback

Solar System: The solar system part of the project has estimated federal and state tax savings due to rapid 5-y depreciation and due to the federal and state ITCs, i.e., up front cash gifts. See table 4.
Table 4/Year1234566-y total

State tax not paid52154947404993822808222431488243371 Fed tax not paid19650135695018815185935838035608916947 Total tax not paid24865545169023808910874310604570961160318 State ITC gift225271225271 Fed ITC gift14440461444046 Total ITC gift16693171669317 Total tax subsidies191797345169023808910874310604570962829635 Battery System: The battery system part of the project has estimated federal and state tax savings due to rapid 5-y depreciation and due to the federal and state ITCs, i.e., up front cash gifts. See table 5.

Table 5/Year1234566-y total State tax not paid38954707623729917036166131112181776 Fed tax not paid14676826660914053164186625934188684875 Total tax not paid18572333737117783181221792065300866652 State ITC gift168257168257 Fed ITC gift10785701078570 Total ITC gift12468271246827 Total tax subsidies1432550337371177831812217920653002113479

With various grants, subsidies, federal and state ITCs, tax benefits of 5-y depreciation write-offs, the rosy illusion is created of a payback of 5 - 6 years (as claimed by GMP), whereas, the actual payback, without grants and subsidies (which are scheduled to be decreasing), is at least 2 times that long, about equal to the 10 - 12 year warrantee life of the batteries. The project, owned by GMP, is not eligible for Standard Offer status, and therefore GMP cannot sell its electricity for about 13.5 c/kWh. However, GMP is allowed to earn 9%/y on assets. That cost is significant, but is not shown in above table, as are several other major costs. Battery System Operation: GMP operates the battery system as follows:

1) Regulation: During regulation mode, its normal operation, the quick-responding battery system may be operated at 75% charged, so it can very quickly absorb or supply electricity from and to the grid, a service traditionally performed by less-quickly responding, small capacity, diesel and gas turbine generators, which release CO2.
GMP was credited about $100,000 for this service by ISO-NE, during the period of 1 January 2017 through June 2017. See URL, page 62. In this article, the credit is assumed at $200,000 for an entire year.

http://legislature.vermont.gov/assets/Documents/2018/WorkGroups/Hou... <http://legislature.vermont.gov/assets/Documents/2018/WorkGroups/House%20Energy%20and%20Technology/Reports%20and%20Resources/W~Anne%20Margolis~Energy%20Storage%20Report~11-15-2017.pdf> The variable daytime solar electricity (DC) is charged into the battery, then discharged, then converted from DC to “smooth” AC acceptable to ISO-NE, then fed into the grid. That process has an electricity loss of about 15%.

At night, the batteries are charged from the grid to continue regulation mode. That process has an electricity loss of about 20%, because of the extra step of grid AC to DC conversion before charging the battery.
2) ISO-NE Charge Reduction: GMP discharges the battery system during the maximum peak hour of the year, as measured by ISO-NE, to reduce its demand on the NE grid by about 2 MW, and thereby reduce the ISO-NE-imposed Forward Capacity Market charges, a saving of about $200,000/y, similar to the Sterling, MA, facility.
The battery system is similarly operated during the maximum peak hour of each month, as measured by ISO-NE, to reduce the ISO-NE-imposed Regional Network Services charges, a saving of about $200,000/y, similar to the Sterling, MA, facility. See table 6 and URL.

http://www.windtaskforce.org/profiles/blogs/solar-energy-does-not-r... <http://www.windtaskforce.org/profiles/blogs/solar-energy-does-not-reduce-grid-investments>

Table 6Charge ReductionRemark

Forward Capacity Market $200,000/yOne peak hour/year Regional Network Services $200,000/yOne peak hour/month Total $400,000/ySimilar to Sterling

3) Loss on Solar Sales: The variable daytime solar electricity (DC) is charged into the battery, then discharged, then converted from DC to “smooth” AC acceptable to ISO-NE, then fed into the grid. That process has an electricity loss of about 15%.
Annual generation 2.5 MW x 8766 h/y x 0.145, CF 3,177,675 kWh/y 

Annual cost of generation 13.5 c/kWh x 3,177,675$428,986/y Heavily subsidized The 13.5 c/kWh is based on the values in table 3.

Annual battery charging/discharging loss 0.15 x 3,177,675 kWh/y476,651 kWh/y 

Annual cost of battery charging/discharging loss 0.15 x $428,986/y$64,348/y

Annual REC sale: $0.03/kWh x (3,177,675 - 476,651)  = $81,031/y

The annual sale of Renewable Energy Credits is based on electricity fed into the grid.

Annual sales at midday wholesale prices: $0.06/kWh x (3,177,675 - 476,651) = $162,061/y
Annual solar net loss: ($428,986/y, generating cost) - ($162,061/y, sales revenue + $81,031/y, RECs revenue) = $185,894/y. That loss is charged by GMP to ratepayers via the rate schedules, as are all other costs.
Summary of Operations: Table 7 shows a summary of the above three items. It is clear, the total gain is grossly inadequate for a $12.5 million project, demonstration or not.
 Table 7/Annual project net gain $/y

1Regulation200,000 2ISO-NE charge reduction400,000 3Loss on solar sales185,894

Total gain414,106

Cost items not shown are:

- The GMP gain of federal and state tax savings. See table 4 and 5

- The GMP gain of federal and state ITCs. See table 4 and 5 - The GMP gain of 9%/y on assets - The amortized cost of borrowed money for the battery system - The replacement of about 75% of the battery system in about year 15, or sooner. - Charges of GMP personnel - Use of other GMP resources - Interest on borrowed capital

If the above cost items were fully accounted for, the project would be drowning in red ink.
Advocating more of such heavily subsidized projects would be great for GMP, but very expensive for already-struggling households and businesses trying to make ends meet in a near-zero, real-growth Vermont economy. Such poor-return projects would build up an increasingly stronger headwind against Vermont’s economy.
Estimating the Annual Cost of the Battery System: This section shows an estimate of the cost of the battery system. GMP likely has a spreadsheet with such an analysis, which should be made public, but is not “for business secrecy reasons”. For my analysis, I assumed:
- The useful service life at a generous 15 years

- 75% of the system to be replaced in year 15 - The cost of the replacement is generously assumed to be 50% of the 2016 cost - GMP, per statute, is entitled to earn 9% on invested capital If replacement capital in year 15 is (5,345,000, battery system) - (0.75 *5,345,000/2, replacement) = $2,004,375 (no inflation), or $133,625/y, allocated to 15 years.

GMP earns 9%/y on the remaining capital, such as (5,345,000 - 133,625) x 0.09 = $469,024 for year 1, or $5,772,600 for 15 years
The cost per year is ($133,625, capital loss) + ($469,024, GMP 9% gain) = $602,649 for year 1, which would be slowly decreasing to $434,281 for year 15. See table 8.
Various grants and subsidies for the battery system would reduce above numbers similar to wind and solar systems. The total reduction would be “socialized”, i.e., spread out over Vermont households and businesses, by means of taxes, fees, and surcharges, and higher electric rates, and higher prices for goods and services. There is no free lunch.
It is abundantly clear, the total of the annual battery costs would completely eliminate the gain in table 7. That gain usually is presented as a press release to the uninformed public, legislators, etc.
NOTE: There are other major costs, as listed under Summary of Operations, but those are not included in this analysis.
NOTE: For comparison, amortizing the $5,345,000 battery system cost, less ITCs of $1,246,734 = $4,098,266 at 5%/y would require annual payments of $509,200 for each of 15 years, which gives some indication of costs not mentioned.
Table 8/Capital cost5345000Return, 9%/yTotal cost

YearCapital usedRemaining cap0.09 11336255211375469024602649 21336255077750456998590623 31336254944125444971578596 41336254810500432945566570 51336254676875420919554544 61336254543250408893542518 71336254409625396866530491 81336254276000384840518465 91336254142375372814506439 101336254008750360788494413 111336253875125348761482386 121336253741500336735470360 131336253607875324709458334 141336253474250312683446308 151336253340625300656434281 Replacement20043755772600