A boom in solar at electric cooperatives is forcing one of the biggest such power providers in the U.S. to answer some hard questions about its power mix.

Tri-State Generation and Transmission still gets half its power from coal, even as co-op owned and purchased solar is reaching nine times what it was in 2013. A new report shows Tri-State, already the biggest supplier of solar among U.S. generation and transmission (G&T) co-ops, can lower customer bills by increasing its procurement of wind and solar, despite its existing coal investments.

"Tri-State has a significant amount of solar and procures almost a third of its power from renewables," Mark Dyson, an Electricity Practice Principal with think tank Rocky Mountain Institute (RMI) and lead author of the new report, told Utility Dive.

RMI found the costs of utility-scale solar and wind are so low that acquiring them can save Tri-State customers as much as $600 million through 2030 "even if it continues paying for and running the coal plants," Dyson said.

Tri-State does not doubt the importance of adding more solar and other renewables to its power mix. It does, however, question the accuracy and value of RMI's conclusions. But if the think tank is right, the remarkable boom in solar for co-ops, their power providers and other load serving entities may be just getting started.

Solar and cooperatives

Interest in solar at co-ops was just emerging in 2014, when the National Rural Electric Cooperative Association (NRECA) used a $4 million Department of Energy grant to launch its Solar Utility Network Deployment Acceleration (SUNDA) project. The project was designed to identify widely applicable tools and strategies co-ops could use to grow solar for their members.

Less than 1% of co-ops had solar arrays bigger than 250 kW back then. Now, the average co-op solar project is over 1 MW, half of U.S. co-ops have solar offerings, and cumulative co-op installed capacity will be over 1 GW by the end of 2018, NRECA reported.

"The transformation happened partly because the installed cost of solar came down so much," NRECA SUNDA team lead and report lead author Debra Roepke told Utility Dive. "But it is also because co-ops learned in the SUNDA project."

Through the five-year, three-phase SUNDA project, co-ops developed a tool set, Roepke said. "That began a transition in co-op adoption of solar as co-op leaders learned what the risks of solar are and how to manage them." By its culmination in September 2017, SUNDA had produced over 30 MW of solar at 13 co-ops, confirming the value of the tools.

Among SUNDA's products were a spreadsheet-based tool for estimating a solar installation's costs and output, as well as a Project Manager's PV Quick Start Guide and a Communications Toolkit. Most significantly, the SUNDA project produced a three-volume PV Field Manual with "the in-depth information the co-ops needed to successfully deploy their solar PV systems," NRECA reported.

Early deployed projects "tested equipment and designs at a cost of $4.50 per peak watt DC," NRECA reported. "The last SUNDA deployment, benefiting from lessons learned and declining costs, came in at $1.30 per peak watt DC."

The early growth driver was member demand for solar, but most co-ops now see it as cost-effective generation, Roepke said. And some now understand its value for meeting challenges like peak load costs or resiliency needs.

Anza Electric serves 5,000 electric meters in the California mountains between Palm Springs and San Diego. It brought a SUNDA-led 2 MW solar facility online in June 2017 and had already initiated a solicitation for a new solar+storage project when wildfires took down the link to its Southern California Edison power supply July 25.

"Resilience is why we started looking at this solar and storage project," CEO Kevin Short told Utility Dive. "In another system-wide outage like the one last month, we will be able to supply at least one circuit at a time. And we hope these two projects are just the beginning because, for long duration events, we will need much more."

If there is a next version of SUNDA, it is "likely to target solar-plus options like storage and distributed energy resources that leverage solar capabilities," Roepke said.

Before SUNDA, a distribution co-op built solar by using the carve out in its contract with its G&T provider that allows for a small portion of local generation, Roepke said. Some were SUNDA-led projects. Now, G&Ts, more familiar with solar, often work with distribution co-ops within the limits of those contracts.

"Many G&Ts were surprised at how much their members wanted solar," Roepke said. "They had been doing small demonstration projects; but the acceptance they got started discussions about larger projects that allowed economies of scale to bring prices down."

Building on the SUNDA experience, Georgia's Green Power EMC has added 20 MW and 52 MW projects and is developing a 200 MW installation. In addition, several other G&Ts have signed similar contracts​, Roepke said.

Before SUNDA, Tri-State's 30 MW New Mexico solar array was the only significant G&T-owned solar capacity, NRECA reported. By the end of 2016, nine G&Ts had partnered with distribution co-ops on solar projects totaling over 370 MW. "G&Ts' entry into solar has been a game-changer," NRECA added.

In an NRECA survey of 42 G&Ts, 75% of those not planning or offering solar said it was only because they had wind or hydro that could beat the cost of solar if members want renewables, Roepke said. That describes much of Tri-State's current portfolio.

RMI's report and Tri-State's response

RMI's case study of Tri-State called for "immediate collective action" between G&Ts and distribution co-ops to take advantage of the new low costs for utility-scale wind and solar. RMI modeled a "typical" Tri-State 247 MW coal plant's economics and reliability with three alternative scenarios.

Because it did not have access to proprietary Tri-State data, RMI used prices from the Xcel Energy Colorado 2016 all-resource solicitation bids for 2022 delivery and based its forecasts on Bloomberg New Energy Finance H1 2018 projections.

In its "Business-As-Usual" scenario, the hypothetical coal plant provides energy and resource adequacy to Tri-State, running at the 51% capacity factor that was "typical" in 2017, RMI wrote. Operating costs would be $40/MWh for fuel, operations and maintenance, and "annualized costs of required environmental compliance upgrades." RMI did not include depreciation.

In a lowered coal use "Fuel Saver" scenario, Tri-State procures 100 MW of wind and 100 MW of solar at Xcel solicitation median bid prices. Integration and transmission costs were added. The coal plant is kept operating to provide firm capacity. Despite its $40/MWh coal cost, the total portfolio's operating costs would be $35/MWh, according to RMI.

In a "retirement" scenario, the coal plant is shuttered, eliminating its fixed and operating costs, and Tri-State procures 342 MW of wind and solar. Another 167 MW of capacity is purchased at median Xcel prices to meet resource adequacy needs unmet by the wind and solar. That portfolio's operating cost would be $32/MWh, RMI found, again beating current costs.

RMI offered eight risk factors that the proposed move away from coal could help Tri-State avoid, quantifying three of those factors, lead author Dyson said.

The first quantified risk is distribution co-ops exiting their contracts with G&Ts to find lower prices and more renewables. RMI estimates taking steps to avoid this could save $13/MWh in rate increases that would follow a customer exit.

That risk is already a reality, Dyson noted. New Mexico's Kit Carson Electric took its 30,000 customers and 1.5% of Tri-State's load to an independent power provider in June 2016. Other members are reportedly negotiating an exit or considering it.

A second risk comes from a 2016 Federal Energy Regulatory Commission ruling that exempts distribution co-ops from contractual limitations on their renewables procurements if they meet federal guidelines, Dyson said. Providing member co-ops with renewables to keep them from doing more renewables procurement could also save Tri-state $13/MWh in rate increases due to lost load.

"If that ruling stands, it would open the door to more self-generation by distribution co-ops across the country and possibly lead to more contract exits," Dyson said.

The third risk quantified by RMI was future carbon pricing. Tri-State could avoid a potential $30/MWh rate increase to its customers by moving from fossil fuels to higher penetrations of renewables, RMI found.

G&Ts and other power providers face decisions about how to manage the cost burdens of legacy coal plants because expert forecasts have consistently underestimated price declines for renewables, Dyson said.

"Everybody thought solar would be way more expensive, which is why there are so many relatively high cost coal plants across the West," he acknowledged. "Solar and wind are now cheap, but the question is how to move forward."

Tri-State rejects RMI's conclusions as based on inadequate data, a complaint frequently made by utilities faced with criticism by those who do not have access to utilities' proprietary system information.

As RMI notes, there are considerable limitations to their study, Tri-State spokesperson Lee Boughey emailed Utility Dive. "It lacks the detailed inputs and complex models necessary to forecast portfolios and arrive at a credible conclusion on costs."

RMI's case study also requires adding some perspective, Boughey said. In 2017, Tri-State member co-ops got 30% of their power from renewables, two new projects will bring that to about 33% in 2018, and they "currently have a request for proposals out for additional renewables."

Tri-State is aware of, and already takes advantage of, the current low costs of market power and renewable resources, Boughey said. By blending lower-cost resources into its generation mix, TriState has retired capacity in one San Juan Generating Station coal unit "and will retire capacity in two others in Colorado in 2022 and 2025."

"Utility resource modeling is complex in scope, and no regulatory body, including our association's board, would substitute the RMI analysis for proper resource planning," Boughey said in a statement. In resource planning, models that include many factors "are run on an hour-by-hour basis over a period of many years to understand how costs and reliability are affected."

The right place for the kind of evaluation RMI proposed is in Tri-State's publicly held integrated resource planning proceeding, he added. "Tri-State encourages RMI to engage in its 2019 process."

Dyson said RMI will consider RSVPing to that invitation when the next Tri-State IRP process kicks off.