Utility planners leverage better accuracy of DOE's renewable cost forecast tool
National lab researchers' methods confirm ongoing price trends, winning users over.
The newest numbers on generation resources' cost and performance for renewables from the Department of Energy (DOE) flagship research laboratory show why the underlying database is becoming the go-to tool for making forecasting decisions.
The long-term falling costs for wind and solar shown in the newest Annual Technology Baseline (ATB) projections from the DOE's National Renewable Energy Laboratory (NREL) are closer to what planners, forecasters and analysts see in near-term marketplace data, which has them turning more frequently to the DOE's new tools.
The ATB can be useful to any analyst without the in-house capabilities necessary to do long-term projections of renewables' costs and performance, NREL Research Scientist Wesley Cole said. "It can be used for integrated resource planning, for analysis of the cost impacts of environmental regulations, or by a grid operator who wants to know how much concentrating solar power will cost if it is added in 10 years."
Until recent years, utility and system planners could choose among generating resources with relatively stable capital and operating expenses, such as coal and natural gas plants. The cost and performance data used to form the basis for those choices were equally stable.
That is changing fast. Renewables penetrations are rising quickly on many of the nation's grids as their falling costs make their appeal too great for utilities to pass up. The costs of contracts for both wind and solar generation dropped 18% in the just first half of 2017, according to the Bloomberg New Energy Finance (BNEF) New Energy Outlook. With the costs of generation changing that fast, the stakes for planners in pursuit of the least-cost, best-fit investments are high.
An example is the "rush to gas" created by historic low natural gas prices between 2005 and 2015, which has put a planned $110 billion into natural gas plant investments by 2025, according to a study released in May by Rocky Mountain Institute. But the plummeting cost of renewables now has planners eliminating natural gas from future portfolios to avoid stranded assets.
Desperately seeking data
Planners have tended to rely on proprietary and in-house data for deriving generation resources' cost and performance projections. The leading publicly-available option has been the Annual Energy Outlook from the DOE's Energy Information Administration (EIA).
"EIA doesn't advise utilities on their investment and portfolio planning," Christopher Namovicz, EIA's renewable electricity analysis team leader, emailed Utility Dive. Most planners likely use input from "a variety of sources," but EIA "strives to provide data and analysis" that can "be useful to a wide variety of stakeholders (including utilities) in better understanding markets."
The ATB was developed in 2015 to provide "a set of cost and performance projections for renewable energy technologies as an alternative to EIA's," NREL Research Scientist Wesley Cole told Utility Dive.
The widely-referenced levelized cost of energy (LCOE) reports from Lazard are only "today's costs," Cole said. "EIA starts with today's cost and performance values and applies a learning algorithm for LCOE projections."
The accuracy of those EIA projections is "dependent on the learning algorithm," he added. "It has been improving, but it has tended to over-project costs for wind and solar PV because the algorithm did not capture the cost declines in the marketplace."
To provide alternative projections out to 2030 and 2050, NREL's ATB "benchmarks today's costs from a variety of NREL and external data to develop high, mid and low projections."
"The ATB projections are not better or worse but different than the EIA projections," added NREL research scientist Laura Vimmerstedt. "The ATB identifies cost reduction potential in ways the EIA's learning algorithm does not."
"We moved away from the EIA because it seemed to not capture as well the changes we were seeing with wind and solar."
Policy and regulatory studies principal engineer, MISO
EIA's LCOE estimates in the Annual Energy Outlook "are forward looking estimates/projections based on well-specified assumptions about future policy and market conditions," EIA's Namovicz said. "They are not unconditional forecasts or predictions of costs, but rather indicators of relative economic value under a baseline, reference set of conditions."
There have been "significant periodic changes" in renewables policies" that EIA did not account for, such as the 2012 changes in federal tax credits, he acknowledged. As a result, EIA did not incorporate the "substantial impact on market deployment and (largely as a result) realized costs" caused by those policy changes. It failed, for instance, to capture the drop in wind prices after the 2012 extension of the production credit, Namovicz added.
Midcontinent Independent System Operator (MISO) stakeholders recommended the ATB to its planners "because they found its forecasts of the falling costs of renewables more accurate," policy and regulatory studies principal engineer Stuart Hansen told Utility Dive. "We moved away from theEIA because it seemed to not capture as well the changes we were seeing with wind and solar."
The 2018 ATB
The 2018 ATB Cost and Performance Summary is the newest projected data for 12 generation technologies. Three key features of its "transparent methodology" target a better assessment of present and future LCOEs, NREL's ATB website reports.
First, to capture "the current state of technology," it compares input assumptions for capacity factor, capital expenditures, fuel cost, operations and maintenance costs to recent historical trends to show how closely they align. Second, it compares its projections to others. And third, it uses "normalized definitions" for those input assumptions.
For the first time, NREL has included financial assumptions used for interest rates, return on equity, and debt-to-equity ratios specific to each technology. The ATB demonstrates how they impact the cost and performance projections.
The ATB's granular definition of LCOE is "the energy component of electric system planning and operation," limited by the technology's "annual average capacity factor." The capacity factor is not detailed enough to capture "capacity and flexibility services," the website adds.
The ATB offers "R&D" scenarios that include LCOEs for only the "evolution of the technology" without the financial factors, as well as LCOEs for "R&D + Market" scenarios, which include the financial assumptions.
The variables that impact performance or reduce risk for using the technology go into the R&D scenarios, and the cost of capital is the "market effect," Vimmerstedt said. "The LCOE is all the technical costs and the cost of capital."
Finally, there are projections for Constant, Mid and Low technology cost scenarios. The Constant scenarios assume the relative costs of technologies will remain constant through 2050. Mid scenarios assume current "likely" technology advances and cost reductions. Low scenarios assume technology "breakthroughs" that take cost and performance to significantly lower levels.
The ATB does not provide analysis, NREL's Cole said. But in working with the data, he observed that the "low end" of the wind cost projections "got lower," suggesting more cost reduction is coming. "This was not a surprising finding," he said. "Wind costs were coming down last year and they are still coming down."
He was surprised that PV cost projections are lower again this year. "It has happened year after year, and I thought they were already fairly low last year, so it surprised me that it keeps happening," Cole said.
Vimmerstedt found that the "default capital recovery period" was extended from 25 years to 30 years this year. "That shows the financial industry sees less risk for recovery of its investments in these assets and more confidence in their value," she said.
An important feature of the ATB is that its cost and performance assumptions "allow high level comparisons" among renewable generating resources, she added. "The parameters for the cost of onshore wind in 2030 and the cost of utility scale solar in 2030 are the same so a utility can compare them."
How to use the data
Utilities have excellent cost and performance data through the periods for which their solicitations provide input, Cole said. "But they typically don't have the in-house expertise to do cost projections over a 10-year to 15-year time frame. That is the data they pull from the ATB."
A 2017 NREL study showed "numerous methodological differences" between EIA, NREL and Lazard projections, EIA's Namovicz said. It showed that, like EIA projections, ATB estimates are limited by being based on policy that can change, he added.
But the ATB's "near-term" case uses a historical year, Namovicz acknowledged. That allows it to "account for more of the as-realized policy and other market developments that are necessarily missing from the projected values from EIA."
The ATB projections for capital cost, operating cost and capacity factor have not been around long enough to assess whether they are definitively more accurate than EIA projections, NREL's Vimmerstedt said. But they differ in two fundamental ways.
MISO uses the ATB for its annual MISO Transmission Expansion Plan (MTEP), Hansen said. "The MTEP starts with a resource forecast that requires the best estimate of what the cost of new capacity will be."
MISO stakeholders proposed using the ATB because it is public, clear and well-documented, Hansen added. The cost projections are not precise but are accurate reflections of what the grid operator sees in utility solicitations.
As the ATB has reported, natural gas technologies are maturing and "don't tend to come down in cost, while wind and solar have been coming down even faster than the ATB projections," Hansen said. "But its numbers have been closer to our current costs than other projections."
Michigan's DTE Energy has used the ATB for developing forecasting assumptions since 2015, Renewable Energy Business Development Manager Terri Schroeder emailed Utility Dive. The ATB is now "a reference for forecasting costs in our renewable energy plans and in the integrated resource planning process."
DTE has found the tool provides "a great baseline for all of the technologies," Schroeder added. Michigan regulators "supported using it in filings as a publicly available third-party resource."
Both Schroeder and Hansen said Cole's personal support has also been a critical asset.
The ATB speaks
The greatest significance of the ATB may be in its ability to more accurately capture the cost trends for renewables and incorporate them into projections that planners use in making choices among resources. As Namovicz acknowledged, the EIA has missed those trends.
In an interview with Utility Dive about the future of larger-scale solar, Regulatory Assistance Project senior advisor Jim Lazar referenced the ATB to make the point that installed costs for larger solar systems are expected to drop more drastically than had previously been anticipated.
"At an installed cost of $1,000/kW [as the graph above shows,] larger-scale solar for commercial-industrial uses will be significantly cheaper than retail energy rates for customers, even for leased systems, sooner than had previously been thought," Lazar emailed. "Based on what we see here, larger-scale solar used on big-box retail stores is something to watch."
One picture, the saying goes, is worth a thousand words.
This story has been updated to correct comments from DTE Energy's Terri Schroeder, which were previously attributed to someone else.