A New Evaluation Paradigm for Flexible Resources
Important policy guidance was issued late last year that could impact the country’s investor-owned utilities, and it largely flew under the radar.
The National Association of Regulatory Utility Commissioners (NARUC), the organizing and policy setting association for the country’s public utility commissions, issued a resolution document that sets the stage for what represents prudent assessment of energy storage and flexible resources.
In the resolution, NARUC weighed in on the unique attributes of energy storage, stating that it represents a “fundamentally a different class of resource.” Accordingly, NARUC argued, utilities should develop “new modeling tools and new planning frameworks that allow for a more complete evaluation of flexible resources, such as energy storage.”
State-of-the-art modeling requires three critical criteria to incorporate and solve for the complexities of a renewables-focused electricity market: 1) sub-hourly dynamics consistent with observed and forecast real-time prices, loads and renewable production; 2) impacts of weather on renewable generation, load and market prices; and 3) imperfect foresight included in unit commitment optimization.
Inclusion of these three critical components address the principal factors necessary to realize the value for more storage and more flexible generation, as we explain below.
Three criteria for a renewables-heavy grid
1. Stochastic sub-hourly forecasting is needed to match real-time volatility.
Determining the cost of a resource mix based on deterministic scenarios leads to a suboptimal selection process. As we convert our resource portfolios to renewable resource options (primarily solar and wind), power production and clearing prices will be increasingly volatile.
We’ve seen this in the California and Texas markets, two leaders in renewable penetration, as day-ahead (DA) and real-time (RT) prices often clear at extremely low or even negative values. For example, in California, as variable renewable energy penetration (wind and solar) grew from 8 percent five years ago to 19 percent last year, DA and RT price volatility (as measured by the standard deviation of prices across major nodes) increased by roughly 200 percent and 50 percent, respectively.
Stochastic approaches allow the testing of up to 100 scenarios, as critical inputs based on historic distributions are allowed to move in an unconstrained manner. When resource decisions are made on an hourly basis, it leads to batteries and other flexible generation resources being undervalued.
2. Weather-induced volatility will be a critical element.
In the past, only load was subject to weather variability. Now, supply is as well, increasing the impact of weather on our ability to meet load obligation. Due to the variable output of renewables, high penetration of solar and wind will lead to highly volatile net load and market price conditions.
Weather-related impacts on generation and load, outages of generation and transmission assets, and congestion all cause volatility in the DA and RT markets, representing a critical risk exposure to ratepayers and investors. Utility modeling practices need to incorporate weather and its volatility characteristics as a primary input to ensure physical and financial risk factors are fully comprehended in a world where weather plays such a significant role.
3. Perfect foresight is highly imperfect.
Most commercially available production cost models incorporate “perfect foresight.” In other words, they know when load will be high or a generation or transmission outage will take place, and minimize costs or maximize value by running the optimal unit commitment and dispatch.
In the real world, we have something called forecast error, which is caused by a wide range of factors from intermittent renewable forecasts diverging from actual production to unanticipated outages. That is, if we could predict load and generation supply perfectly, there would be no market volatility. But since that will never be true, we should stop relying on models that pretend to know exactly how load is shaped, transmission congestion will impact power imports, and variable resources produce energy.
A paradigm shift
Utilities have been tasked by policymakers with transforming our electricity grids to be highly dependent on renewable supply resources in a few short decades, all while electrifying the transportation sector and other components of the economy. Achieving those extraordinary goals will require a paradigm shift in the way we assess and operate our resource options.
We believe NARUC has correctly steered to the horizon, guiding our public utility commissions on emerging prudency requirements. We would encourage PUCs and the utility industry to take on the innovation task at hand and put in place analytic solutions that are consistent with capturing the complexity of these changing dynamics.
Mike Mendelsohn is the manager of marketing and analytics at Ascend Analytics, with 25 years of experience in utility regulation, wholesale electricity markets, and thought leadership in solar finance and capital acquisition.
Dr. Gary Dorris is co-founder and president of Ascend Analytics, a software and consulting firm with integrated approaches to physical and financial markets. Dr. Dorris has conducted resource planning, portfolio management, and asset valuation services for over 50 utilities across the country and led valuation analyses for three dozen storage projects.