Resource adequacy (RA) programs — from PJM’s Reliability Pricing Model to California’s Resource Adequacy Program — are facing pressure from multiple directions: load growth, changes in generation mix, emerging technologies, interdependencies with fuel supply operations and regulations, extreme weather, energy policy, and interconnection bottlenecks. In response to these stressors, many RA programs, already complex, are becoming even more complex, confusing, and harder to manage. At their worst, they end up encouraging rent-seeking behavior rather than reliability.
Rather than continuing to reactively and continually tweak the current, complex models for RA, we argue that it would be helpful to take a step back and think about the design of RA programs from the perspective of what they are meant to do — ensure real-time reliability.
We don’t have all the answers and there is still much we all don’t agree on. Our goal is instead to offer an alternative perspective on RA, walk through some of its implications, and move the conversation forward in a few priority areas.
A Real-Time Reliability Perspective on RA
RA programs have their origins — whether via power pools or utilities — in long-term system planning. As a result, they tend to look forward from a planning period to real-time operations, and the reality and details of real-time operations are often ignored. An alternative way to think about RA is to start with real-time operations and work backward toward the planning horizon, a bottom-up approach in the figure below. Real-time operations is, after all, where the rubber meets the road with RA.
The “funnel” figure below illustrates the progressions from these different perspectives.
Figure 1: An Alternative, Bottom-up Approach to Resource Adequacy: Starting with Real-Time Operations and Working Backward Toward the Planning Horizon
Source: GridLab
The traditional, forward-looking (top-down) perspective starts with long-term portfolio planning and medium-term RA planning, and sets requirements and incentives based on an assessment of what operating conditions are assumed to be a year or many years in advance. The forward perspective tends to focus on “getting the modeling right” — with the hope that, with more modeling precision, we can accurately anticipate what will happen in real time. Unfortunately, “the hope on getting the modeling right” depends upon a bunch of heroic assumptions about what real-time conditions will be in an RA event.
The real-time reliability perspective on RA turns the traditional paradigm on its head. It starts from real-time operations (“you are here.”) and works backward to the kinds of incentives, requirements, and planning needed to make sure that the system can be reliably operated in real time (“how did we get here?”). This more bottom-up perspective acknowledges that planning models will always be wrong because they depend on forecasts and assumptions. What matters most is adaptiveness, risk management, market incentives, and feedback among operations, operations planning, and RA planning.
Rethinking RA Program Design
Our proposition is that a real-time reliability perspective on RA is better suited to managing emerging RA challenges. It implies several changes to RA program design, including the following. This list is not meant to be comprehensive.
- Performance incentives: We recommend moving back to real-time market prices as the primary incentive for generators to be available and dispatchable. There is no substitute for well-designed real-time markets with reserve shortage pricing, energy and reserve co-optimization, and reserve products that adequately capture real-time operating uncertainty.
The core problem with RA performance incentives is the challenge of defining stable capacity obligation periods that market participants can plan around. Availability incentives that have obligation periods defined over a longer time horizon may be more workable.
We believe new reserve products such as those already implemented in the Midcontinent Independent System Operator, the Southwest Power Pool, and the Electric Reliability Council of Texas are potentially more effective than RA performance incentives or RA programs at achieving the most important objective of RA programs — incentivizing behavior that contributes to reliable operations, given the myriad analytical uncertainties that plague the current top-down RA framework.
- Demand-side participation: Historically, RA paradigms are fraught with supply-side thinking. Increased demand-side participation in RA programs and real-time operations, beyond traditional demand response, enhances flexibility and reliability. Ideally, RA programs would be designed to allow load-serving entities to avoid RA charges by reducing metered net load during periods of system stress.
- Capacity crediting: Rather than attempting to develop “ever-more precise” yet overly complex estimates of effective load-carrying capability (ELCC), we recommend considering more practical approaches that incorporate real-time operating experience and account for changed market rules and operational practices to which resources have responded historically. More practical approaches imply finding a balance between simplicity and meaningful accuracy — what is the level of accuracy in capacity crediting that system operators need in order to be able to manage real-time reliability?
- RA visibility: Markets operate best with increased information to improve decision-making. Providing regular (e.g., annual, monthly), publicly available, ex-post analyses and forecasts of system RA along with linkages to pricing outcomes can enhance reliability outcomes and improve market performance.
- Planning-operations coordination: Coordinating planning and operations means creating feedback mechanisms between RA planning, month- or months-ahead operations planning (e.g., outage planning), and day-ahead and real-time operations. More adaptive planning and operations requires more communication and coordination between planning and operations departments.
These changes would all need to be negotiated through stakeholder processes. Some changes would require more departure from current practice, whereas others could build on existing practice. But change will take time. It’s important to begin the discussion now. Venues such as ESIG’s task forces may be suited for this discussion, building on previous efforts such as the recently released Electricity Market Visions to Support a Reliable and Affordable Electric Grid Under Electricity Decarbonization.
Fritz Kahrl
Managing Partner
3rdRail
Paul Sotkiewicz
President and Founder
E-Cubed Policy Associates
Michael Milligan
Power System Consultant
Milligan Grid Solutions
Priya Sreedharan
Senior Program Director
GridLab
Mark Ahlstrom
VP, Renewable Energy Policy
NextEra Energy Resources
Leave a Reply