Beyond Capacity Adequacy

Imagine a future where all physical power system requirements needed for reliability are met through efficient market mechanisms. Procurement and compensation for the full set of system services would be signaled through a wide range of discretized, multi-timescale market products. Both demand- and supply-side resources could equitably and transparently participate. Real-time pricing, with fully responsive demand that can express its true value of lost load, would be a reality. Reliability would no longer be an ad-hoc process but would instead be integrated into the very core of an efficient supply-demand market balance.

This is a vision shared, at least in part, by many power system stakeholders. It would require a major overhaul to power system markets. Perhaps more importantly, it would necessitate a fundamental shift in how we view and implement reliability – requiring a move from adding up supply-side megawatts (MWs), to utilizing efficient market mechanisms to serve physical system requirements with the full suite of demand- and supply-side resources across multiple possible system configurations.

Broadly speaking, reliability consists of two parts: (1) capacity adequacy, which ensures that sufficient resources exist to supply power and energy requirements at a future time and location, with a certain probability of failing to do so, and (2) operating reliability, or security, which consists of prescriptive requirements to enable the grid to withstand sudden disturbances or unanticipated losses of system components. This blog post focuses on capacity adequacy, or resource adequacy (both are assumed to be equivalent here). Capacity adequacy is typically estimated with planning reserve margins or, more ideally, with probability-based metrics like expected unserved energy (EUE) that quantify the risk of a shortfall. This accounting process typically includes supply-side resources and accounts for the transmission network to transport that generation to load centers; demand-side resources, such as demand response and storage, have historically not been included, or have been represented in very simplistic terms. There is no universal capacity adequacy target; instead, each planning area sets its own target, which is often the result of a somewhat arbitrary selection process. In North America, the North American Electric Reliability Corporation (NERC) annually assesses, but does not enforce, seasonal and long-term planning reserve margins.

With so much recent news focused on the state of reliability in the United States, one might assume that capacity adequacy is a critical issue in today’s power system. But as reflected by large reserve margins in nearly every operating region in the United States, we actually have excess capacity. This has been driven by a number of factors, including a lack of incentives to exit the market, as well as incentives, including those signaled by well-meaning policies, to build more capacity.

However, this doesn’t mean that capacity adequacy is useless. It means that we need to view capacity adequacy differently. We need to think beyond many of the entrenched rules-of-thumb and heuristics that simply add up MWs, derated or otherwise, to meet a predetermined capacity target. Rather, we need to incorporate additional dimensions: consider the full set of system services required across multiple timescales, design efficient mechanisms to access resources with the capabilities to supply those services when needed in real time, and procure resources with those capabilities in the first place, which takes place years before the units are ever called upon. This means we think more deeply about ramping, frequency, and voltage support during the capacity assessment and procurement process. It means we account for the interactions across a fuller set of system components when making investment decisions in order to meet a given EUE target; for example, the interaction of chronological storage charging and discharging with wind, solar, and conventional generators. Some entities are already thinking about some of these considerations. For example, EirGrid’s DS3 Programme has implemented 14 explicit market products to signal grid services across multiple timescales, and multiple operating regions in the United States have or are developing plans for flexibility and/or inertia market products.

So how do we do this analytically? As with any analysis, the proper approach is to define the problem, ask the right questions, and then use the appropriate tools to inform decision making. As established above, the problem of achieving a reliable system via “capacity adequacy” must be viewed differently. With an evolving grid and a dynamic market landscape, the questions and tools we use also need to change. Our questions should shift from “how many MWs do we need?” to “what resources do we need to provide the full set of required system services under a wide range of possible futures?” This means we need a broader suite of tools with more extensive data inputs to capture the nuances of reliability across different temporal and spatial scales. This requires evaluating a broader set of scenarios, including those with correlated outages and fuel supply shortages. There is a particular need for data to better understand low-probability, high-impact tail events that our independent failure assumptions may miss. In addition, our capacity adequacy tools need greater representation across different timescales to capture potential reliability concerns across a wide range of operational scenarios; this could mean more seamless linkages and enhanced compatibility with other capacity adequacy tools, as well as linkages to other tools, such as capacity expansion models, production cost models, and AC power flow tools. This also means we need to include institutional considerations, such as market design, in linking the physical requirements of the system with the resources that can supply those capabilities. Ultimately, the choice of which tool(s) to use depends on the question at hand; there are model resolution (temporal and spatial), fidelity (physical accuracy), and computational tractability tradeoffs with every tool.

A future power system with efficient and holistic capacity adequacy processes is possible. It will require moving beyond traditional methods to an integrated approach that explicitly captures engineering, economic, and institutional considerations. In a world where reliability is the name of the game, we must evolve our capacity adequacy processes to value all services needed to support that end goal.

Bethany Frew
NREL