Jonathan Searls urges multi‑metric overhaul of resource adequacy in FERC presentation

Jonathan Searls, a presenter from Energy Exemplar, told a Federal Energy Regulatory Commission audience that resource adequacy methods must evolve to match changing generation mixes and accelerating load growth.

Searls opened by reading a standard definition from the North American Electric Reliability Corporation (NERC) — “the ability of the electric system to supply the aggregate electrical demand and energy requirements of the end‑use customers at all times, taking into account scheduled and reasonably expected unscheduled outages” — and then framed the practical test as whether a region has more supply than demand when scheduled and forced outages are considered.

He said traditional capacity‑based planning, which relies on fixed reserve margins and capacity‑outage probability tables, was effective for largely thermal systems but misses new risks in grids with more renewables, storage and rapid net‑load changes. To illustrate the classic approach he used a simple example of a 100 MW generator with a 10% forced outage rate and described how those probabilities combine across units to produce metrics such as loss of load expectation (LOLE).

Searls cited two broad shifts that complicate those methods: changing resource mixes and larger projected load growth. He showed national and regional charts indicating faster recent growth in solar and wind and noted regional contrasts — for example, he described ERCOT as having much higher renewable penetration, and SPP where wind is the largest producer. He said load‑growth projections for 2030 have risen sharply in recent planning work, citing an earlier 24 GW 2030 estimate versus a more recent roughly 166 GW projection tied to AI and data‑center demand and referencing EIA projections.

Pointing to the January 20–27 winter period in ISO New England, Searls described how constrained natural‑gas supply for generation (because gas went to heating) caused generators to shift to petroleum and dual‑fuel operation, underscoring how fuel‑supply limits can change adequacy outcomes.

From those examples he listed common modeling assumptions that may no longer hold: independent outages, firm fuel availability, focus on a single daily peak hour, copper‑plate transmission assumptions and perfect foresight of resource behavior. "Old tools won't solve new problems," he said while summarizing a DOE response that used multi‑metric frameworks and multiple historical weather years.

To better capture risk, Searls recommended a multi‑metric approach that includes frequency (LOLE), duration (loss of load hours) and magnitude (expected unserved energy, EUE). He argued models should be chronological and sequential — for example, sequential chronological Monte Carlo simulations — to capture intertemporal constraints, fuel and transmission limits, and correlated weather‑driven outages. He highlighted a CAISO net‑load example where net load rises from about 15,000 MW to roughly 35,000 MW over a handful of hours, emphasizing the need for fast ramping and flexible resources.

Searls pointed to several industry responses as examples: a DOE study that ran multiple scenarios and weather years and warned that status quo retirements are unsustainable; CAISO's recent proposals to revise capacity accreditation and planning reserve defaults (moving from a simple 15% default margin and noon‑to‑6 PM renewable performance windows toward Monte‑Carlo‑based accreditation targeting a 1‑in‑10 LOLE and using extensive weather/forced‑outage sampling); MISO's move to seasonal constructs and new risk‑analysis requirements; ISO New England's plans to shift toward seasonal auctions and accreditation that considers fuel‑supply firmness; and Hawaiian Electric's integrated planning work on small island systems that favors ELCC‑style accreditation.

He concluded that regions will likely need to adopt multi‑metric frameworks, focus on critical hours rather than only peak hours, model chronological constraints and operational detail, and use robust sampling across weather years. After summarizing those points he invited questions from the audience.

The presentation primarily described analytical and modeling recommendations rather than proposing a specific rule or vote; next procedural steps depend on each region's regulator or planning authority.