Best Practices for Modeling Extreme Weather in Power Systems

Five Do's and Don'ts for Incorporating Weather in Resource Adequacy Analysis

Image Source: The Weather Channel

Winter Storm Elliott has brought a devastating week of weather across the United States and power grids are feeling the strain. The extreme weather is not just impacting local distribution grids, but leading to shortages across the bulk power system as well. Grid operators in PJM, TVA, ISONE and elsewhere either shed load or came close to doing so during the last week - an extremely rare occurrence across North America.

While there will be plenty of time over the coming weeks and months to determine root causes, identify what went wrong, and opine about which resources were to blame - power system planners everywhere need to consider this another warning shot about our grid's vulnerabilities.

Over the past few years I have spent countless hours working with grid planners about how we need to redefine our collective approach to resource adequacy. Resource adequacy is the way we determine whether the power system has sufficient resources to meet load, across a range of uncertainties. These uncertainties include equipment failures, unexpectedly high load, and variability in wind, solar, and other resources. And the fact is while power system reliability was always dependent on the weather, it is becoming increasingly true.

To maintain reliability and resource adequacy during the energy transition, power system planners need to improve the way we model weather and plan for extreme events. Two weeks ago I shared some thoughts about Best Practices for Modeling Extreme Weather in Power Systems with the New York State Reliability Council. In that presentation I developed Five Do's and Don'ts for Modeling Extreme Weather.

1. Do consult a meteorologist and Don't go it alone.

Power system planning, particularly resource adequacy analysis, is at the intersection of engineering and meteorology. Too often I see power system engineers and planners (myself included!) develop assumptions on weather without consulting with a meteorologist or atmospheric scientist. The truth is cross-disciplinary analysis is required and power system engineers need to exercise caution when bootstrapping datasets, especially for outlier events like Winter Storm Elliott.

2. Do evaluate "Black Swan" events, but Don't try to assign probabilities

I have been increasingly vocal about stress testing the power system against potential high-impact low probability (HILP) events. This is needed for both the current and future power grids. But planners need to proceed with caution. These events may not fit neatly in the conventional planning reserve margin and the one-day-in-ten loss of load expectation (LOLE) framework our grids are planned to. These events shouldn't just drive investment in more resources, but rather investment in a more resilient grid generally. Stress testing, in conjunction with typical probabilistic analysis, is needed.

3. Do use real data, and Don't just evaluate a doomsday planning scenario

When trying to prepare for worst-case scenarios, the knee-jerk reaction is to develop an infeasible "what if" situation. Rather than utilize robust atmospheric and meteorological analysis (see #1) too often I see power system planners develop a doomsday scenario where everything goes wrong simultaneously. Load spikes, all wind and solar drops to zero, and transmission interconnections with neighbors are unavailable. While there are certainly correlations across these stressors, it is important to base analysis on likely or potential weather conditions rather than synthetic stress events.

4. Do consider weather in neighboring grids, and Don't assume each power system is an island.

While the North American power grid is made of a smorgasbord of ISOs, RTOs, utilities, and various balancing authorities, transmission links most of them across the Eastern and Western Interconnections. During extreme weather events, these transmission links can offer significant reliability benefits, but unfortunately many ISOs, RTOs, and utilities plan for islanded conditions without support from neighboring balancing authorities. This leaves a lot of value on the table and makes interregional transmission development difficult. Ongoing discussions at FERC and elsewhere are rightfully considering ways to improve these interconnections to help support reliability and resource adequacy.

5. Do model stressors to all resource types, and Don't assume extreme weather only impacts renewables.

Winter Storm Uri in 2021 and early indications of Winter Storm Elliott have shown that all resource types are affected by the weather. Natural gas is susceptible to fuel scarcity, wind and solar are dependent on underlying atmospheric conditions, coal piles freeze, and all equipment sees increased outages during extreme conditions. All too often, however, system planners assume only variable renewables (like wind and solar) are affected by the weather and assign "capacity credits" to discount their capacity counted towards resource adequacy. This capacity accreditation needs to apply to all resource types to ensure we have sufficient resources to meet load across a wide range of conditions.

While there is always uncertainty in the weather, there is one thing we can be certain of: our power system is becoming increasingly dependent on the weather, and we need to improve our analysis to plan for it both today and in the future.