Accurate Models, Accurate Forecasts: Improving Resource Adequacy Planning

Navigating the energy transition requires innovation both in terms of technology and methodology. At the Ascend Summit 2023, Wesley Cole, PhD, Senior System Modeler at the National Renewable Energy Laboratory (NREL), Jared Hansen, Director of Resource Planning for the Idaho Power Company, and moderator Zach Brode, Manager of Analytics at Ascend Analytics, discussed the future of resource adequacy in a rapidly evolving energy landscape.  

Key Takeaways

• As extreme weather and prolonged outages become more common, resource adequacy planning will require a granular approach that incorporates hourly data.
• Variable resource and demand side management programs must be valued differently under different scenarios.
• Resource adequacy is a complex, dynamic problem. Decision makers must not limit themselves to one model or metric, and instead should adopt a broad, systems-based approach.
• The Ascend Analytics PowerSIMM platform helps balancing authorities capture a wide range of realistic futures and make informed decisions about resource adequacy.

Better Energy Resource Management: The Case for Thorough Resource Adequacy Methods

In his technical presentation, Dr. Cole argued that resource adequacy modeling must consider the timing of outage events. For instance, loss of supply causes relatively few outage events. Often during extreme weather conditions, however, an outage proves disastrous, as seen with winter storm Uri in Texas in 2021, which left millions without power and caused over 200 fatalities, many from hypothermia due to interrupted energy access. Outage length matters, as well. In one example, Dr. Cole highlighted that customers may be content with many one- to two-hour outages in a given decade but will not see a multi-day outage as acceptable, even if they occur very rarely. Despite utilities hitting their long-term outage targets, winter storm outages have been much longer than predicted in recent years. As the climate changes and weather becomes more extreme, Uri shouldn’t be written off as a 'black swan' event. These two parameters (outage length and timing) illustrate the need for updated, transparent resource adequacy criteria. A traditional loss of load hours (LOLH) threshold of 24 hours per decade makes sense on paper; however, it does not capture the harm done by prolonged outages, especially during extreme events.

“We need to incorporate more thorough methods of thinking about resource adequacy,” Dr. Cole asserted. Specifically, the vast amount of readily available data allows for a more granular consideration of the impact of weather, demand changes, and transmission availability than in years prior. “Every hour of the year matters. You have to look at all the hours, even if you normally wouldn’t look at them as ‘high risk,’ so you don’t get caught unprepared when you weren’t expecting it.”

Resource Adequacy Modeling in Practice: A Utility’s Perspective

Indeed, proper resource adequacy modeling requires accurate forecasting. Mr. Hansen corroborated and underscored the importance of integrating hourly data into forecasting and modeling rather than exclusively looking at annual data, as has been the practice historically. As Director of Resource Planning at the utility Idaho Power Company, which serves about 620,000 customers in the state, Mr. Hansen identified three major challenges within the context of resource adequacy modeling at Idaho Power:  

• Winter peaks prove increasingly difficult to cover because solar generation doesn’t cover morning or evening peaks.  
• Dismal, calm winter days, with high heating loads and little sun or wind, require large blocks of dispatchable energy from other sources.
• Difficult coverage days may occur more often than forecasted.  

“A couple of the resources we thought were solid are not as solid when you look at hourly dispatch,” Mr. Hansen noted.  

As a result, variable resource and demand side management programs must be valued differently under different scenarios. This uncertainty calls for a probabilistic approach to calculating the effective load carrying capacity (ELCC) of resources, or how much load the resource can effectively serve. Mr. Hansen’s team calculates ELCC based on how much of a resource is required to hit LOLH targets under a variety of weather forecasts, yielding an ELCC distribution that provides better fuel for decisions than a single number based on a single weather forecast.  

Thus, developing accurate forecasting models also requires seasonal considerations. For example, energy demands in winter may be higher than those in fall due to increased heating loads, and energy procurement in the summer may be higher than in the spring due to more solar hours. “That separation is simple, and is easy to understand,” Dr. Cole explained; this seasonal separation informs decision making so that the right resources are chosen at the right time.  

Even with the right resources, and even with conservative planning, Mr. Hansen and Dr. Cole underscored the importance of resource diversity and sharing between utilities.  

“There is a lot of value in connecting resources to share when one [utility’s load] is peaking and another is not,” said Mr. Hansen. “There will always be resources that are utilized, and if you can get them to where they can be used—there is a lot of opportunity for that.” He added that Idaho Power has been looking into connection and transmission expansion to tap into that diversity.

The Case for Accurate Forecasts and Robust Models

The growing need for accurate forecasting in resource adequacy requires changing modeling approaches and developing unique solutions that fit a specific need. Different solutions should be considered for different types of outages, and resources should be considered to bolster reliability. As Mr. Hansen and Dr. Cole described, the complex, ever-changing nature of resource adequacy means that decision makers must not limit themselves to one model or metric, and instead should adopt a broad, systems approach to address the nuances of the field.  

Ascend Analytics has forecasts and models readily available to tackle the challenge of resource adequacy modeling. Ascend Market Intelligence creates detailed forecasts for power and fuel prices at nodes across the country, based on historical trends and both current and upcoming policies. These forecasts are fed into the Ascend PowerSIMM model, which has two critical capabilities for resource adequacy calculation. First, PowerSIMM can run a stochastic resource adequacy simulation with hundreds of possible weather futures, essential for identifying the worst-case reliability scenarios that define resource adequacy. Second, weather drives loads, prices, and renewable generation in PowerSIMM. This correlation captures the reality that certain weather conditions create critical conditions for resource adequacy. Robust resource adequacy modeling with PowerSIMM helps balancing authorities capture a wide range of realistic futures and make informed decisions.