Utility Planning Lessons from Texas: Four Recommendations

Recommended new approaches to ensuring reliability and resilience in planning for high-impact common mode events.

Guest post by Dr. Ira Shavel

The events of February 15-19, 2021 in Texas and parts of the Midwest demonstrate the vulnerability of the power system to high-impact common mode events. The key feature that made the situation in Texas particularly disastrous was that cold weather and precipitation affected virtually all types of generation (gas, wind, coal, and nuclear). The weather affecting multiple generation facilities at the same time resulted in a major common mode event.

The events of February 15-19 point out the criticality of electricity in today’s society. In large parts of Texas, customers lost not only power, which for many also meant heating, but also lost water supply, and some experienced disastrous flooding from broken pipes. It is fair to say that the events of February 15-19 go well beyond what we usually think of as reliability. The failures exposed a lack of resilience in overall electric, natural gas, and water infrastructure.

Common mode events like these are not in the scope of utility planning. For example, in its November 5, 2020, Seasonal Assessment of Resource Adequacy, ERCOT considered extreme load conditions but did not consider the failure of a large number of generating units due to a common cause.

My colleagues Paul Centolella, Mark Gildersleeve, Alex Rudkevich, Richard Tabors, and I wrote a January 2021 report for the Electric Power Research Institute (EPRI), Exploring the Impacts of Extreme Events, Natural Gas Fuel and Other Contingencies on Resource Adequacy, in which we described the importance of planning for extreme weather (Chapter 4) and recommended strategies to enhance resilience (Chapter 6). The report also discusses cyber threats, which should not be discounted as a real possibility, especially in light of the recent SolarWinds hack that we know affected critical infrastructure.

We began our paper with the observation, “It is human nature to under-estimate the likelihood of extreme events. Across topics varying from weather to fuel supply and cybersecurity, today’s power industry employs planning methods that tend to understate the probability of supply disruptions affecting multiple units and their impact on consumers and the system itself.” The February 15 -19 events in Texas are not dissimilar to what occurred in February 2011 in Texas. Fortunately, at that time the results were not as disastrous, although 4,000 megawatts of customer load were shed. It may have seemed at the time a rare event, but it has reoccurred this year with even greater intensity. The event of February 2011 was not the only recent warning that this type of event could happen again in Texas. The “polar vortex” event of January 5-7, 2014 dipped into Texas but did not lead to outages. It should not have been a shock that extreme weather would occur again during Texas winters.

The polar vortex is a wind pattern that circles the North Pole and keeps extremely cold air confined near the pole. When those winds become unstable, the polar vortex can dip south bringing Arctic air with it. Some scientists believe that global warming is resulting in a less stable polar vortex, but this has not been proven.

The electric grid is moving into a new era in which a greater fraction of generating assets are wind and solar, and an increasing portion of the generation occurs behind the meter and is not visible to the grid operator. At the same time, the economy has become increasingly dependent on a highly reliable supply of electricity. With many people now working from home, a pattern that will probably continue in the future, highly reliable electricity to all consumers, not just the business centers of major cities, has become very important. The industry needs to plan to make electric service highly resilient to significant supply disruptions.

Report Findings

In our report, we observe that:

• Electric industry planners tend to understate the probability, depth, and geographic scope of many high-impact common mode events. The recent event lasted days, affected a broad area that included all of Texas, and came at a time when regional weather is usually starting to warm.
  
  
• The economic impact of weather events has been rising rapidly. Whether or not this is due to climate change can be argued, but the trend is clear.

Billion-dollar event frequency by type

• Natural-gas-based generation is a critical supply technology that is generally assumed to be available, but operational problems and regulatory issues sometimes result in capacity unavailability. Gas now makes up a larger fraction of the fleet than it ever has, and that fraction is growing over time as coal plants retire.

U.S. Generation Mix 2009 vs 2019

2009

2019

• The industry’s methodologies for calculating resource adequacy assume that outages and reductions in output are largely independent. Increased dependence on renewable technologies combined with a recognition of common mode events that affect multiple generators make it clear that the assumption of independence is no longer valid. As we saw in 2011 in Texas, in 2014 in the Midwest, and in February 2021 in Texas, an assumption of independence should not be made for planning purposes.

New approaches to ensuring reliability and resilience need to be developed that incorporate high-impact common mode events.

• To project future disruptive event probabilities, historical probabilities for the frequency, intensity, geographic scope, and duration of weather events should be adjusted to take account of recent climate trends. Probabilistic weather forecasts can help deal with the rising frequency, intensity, and duration of extreme weather events.
  
  
• The resource adequacy framework needs to be modified to reflect the depth, duration, and economic costs of unserved energy and supplemented to account for common mode events. Planning should include scenarios that are relevant to the specific region and should consider both investments and potential operational responses.
  
  
• The interaction between the natural gas and the electric power markets needs to be restructured to remove the operational inefficiency due to the nonalignment of the market cycles of the two industries.
  
  
• Planning in the power industry needs to evolve to acknowledge the stochastic realities associate with increased penetration of weather-dependent variable resources and changing consumer behavior. These changes can be addressed by the development of probabilistic metrics and analytic/modeling systems that can measure, probabilistically, the economic impacts of these changes beginning with the development of scenario planning methods of extreme events.

Key Recommendations

Among our key recommendations are:

• Develop regional scenarios of high-impact common mode events and estimate the probability distributions of the scenarios’ physical impacts and associated economic costs. The process of developing these scenarios is critical for planning systems that are resilient. The process of thinking through scenarios informs broad system planning as well as mitigation measures that can be taken in advance.
  
  
• Develop a modeling framework to combine an operational model of the natural gas pipeline network with a production cost power system model to better understand the joint vulnerabilities of the electric and natural gas systems. The natural gas and the electric system are planned and operated separately. As we describe in detail in the report, the day-to-day coordination of providing gas to generating units is a major system vulnerability.
  
  
• Improve existing metrics for measuring resource adequacy and resilience, and, where necessary, develop new metrics to address the shortcomings of existing metrics.
  
  
• Develop new modeling techniques that incorporate high-impact common mode events in power system resource planning. Most models use the notion of reserve margin, which is the excess capacity that is available at the time of system peak load assuming all generation resources are independent. That metric would have predicted a vanishingly small probability for the outages that occurred between February 15 and 19.

Guest contributor Dr. Ira Shavel is a Senior Consultant at Tabors Caramanis Rudkevich.

The opinions expressed herein are those of the author and do not necessarily represent the views of the Boston University Institute for Global Sustainability.