Ron Jenkins via Getty Images
The following is a contributed article by Hala Ballouz, owner and president of Electric Power Engineers, in Austin, Texas; Joel Mathias, a PhD candidate in the Dept. of Electrical and Computer Engineering at the University of Florida; Prof. Sean Meyn, Robert C. Pittman Scholar Chair at the Dept. of Electrical and Computer Engineering at the University of Florida, and the International Chair at INRIA, in Paris, France; Robert Moye, senior vice president, energy management, at Tyr Energy; and Joseph Warrington, senior staff research engineer at Home Experience in Cambridge, U.K.
We are at a critical juncture in the energy industry, and a great responsibility lies on the shoulders of researchers and industry leaders to ensure that we correctly understand how the energy market in the Electric Reliability Council of Texas (ERCOT) failed during the 2021 electricity crisis, just as it did in the 2011 crisis.
The researchers and industry experts behind this article are strongly concerned that contributors to several publications have proclaimed that the markets essentially worked, such as the recent panel and the recent opinion article by Peter Cramton published in Utility Dive. Prof. Cramton’s experience and contributions in the power sector are significant, and while he makes many good points, the conclusion presented is alarmingly misleading: “The conclusion from the crisis is not ‘markets don’t work.’ Instead, it is ‘markets need to work better.’ Continuing the constant improvement of the market rules is the best path forward.”
Our scientifically documented research and substantial experience is in sharp divergence with this conclusion.
While much remains to be learned about the causes behind the failures in Texas this past February, it is obvious that some combination of the following occurred:
- There was insufficient installed generation capacity to meet load requirements, resulting from inadequate reserve capacity and generation outages that were not adequately planned for or managed.
- The amount of energy that could be imported from neighboring regions was limited.
- Generation equipment (thermal and wind) was not sufficiently “winterized”, despite Texas’ experience with such failures just 10 years prior. Following the crisis in 2011, FERC/NERC had recommended “the adoption of minimum, uniform standards for the winterization of natural gas production and processing facilities.” Unfortunately, these recommendations were only issued as voluntary guidelines in Texas: most companies failed to adhere to these guidelines.
- There were insufficient resources with alternative fuel supply (e.g., diesel) despite knowledge of the uncertainty of natural gas supply during extreme weather events.
- Too many resources were out for maintenance.
- A failure to coordinate the operational requirements of natural gas pipelines with load shedding protocols — power was cut to compressor stations leading to a cascading set of problems.
Many of these deficiencies are a consequence of the way deregulation was enacted in Texas. In light of the tragic failure in Texas, primarily as a consequence of following an approach to resource investment that relies entirely on the use of (short-term) markets, the claim that these markets “worked” is disputable.
Consider any engineering solution to solve a complex task. For example, a poorly designed flight control system may also work until it fails, but the observation that it worked for some time does not excuse the fact that the system causes a crash in stormy weather.
This is not hypothetical: The MIT rule for adaptive control systems was considered a breakthrough while it was working – three X-15 aircrafts flew a total of 199 times. However, the control system was discovered to be highly non-robust, resulting in the death of a pilot. As a consequence, the control discipline now puts much greater emphasis on offline testing through simulation prior to implementation. A similar approach is required in power systems, especially in power economics.
As we explain in our recent article, there appears to be little hope that one can incentivize investment in costly grid infrastructure based on extreme short-term price signals. In particular:
- $9,000/MWh is synonymous with blackout or shortage conditions. Therefore, any consistent argument that it forms an incentive for investment comes with an admission that Texans must tolerate blackouts and shortages like those experienced this past February.
- The $9,000/MWh shortage price is claimed to be a long- and short-term incentive for generators to be installed in sufficient numbers and with sufficient reliability to avoid the problems Texas experienced. However, if there are sufficient, reliable resources, the market will not experience shortage pricing and the resource owners will never be fully compensated for their investments.
- Even if generators obtain the shortage price on some occasions in practice, no financial model for a newly built generator, or life-extension project for an existing one, would place sufficient weight on the potential to earn income this way over the lifetime of the asset. Such events are too infrequent, and their durations are entirely unpredictable on timescales consistent with investment planning.
- A theme in this is that users must be exposed to real-time prices, and yet protected from risk. Presumably, it is possible to invent a “de-risked” real-time tariff such as the widely reported Griddy product. However, relying on an approach that counts on consumers responding to market prices, only to recommend tools (hedges) that shelter consumers from real-time volatility, appears contradictory.
- Consumers act as “virtual power plants” when they shed load during a $9,000/MWh grid event. However, neither the consumers nor the generators dispatched in the day-ahead market are compensated at this rate when the shortage occurs in the real-time market: only a handful of generators (mainly those providing operating reserves) stand to make massive gains during a crisis.
The relationship between electric power and the ultimate use of electricity is only loose in most cases. The foundations of price-based demand response are rooted in the theory developed by Dupuit in the 1840s, Hotelling in the 1930s, Coase in the 1940s, Vickrey in the 1950s, and Schweppe in the 1980s (see our recent article for references and further history).
In contrast to the direct value to the consumer for moving a product on a railroad, or the value of crossing a bridge versus taking longer alternate routes, electricity has always presented a challenge because the product that the customer values is not electricity per se, but the refrigeration, air conditioning, hot water, or lighting that comes from it. Over time, this important nuance appears to have been forgotten in power economics, leaving us with the crude and largely inaccurate assumption that demand responds smoothly to prices.
The most obvious examples of the disconnect between prices and consumer value can be found in the refrigerators and hot water heaters that consume power intermittently to maintain their temperature within predefined bounds. Fig. 1 illustrates the consequence of this disconnect, showing how each load class prepares for a critical peak price event, and the entire population shuts down at the onset of the event. Based on these findings, it is now known that the use of critical peak pricing will destabilize the grid if there is sufficient participation.
This conclusion is common sense: consider a typical residential water heater that is well-insulated, meaning it may remain off for six hours before heating is required. Isn’t it obvious that there is no loss to the consumer if the plug is pulled for 90 minutes? The figure illustrates this point for commercial and residential water heaters among several other common loads. Whether it is a 1% or 100% increase in price, the optimal consumer response will be approximately the same, because the change in power consumption results in no loss of value to the consumer.
Solutions: We propose the creation of a reliability system operator (RSO) that acts as a central planner and develops an optimal resource expansion plan across the entire market. The RSO would take on many of the responsibilities of today’s RTO or balancing authority. In addition, it would create contracts with generation companies and distributed energy resource aggregators to ensure reliability at low cost. As in most other countries, and many parts of the U.S., services would be obtained through carefully constructed, financeable contracts.
We must preserve the scope for innovation that comes with open electricity markets. But the “genius” of competition is still a prominent feature of coordinated procurement mechanisms. Those who are able to build competitive and efficient infrastructure will be the ones rewarded with contracts for generation capacity and ancillary services. We consider this a markedly better prospect than the model used today.