|
|
| Fig. 1: Electricity generation by source in Texas. [3] (Reported by the New York Times using data from the U.S. Energy Information Administration. Image source: B. Alexander) |
In February 2021, the major Winter Storm Uri precipitated a major power crisis in Texas. Beginning the evening of Feb 14, the severe storm began to disrupt power generation capacity. By the peak of power outages on the 16th, it is estimated that slightly over 4 million customers in the state (31%) lost power. [1] The organization that manages the Texas power grid - the Electric Reliability Council of Texas (ERCOT) - reported an end to emergency conditions on Feb 19, which can be understood to be an end to the worst of the outages. Outages were accompanied by major surges in pricing.
The goal of this report, then, is to explain what happened during the crisis. This is made more complicated by a highly politicized discourse surrounding the crisis: attempts were made to blame either wind or natural gas generation failures for the crisis. [2] Power outage data can be presented in such a way as to appear to support either narrative, generating further obfuscation. This will be explored in some detail below in an attempt to bring clarity.
In order to understand the events of the crisis, we may examine reported power generation (Fig. 1) and total generator outages (Fig. 2). These two datasets tell a closely related story about power generation failures. While the total power supplied to the grid is a somewhat straightforward number, the meaning of an "outage" is less obvious. What is perhaps most confusing is that outage numbers reported by ERCOT don't mean precisely the same thing for every generation source listed here. While this subtly is motivated by very real differences in the power generation sources, it means that it can be helpful to focus on Fig. 1 in our first pass of putting together a picture of what happened.
 |
| Fig. 2: Electricity generation and storage outages reported by ERCOT from Feb 10-19, 2021. [6] Average outage generation capacity is also shown for each category Outage is computed as the difference between capacity (seasonal max) and produced power. Shown is the sum over unit reported outages. (Image source: B. Alexander). |
At around 6 PM on the evening of Feb 14, 2021, Winter Storm Uri hit Texas. [2] Shortly thereafter, the total load being supplied by the system falls by approximately 20 GW, about a third of its per-storm value. [3,4] The onset of the storm is shown in Figs. 1 and 2 as a vertical dashed line. Around this time, we can observe two important events. First of all, as the storm begins, total power generation actually increases, with most fulfilled by natural gas generation. This increased electricity consumption may be predominantly explained by customers turning on the heat as outdoor temperatures dropped. [2] Then, as the 14th of February turns into the 15th of February, energy generation from a number of sources begins to rapidly decrease. This is the onset of the power crisis.
The fall in electricity generation after the onset of the storm could mean only one thing: less electricity was delivered to customers. That is, blackouts. On the 15th, this meant 4 million households and business without power - or about a third of those served by ERCOT. [1] That this is a similar fraction to the drop in generation capacity is no accident. The energy to supply demand must come from somewhere: when that supply is reduced, the only real option is to remove some customers from the grid via blackouts. If that is not done, the entire grid will fail rapidly and catastrophically.
Resulting blackouts occurred statewide. Major blackouts, defined as those where at least 20% of customers lack power, occurred in 75% of counties, with 23% of counties experiencing such major blackouts for 48 hours or more. [1] That these blackouts were spread across the state is not surprising, given the large scale of the storm and the fact that most of the Texas grid is unified under the management of ERCOT.
Texas pricing structures also meant that some of the power that was delivered was vastly inflated in price. When the grid failed to meet demand, pricing structure regulations kicked in. In the San Antonio Zone, for instance, wholesale prices spiked from normal values of tens of dollars per Megawatt-Hour (MWH) to the cap of $9000/MWH. [2] It should be emphasized that there was variability in how this translated to the price paid by consumers: while it is most common in the United States for household rates to be disconnected from the wholesale rate (at least on short timescales), Texas had a uniquely high proportion of households paying rates that were pegged to the wholesale rate. This was a major topic of discussion in the subsequent media firestorm surrounding the power crisis.
When we look at the series of events surrounding the 2021 crisis, a particular question arises: What went wrong? It is clear that a lot of power generation failed when Winter Storm Uri hit Texas. It is also clear that the extreme weather conditions were the direct cause of these failures. What has remained a bit more difficult to parse is what precise pieces of the power generation system failed and how.
ERCOT, the organization responsible for managing the grid in Texas, has attempted to answer this question. In the immediate aftermath of the crisis, they compiled reports of derates (reductions in generation ability at specific generators) and outages (100% derates) from the vendors who supply electricity Texas. Fig. 2 shows these. As expected, these data show a large increase in outages after the onset of the storm.
Yet it is not trivial to interpret this data, especially for to wind and solar outages. Most of the wind "outages," for instance, are already in place before the onset of the storm. This can be explained by noticing that a wind outage was defined here with respect to "seasonal maximum capacity." For wind power, unlike natural gas, this is different from expected available generation. Even so, it is a bit unclear what this "seasonal maximum capacity" means in detail. We might expect
but that is not precisely what is happening here. This is most obvious if we examine solar power (not that for sources that don't consume fuel, they are typically generating all of the power they could be generating). If outages were defined using existing power generation, we would expect to see solar power outages spike every night, which does not happen. In fact, even though solar outages are reported to increase during the storm, the solar energy generated, as reported in Fig. 1, actually increases (one might hypothesize that fewer units were online but those that were online were more productive).
Despite these complications, one may observe some relevant facts about which generators failed. First, even before the storm, wind power is producing well below its seasonal maximum. Secondly, the increase in outages after the onset of the storm is dominated by natural gas generators, where outages spike by close to 15 GW. This accounts for roughly three quarters of the previously discussed initial drop in energy production. For comparison, at the onset of the storm, natural gas accounted for about 60% of electricity produced in Texas.
Perhaps more important than what specifically failed is why those generators failed. A rough relative accounting of this was given by the Federal Energy Regulatory Commission; these results are shown in Fig. 3. The largest causes of failure were reported to be freezing (including instrumentation, wind turbine blade icing), fuel shortfalls, and the frustratingly ambiguous mechanical/electrical issues (at least some portion of which was related to cold weather). [5] Much of this is simply a detailed way of saying that power generation equipment was not designed to be sufficiently robust to low temperatures.
 |
| Fig. 3: Number of outages by cause during the crisis. [5] Average offline generation capacity of outages is also computed and shown. Note that a single generator may have experienced multiple outages. (Image source: B. Alexander) |
Given the central role played by Natural gas in Texas' power generation during the crisis, the issue of fuel availability if of particular interest. Fuel availability became an issue because of reduced production and increased use of natural gas in the days before the crisis. [2] Gas production facilities in Texas' productive Permian Basin reached an 85% reduction in gas output as issues peaked on the 16th. [2] For processing plants in this region, 58% of this decrease was due to upstream supply issues, 17% attributed to weather-related mechanical issues, and, perhaps most importantly to our story, 25% resulted from power outages. [5] That is, as the crisis went on, the initial outages cascaded into pressure on the gas supply chain.
The failures of natural gas power generation was particularly painful for Texas because of how its grid relied on those facilities. Natural gas turbines are particularly good at responding to changing system loads: they can be spun up or down quickly (unlike, say, nuclear power plants) and are expected to operate in all weather conditions (unlike wind or solar energy). In Texas, as in many other states, they therefore formed the bulk of the reserve generators: designed to move into action in the event of load spikes or unexpected outages from less reliable electricity sources. Thus, when the gas plants themselves turned out to be susceptible to the cold brought on by Winter Storm Uri, there was nothing left to fill in the gap. When resulting outages hit the fuel supply chain, it created additional challenges with returning the grid to normal operations.
It is clear that the 2021 Texas power crisis was a major failure of grid infrastructure. Winter Storm Uri disrupted electricity production across the state. As a consequence, blackouts were similarly widespread. These massive disruptions were initially the result of weather-related equipment failures. These failure became a crisis when the natural gas generators that were intended to be the all-weather, adaptable component of the power system also failed. These initial failures caused major disruptions to gas fuel supply: processing facilities lost power and were unable to produce gas for the generators to burn, intensifying and extending the crisis.
© Ben Alexander. The author warrants that the work is the author's own and that Stanford University provided no input other than typesetting and referencing guidelines. The author grants permission to copy, distribute and display this work in unaltered form, with attribution to the author, for noncommercial purposes only. All other rights, including commercial rights, are reserved to the author.
[1] N. M. Flores et al., "The 2021 Texas Power Crisis: Distribution, Duration, and Disparities." J. Expo. Sci. Environ. Epidemiol. 33, 21 (2023).
[2] C. W. King et al., "The Timeline and Events of the February 2021 Texas Electric Grid Blackouts," University of Texas Energy Institute, July 2021.
[3] V. Penney, "How Texas Power Generation Failed During the Storm, in Charts," New York Times, 19 February 2021.
[4] J. W. Busby et al., "Cascading Risks: Understanding the 2021 Winter Blackout in Texas," Energy Res. Soc. Sci. 77, 102106 (2021).
[5] "The February 2021 Cold Weather Outages in Texas and the South Central United States." U.S. Federal Energy Regulatory Commission, November 2021
[6] "Generation Resource and Energy Storage Resource Outages and Derates, February 10-19, 2021," Electric Reliability Council of Texas, February 2021.