A Free-Market Energy Blog

Energy Emergency Alert! ERCOT’s Close Call of September 6 (Part I)

By -- September 12, 2023

Editor Note: This study by the Energy Alliance of the Texas Business Coalition is in two parts, with Part II tomorrow. MasterResource appreciates the opportunity for publication.

“Where only a few years ago modern natural gas peaker plants would have been called on to meet peak demand at 5:55 pm on September 6, unreliable solar generation is now filling much of that role.”

“The Texas grid almost failed because $29 billion of federal, state, and local subsidies over the last 18 years have left Texans relying on renewable energy sources that cannot generate electricity when it is most needed.”

When the Electric Reliability Council of Texas (ERCOT), the operator of most of the Texas electric grid, declared an Energy Emergency Alert 2 on September 6 for the first time since Winter Storm Uri, it exposed how renewable energy and renewable energy subsidies are rapidly undermining the reliability of the Texas electric grid.

Despite efforts of promoters of renewable energy to deflect blame in the aftermath of the winter emergency, the recent near grid failure despite declining demand points directly to wind and solar generation as the cause behind the grid’s reliability problems. Where only a few years ago modern natural gas peaker plants would have been called on to meet peak demand at 5:55 p.m. on September 6, unreliable solar generation is now filling much of that role. The problem with this is that the ability of solar generation to meet demand, unlike natural gas generation, rapidly diminishes as the sun goes down.

By 7:25, demand on the grid had dropped by 4,680 megawatts (MW) from the peak. But solar output declined by almost twice as much, 8,498 MW. Wind generation, already operating well below expected capacity during the day, also declined by 592 MW. At that point, there was nothing left for ERCOT to do but declare an emergency when the margin between demand and available generation capacity dropped to 1.6 percent.

The Texas grid did not almost fail on September 6 because it was too hot outside. The Texas grid did not almost fail because Texans were using too much electricity. The Texas grid almost failed because $29 billion of federal, state, and local subsidies over the last 18 years have left Texans relying on renewable energy sources that cannot generate electricity when it is most needed.

The Energy Emergency

At 4:45 Wednesday afternoon, September 6, ERCOT issued a Conservation Appeal asking “Texans to conserve electricity” beginning at 6 pm. The reason for the request was because of “low wind and declining solar generation into the evening hours” as the summer heat kept demand for electricity high.

                Source: ERCOT

Wind generation was performing well below expectations, with output having peaked about midnight 16 hours before. After a slight uptick at 8 a.m., about the same time solar generation came online, it continued its decline. Total renewable generation, the combination of electricity from wind and solar, peaked about 9 a.m. as demand was still on the rise. The increase of solar generation as the sun rose could not keep up with the decline of wind. Renewable output declined until about 1p.m., then held steady the rest of the afternoon until 6 p.m.

Temperatures increased throughout the afternoon, reaching their peaks between 4 p.m. and 5 p.m. It wasn’t a particularly hot day, however, compared to earlier in the summer. Dallas and Austin topped out at 102 degrees, Houston at 99. Demand for electricity followed a similar pattern. The need for electricity to power home air conditioning as workers headed home around 5 p.m. extended demand somewhat, but by 6 p.m. it also was declining. Crisis averted—or so it seemed.

                                    Source: ERCOT

It was at this time that the conditions which led ERCOT to issue its Conservation Appeal from 6 p.m. to 9 p.m. began to unfold. In the next hour, renewable generation dropped 16 percent, down to 13,576 MW. Demand was falling as well, but not as fast as renewables, particularly solar. As the sun neared the horizon, solar output followed it down, falling 31 percent.

At 7 p.m., demand was 96.9 percent of “committed” capacity. Given the high temperatures this late in the year, the 3.1 percent reserve margin was reasonable, though not ideal. But it would not hold at that level. By 7:15, even though demand had dropped by 1,219 MW, the reserve margin had fallen to 2.3 percent because renewables were dropping faster.

By 7:25, demand on the grid had dropped by 4,680 megawatts (MW) from peak. But solar output declined by almost twice as much, 8,498 MW. Wind generation, already operating well below expected capacity during the day, also declined by 592 MW. Reserves fell to 1.6 percent.

With conditions deteriorating to the minimum threshold, ERCOT declared an Energy Emergency Alert 2 at 7:25 for the first time since Winter Storm Uri. While there were no controlled outages, or blackouts, ERCOT asked Texans to “safely reduce electric use and have a plan to stay safe in case outages become necessary.” The impetus for the Emergency Alert was the narrowing of the margin between capacity and demand.

                                       Source: ERCOT

Not only were supply/demand conditions very tight at this time, but ERCOT’s ability to maintain the frequency of the grid was stressed. In its announcement, ERCOT explained that the “frequency of the entire grid must be maintained between 60.1 hz and 59.9 hz at all times.” Failure to do so could lead to the destruction of equipment and generation necessary to maintain grid stability.

After the announcement, grid conditions began to improve. In large part this was because ERCOT could obtain additional capacity because of the emergency. Additionally, demand continued to fall as the sun set and temperatures declined. By 7:40 p.m. the reserve level was back up over 3 percent. ERCOT announced the end of the emergency at 9:40 p.m.

Disappearing Renewables

Even before the Emergency Alert ended, renewable energy advocates were already looking for other culprits on which to place the blame. On X, Doug Lewin, founder of STOIC Energy, claimed the cause “was *likely* a big power plant tripping offline; see the frequency drop below. Storage set an all-time record when it was needed most, almost certainly preventing rolling outages.”

The problem with this take—even if a power plant did trip offline—is that Lewin and other supporters of renewables must ignore virtually every fact available that points to renewables as the major cause of the Texas’s wounded grid… including the events of September 6.

Perhaps the best way of understanding the harm caused by renewables is through examining the lengthening of the time during which the Texas grid is most at risk. Before wind and solar generation came to dominate the Texas grid, the tightest reserve margins almost always occurred at peak demand, generally between 5 p.m. and 6 p.m. This is about the time when temperatures peak and the migration of workers from office to home means that air conditioners across Texas are doing duty at work and at home. Yet this is not when emergency conditions occurred on September 6.

Woody Rickerson, vice president of ERCOT System Planning and Weatherization, explained how a reliability risk model they used shows that renewable energy has shifted the reliability risk on the grid at a June 19 reliability committee meeting.

“Due to higher solar penetration, EEA (Energy Emergency Alerts) risk shifts from late afternoon to early evening as solar production diminishes,” said Rickerson. “If we had this model and ran it three years ago, the [risk] would have peaked at 5 p.m.”

As we have seen, there was no emergency on September 6 when demand hit its peak around 6 p.m. The emergency came an hour and a half later when demand was on its way down. The problem was that wind and solar generation was falling even faster as renewables were operating far below their installed and expected capacities. The increase of wind and solar generation in Texas has extended and shifted the period of greatest risk to grid reliability to a time when the capability of available generation to meet demand is significantly diminished.

Intermittency on September 6

The inability for wind and solar generation to reliably perform when needed is known as the problem of intermittency. It is a simple concept—though many American politicians appear to have a hard time understanding it. For natural gas, coal, and nuclear plants to successfully operate, all they need is maintenance of facilities and fuel to generate electricity. Proper management of fuel supplies allows these plants to run as needed, with only periodic shutdowns needed. A typical baseload plant might operate 95 percent of the time if needed.

Wind and solar plants operate very differently. They do not use fuel. Instead, they use the wind and sun to generate electricity. While this sounds great to many people, the problem is that the wind does not always blow and the sun does not always shine. Thus, when the weather and nature do not cooperate, e.g., on cold winter days when clouds obscure the sun or hot summer afternoons when there is little wind, output from renewable generation fluctuates and/or diminishes. It may generate well below its installed or expected capacity or may not even generate at all. Renewables cannot be expected to generate electricity on any given day at any given time. This is intermittency.

Renewable Generation Contributions Sept. 6 (MW)
HourWindSolarCombinedExpected Capacity%Installed Capacity%
Avg.    61.88% 23.19%

Source: ERCOT

The intermittency of renewable generation was the primary cause behind ERCOT’s September 6  Level 2 Energy Emergency. To explore this, we will look at intermittency in two ways. One examines the output of renewable generation compared to its installed capacity. For instance, during the daylight hours of September 6, renewable generation operated at only 23.2 percent of installed capacity. In other words, intermittency reduced the output of renewables by close to 75 percent compared to thermal (natural gas, coal, and nuclear) generation.

Defying the concept of periodic stopping, this type of intermittency is actually very predictable. For example, though there are 39,411 MW of installed wind capacity in ERCOT, system planner know that most of that cannot be counted on. In the summer, the expected output of wind is just over one quarter of installed capacity at 10,427 MW.

This leads to the other way of understanding intermittency by comparing it against expected capacity. This is done by measuring the output of wind and solar generation against the expected, or projected, capacity during a season, day, or even hour. Forecasters look at expected wind speeds and cloud cover to provide a projected amount of generation. Using this method, renewable generation looks better. On September 6, it produced 61.9 percent of expected summer capacity. Yet this still pales in comparison to the average 95 percent run rate of natural gas plants; even though renewables are producing close to what is expected they still greatly underperform compared to other generation sources.  

The expected capacity of renewables can also be measured on a daily or even hourly basis. Grid operators in Texas spend a lot of time forecasting how much generation will be available on the grid in the days ahead. Renewable energy advocates proclaim these attempts to predict renewable output as the way to increase its reliability. Yet for all of their efforts, the forecasters often get renewable output wrong. But even when they are right, this does nothing to improve grid reliability. For instance, the reason ERCOT issued its conservation request was because it had correctly predicted the underperformance of wind and solar generation as the evening progressed. But this did nothing to avert the onset of the energy emergency. It simply served as a notice that renewable energy had extended the highest period of risk to the grid from the late afternoon into the early evening.

Renewable Performance: Actual vs. Expected – Sept. 6
Wind% ExpectedSolar% Expected
5:00 p.m.4,01538.51%10,98286.91%
6:00 p.m.5,69454.61%9,81377.66%
7:00 p.m.6,20059.46%4,18933.15%
7:25 p.m.5,12849.18%1,46211.57%
8:00 p.m.5,61253.82%320.25%

                                           Source: ERCOT

Part II tomorrow concludes the analysis.


Bill Peacock is policy director of the Energy Alliance, a project of the Texas Business Coalition. The Coalition highlights the aspects of the energy market that matter most to consumers: Reliability, Affordability, and Efficiency. Our research can be found here.


  1. Thomas Tanton  

    Kudos to you Mr. Peacock for a well reasoned and written piece that unfortunately tells a tale that is becoming more prevalent Few have recognized that hot weather strongly correlates with “no breeze.’ You also briefly mention the need for frequency control, lest rotating equipment overspun and destroy themselves. This is particularly troublesome for some grids when coupled with the cube rule and gusty winds. So to a large extent it’s dam*ed if the wind is not blowing and dam*ed if it is.) But that’s not the most important issue: the hubris of central planners and regulators who believe they can grok the complexity of something ‘like’ the grid.
    I so look forward to your Part II tomorrow.


  2. Kevin  

    This was obviously due to improper winterization of generators and the grid.


Leave a Reply