Among the more memorable presentations I have seen over years of energy reporting are those where the ‘whole system’ has come back to bite those who have done resilience planning that has been too tightly focused. A power plant control room that has passed through a major earthquake completely untouched – but is standing empty because the local access road is just a pile of rubble. A transporter full of flood barriers, stuck in traffic far from their intended site. Even a drinking water reservoir with a fully backed up power supply, emptied overnight, bowser by bowser, because a neighbouring supply area has a power outage for its own water supply.
February’s weather-induced blackout in Texas revealed once again how interlinked systems often thought of as ‘backup’ can, if not fully considered, equally be an Achilles heel.
An article published in Science included a number of ways that the state’s extensive gas network proved to be a hindrance, rather than a help, in solving the state’s power problems.
Texas is a centre of natural gas production and half of its production is used close to home, fuelling gas-fired power plants within the state. Those gas-fired plants provided 46% of the state’s power in 2020. But during the February event gas-fired power represented two-thirds of the generation deficit.
Overall, gas generation availability was 18GW below expectations and that included 8GW of gas plants that shut down on 15 February, as the state approached its coldest day, either because of gas shortages or because they had frozen equipment.
Low gas pressures in the pipelines were one reason for the shutdowns. During the crisis, 40% of Texan gas production capacity was out of action. The reasons included electrical compressors in the gas network that were out of action, as well as because wells or ‘gathering lines’ had frozen. The more modern production facilities were more highly electrified and therefore more at risk during electricity blackouts. Some gas producers had interruptible power contracts and had not registered as being critical infrastructure – so they stopped producing gas, because they were blacked out.
With those processing and production facilities shut off, gas pipeline pressures were driven down further, putting more power plant at risk.
There has been speculation that a Capacity Market like that of the UK would have helped solved the problem of lack of supply. But the reason gas-fired plants were out of action was that gas was not available. A CM would not solve that problem – unless it were one for gas.
It is worth noting that the gas and electricity sectors in Texas are overseen by different regulators. So far, only the electricity regulator has seen top level resignations and the start of discussions about whether and how to share the hike in bills that will otherwise land on consumers. Even within the electricity sector, unwinding price rises is complex. Electricity system operator Ercot activated a price cap that allowed for prices to rise up to $9k/MWh and kept it in place for two days after it stopped shedding load. Some parties say that was too long, and customers suffered from the high price, but others note that Ercot still had to shed load during that period. Ancillary services prices, in contrast, were not capped, and reached $22k/MWh. Gas prices also peaked.
Interestingly, renewable energy was not so unpredictable as gas. That does not mean it could be called on unexpectedly, but Ercot was prepared for low wind – it only expected 6.1GW of wind availability from 28GW of capacity. Some 2GW of wind went offline and due to low wind, ice formation on the turbine blades or frozen equipment (eg substations) – at least one of those could have been addressed by ‘winterisation’.
Uniquely, solar generation contributed its full predicted capacity (network issues also saw a nuclear plant go offline). That may have been a relatively small contribution – but it could have made a big difference for some customers, especially if combined with a battery. Many domestic gas users were unable to start up their in-house boilers (or ‘furnaces’) because the equipment had electrical controls that did not work when the power out – and so did a number of the central refuges.

Read the full paper here