The Cold Trap: When Demand and Supply Fail Together

Cold weather is the purest teacher of correlated failure the grid has, because it attacks from both ends of the balance sheet at once, and it does so through a single mechanism you can actually name. Understand the cold trap and you understand the general principle better than any abstract lecture on statistics could deliver.

Begin with demand. As temperature drops, electric heating load rises, and not linearly — heat pumps lose efficiency in deep cold and switch to resistance heating, which draws far more power for the same warmth. Millions of homes doing this at once, in the coldest hours before dawn, drive electric demand to levels the system may never see in any other season. So far this is merely a big load, and big loads are survivable if supply holds.

Supply does not hold, and here is the mechanism. A large share of modern generation burns natural gas, and natural gas is delivered, not stored on site. It arrives by pipeline, in real time, from wells that are producing in real time. Deep cold does two things to that chain simultaneously. At the wellhead, water in the gas stream and in the gathering lines freezes — a 'freeze-off' — and production drops precisely when it is needed most. Downstream, residential and commercial heating, which in most jurisdictions holds legal priority over power generation, draws the pipeline down. Pressure falls. The gas plant that assumed firm fuel discovers its fuel was never truly firm; it was firm on ordinary days and interruptible on the one day that mattered. The plant cannot run, or cannot run at full output, and it does so at the exact hour demand is peaking.

Now layer the other failures, all sharing the same cause. Wind turbines ice and shut down to protect their blades. Instrumentation freezes — the sensing lines, the water traps, the small ancillary systems that a power plant needs to operate but that nobody weatherized because weatherization costs money against a risk that only materializes rarely. Coal piles freeze solid. Even nuclear units, robust as they are, can trip on frozen sensing lines in equipment that was never expected to see such cold. Each of these is a separate plant, a separate technology, a separate company — and every one of them fails for the same reason, in the same hours, because a single mass of cold air is the common cause behind all of it.

This is why the reserve margin from the previous module evaporates. The planner modeled each unit's forced-outage rate independently and summed the risk, concluding that simultaneous failure was astronomically unlikely. The cold does not know it is supposed to be unlikely. It removes a quarter, a third, sometimes more of the fleet in a single stroke, while demand sets a record, and it does this through the plainest physics imaginable: water freezes at a known temperature, and the entire system was quietly full of water that nobody had protected.

The cruelest part is the feedback loop between electricity and gas. Some of the equipment that keeps gas flowing — compressors, processing plant, wellsite controls — now runs on electricity rather than on gas engines. So when the grid sheds load to survive, it can cut power to the very facilities producing and moving the gas that the grid's own generators need. Electricity depends on gas; gas increasingly depends on electricity; and in a deep freeze each starves the other. Rolling blackouts intended to protect the grid can deepen the fuel shortage that caused them. The two systems were built by different industries, regulated by different bodies, and coupled by physics that neither fully owned.

The practitioner's takeaway is to stop treating 'a cold snap' as a demand event and start treating it as a correlation event. The question is not 'how cold will it get' but 'what single cause could take down supply and lift demand at the same time, and how much of my system shares that cause.' Weatherization — heat tracing on sensing lines, enclosed wellheads, winterized controls — is unglamorous and looks like money spent against nothing, right up until the morning it is the only thing standing between you and a multi-day blackout in lethal cold.

Faraday's read: the cold trap is not a failure of any one plant. It is a failure of imagination about shared causes, dressed up as a run of bad luck. There is no such thing as bad luck at that scale. There is only correlation you declined to model.