The intermittency problem

A gas plant has a dial. The operator turns it up when the evening rush hits, eases it down when the country sleeps — burning more or less fuel, on command, to follow demand. That dial is the quiet hero of the last item: how the grid keeps supply matched to demand every second.

Now picture a solar farm. There is no dial. At noon it pours out power whether you want it or not. At nine in the evening — when everyone is home, cooking, lights on — it gives you nothing, no matter how hard you turn. The sun has set. That gap, between when clean power arrives and when people actually need it, is the hardest problem in the energy transition.

Some sources you command; some you take as they come

The plants we have leaned on for a century share one prized trait: you can dispatch them. Dispatchable means switch-on-demand — coal, gas, and nuclear all sit there with a dial, ready to ramp up or down on the operator’s word. Demand climbs at six in the evening, you burn more gas. That is exactly what the grid needs, because, as the last item showed, supply must equal demand at every single second.

Wind and solar do not work like that. A solar panel makes power when the sun shines on it — and only then. A turbine spins when the wind blows — and only then. The sun sets every night, clouds roll across at midday, the wind drops for a calm week. None of this answers to a dial. The power arrives when nature allows, which rarely lines up with when people ask for it.

This is intermittency — sometimes called variability: the source produces on its own schedule, not on demand. It is not unreliable in the sense of breaking down. It is reliable in a way you cannot control. The sun is utterly dependable about setting at night, and that dependability is the problem.

The hard part used to be cost. Now it’s timing

Here is the reframing that catches most people out. For decades the objection to wind and solar was simple: too expensive. That is no longer true. The cost of building solar panels and wind turbines has fallen dramatically — solar especially, by something like ninety percent over a couple of decades. The price of generating one unit of electricity from a solar or wind farm is now often the cheapest there is, cheaper than new coal or gas.

So the hard problem moved. It is no longer the cost of making the electricity. It is the timing — getting that electricity to exist at the moment people want it. You can have the cheapest power in history and a dark city at 9 p.m., because the cheap power all showed up at 1 p.m. and went home with the sun.

The mismatch, drawn as a day

Lay one day flat and the trouble is obvious. Solar output is a hump: zero at dawn, peaking around midday, back to zero at dusk. Demand has a different shape — a dip overnight, a morning bump, then the big one: a tall evening peak as a whole country gets home and switches on at once. The solar peak lands at midday; the demand peak lands hours later, after the sun is gone.

Grid operators have a name for the leftover demand they must cover with other plants: the duck curve. Through a sunny midday, solar floods the grid and that leftover demand sinks into a deep belly. Then the sun drops just as people come home, and it rockets up the duck’s steep neck into the evening peak. Plenty of clean power at noon, when it’s least needed; almost none at eight, when it’s needed most.

Storage is the missing piece

If the surplus and the shortfall happen at different times, the fix is a way to move energy across time. Hold the midday flood and release it into the evening gap. That is storage, and it is the missing piece of the whole transition. The main tools are batteries — giant versions of the one in a phone — and pumped hydro, which uses cheap midday power to pump water uphill, then lets it fall through a turbine after dark. Both bank energy when it’s plentiful and spend it when it’s scarce.

Recall two ideas the course taught. From item 1, energy can be stored as chemical energy in a battery or gravitational energy in lifted water — it never vanishes between midday and evening, just waits in a new costume. And from item 2, a battery has two separate ratings: how much energy it holds (watt-hours) and how fast it can deliver it (watts — its power). Both matter here.

Walk one honest day. A city’s evening peak runs at 4 gigawatts for four hours — 16 gigawatt-hours the sun isn’t supplying. To cover it from storage you need a bank rated for 4 GW of power (to push hard enough to meet the peak) and 16 GWh of energy (to keep pushing the full four hours). Too little power, and you supply the evening but not all at once. Too little energy, and you meet the first hour, then run flat. You need both — the two-number truth from item 2.

That is why this is the hard problem. Storage at that scale is expensive and, today, limited. Other tools help: overbuilding solar so even a cloudy day covers demand, long transmission lines pulling wind from a windy region into a calm one, flexible demand that shifts the dishwasher to noon. But each has limits, and none escapes the core fact: cheap panels alone leave a gap, and something must fill it.

The grid you live on was built around a dial — plants that followed your demand wherever it went. Wind and solar hand you something stranger: power that is nearly free but arrives on the weather’s schedule, not yours. Closing the distance between when clean energy shows up and when you flip the switch is now a central engineering problem of the century — mostly a matter of holding energy still across a few hours of one evening. You sit at the end of that gap every night, between a sun that has set and a peak that hasn’t, depending on whether enough of noon’s surplus was saved for you. Seeing that the trouble is timing, not price, is the honest place to stand before asking what a clean grid really costs.