Constraints On The Electricity Grid

I recently wrote about Moss Landing, the biggest grid battery storage operation in the world. I discovered from talking to a friend recently that most people have no idea what constraints the electricity grid operates under. I think most politicians are the same, and they assume that if we build enough windmills and solar panels then we can live in some sort of eco-nirvana. But that’s not going to happen, and this post will explain why.

An electricity grid is synchronous, in the sense that the 60 cycles per second (or 50 for Europe and many other countries) are synchronized across the whole grid. There is not just one grid in the world, but not many. Every color in the above diagram is a synchronous grid. You can see (for instance in Europe and Africa) that many of the grids cover multiple countries. The US is divided into several separate regions, as you can see in the map below.

For weird historical reasons, Japan is divided into two incompatible grids, with a standard voltage but running at 50Hz in the West and 60Hz in the East. I say “incompatible” but these days it is possible to connect different grids, but it is fairly expensive. The AC has to be turned into DC on one grid, and then back to AC synchronized with the other grid.

Electricity In = Electricity Out

One thing that most people don’t seem to realize is that each grid has to be balanced, in the sense that exactly the same amount of electricity is being produced as is being consumed at any moment in time. Of course, batteries like Moss Landing makes this a little less true, but in the grand scheme of things, their capacity is too little to make any difference to the basic premise. Also, this can apply in both directions. If more power is required, then more power must be delivered. But sometimes (solar in the early afternoon) more power is being created than consumed and so some of it needs to be disabled.

Controlling how much power is produced operates on two different scales. The big scale is that power stations can be turned on and off, or ramped up and down. This doesn’t apply to all power stations—you can’t turn a coal-fired or nuclear power station off on any short timescale. This sort of power station is known as “base load.” There is an inbuilt assumption that the amount of power demanded never drops below all the base load stations together. But we can ramp some gas-turbine power stations. They can be brought online or have their power ramped up significantly. These are known as “peaker” stations and they are the last ones to be turned on and the first to be turned off. That makes their economics a bit sketchy, to say the least.

The smaller way power is controlled is in response to changes in load. If you turn your air-conditioner on, it’s too small to affect the grid. But if it is a hot day and everyone turns their air conditioners on, then more power will be required. All those air conditioners will put more load on the generators in the power stations and cause them to run a tiny bit slower, and so the frequency will start to lag a tiny amount. In response, more fuel is applied to speed them up again (or more water in a hydro plant and so on). It is a bit like the cruise control in your car. When you hit a steep hill, the car slows down, the cruise control notices, and it puts its metaphorical foot on the gas. Of course, it works the other way around when a lot of load is removed. If the difference is too large to be made up this way, then peaker stations need to be turned on too.

If the situation gets extreme, and there is more demand for power than can be supplied, the only thing to be done is to change demand by instituting rolling blackouts. You probably read about the near-total failure of the Texas grid last year. In order to cut demand fast, the grid operators cut power to the fabs among other places, which you probably know use a lot of power. That certainly got noticed in the semiconductor industry since a lot of wafers were lost as a result. But those were the extreme measures necessary to avoid totally powering down the grid and restarting from scratch (without relying on having any electrical power to get the power stations started).

Renewables

Renewables (wind and solar) don’t actually replace any power stations since there needs to be enough capacity in the system to supply all the required power even when the wind is not blowing and the sun is not shining. The way to think about renewables is that they replace some of the fuel that would have been used in the power stations, and the accompanying emissions, but they don’t replace the power stations themselves. Renewables are also largely uncontrollable (not dispatchable, in the jargon). You can increase the power to a gas turbine power plant by increasing the fuel, but wind turbines just produce what they produce, and the same for solar. That’s not strictly true since I think wind turbines can feather (rotate) their blades to capture more or less of the wind, but if there is not enough wind they won’t produce much power. In fact, if there is too much wind they won’t produce much power since they have a maximum operable windspeed too.

The economics of renewables are terrible. First, there is a level of dishonesty in the way renewable power stations are rated. The capacity factor of a wind farm is about 35%. This means that if it is rated for 3GW assuming the perfect wind blows all the time, it will actually produce 1GW on average. Modern solar capacity factors range from 10-30% depending largely on where they are sited.

The second bit of dishonest accounting is that nobody has really addressed the fact that in 10 to 20 years all these wind turbines need to be dismantled and new ones erected. Similarly, for replacing all the solar panels. This will be a huge expense. If you’ve ever been to Palm Springs or the south end of the Big Island of Hawaii (see pic) you will have seen previous generation wind turbines just rusting since nobody wanted to pay to take them down.

But the biggest dirty secret is the need for backup power. You will read that solar is now cheaper than other fuels, and this might even be true. But to do a proper accounting, the backup power needs to be accounted for as part of the cost of the renewable power. As I said above, wind and solar largely only save some fuel, not the capital costs of the power stations that stand behind them.  As we build more and more renewable energy, I’m not sure how the economics of that are going to work. If there is enough solar, for example, there may not be much need for power from the backup stations. But then the backup stations don’t get any revenue, since they get paid for electricity, not for being on standby. If the backup stations are only required, say, once a year for 3 days then they would need to be charging thousands of dollars per KWh to recover the costs of sitting idle and depreciating for the other 362 days of the year.

It’s possible that in the future, new technologies will come along that change the economics. For example, if surplus renewable power is used to create hydrogen, that can be stored and then used to generate electricity when the sun is not shining and the wind is not blowing. But we can’t do that today at grid scale.

It all comes back to the fact that renewables are very variable, out of control of the grid operator. But the golden rule, that electricity being generated has to exactly match electricity being consumed, cannot be broken.

With Diablo Canyon soon to be offline, it remains to be seen if the California grid can cope with a calm day without a lot of sun. Enjoy the rolling blackouts!