How to Define Nuclear Success

Donald Trump says coal is clean and beautiful, and on April 8 he signed an executive order to encourage its use. Whether or not you share his taste in fuels, the order appears to overlook some critical details of how the power system works, details that nuclear advocates should bear in mind when they talk about what will constitute a “renaissance.” The lesson is that government policy can help identify political priorities, but when policy conflicts with the market, the market will win.

Coal has some attractive attributes. It’s plentiful, here and abroad, its price is fairly stable, and a plant manager can look out the office window and see where the next few months of fuel are coming from. In contrast, natural gas is delivered just in time, wind and sun come when they want to, and even water for hydroelectric dams is variable.

But, the amount of coal we burn does not depend on how much we can mine or how many power plants we have to burn it in. It depends mostly on its cost per megawatt-hour of electricity, an area where it has been overtaken by natural gas. There have been substantial improvements in coal mining, and there are potential improvements in coal plant thermal efficiency, but they cannot match natural gas.

Nuclear energy is feeding into the same market. And depending on whether the market is traditionally regulated, with ratepayers compensating utilities for the cost of utility investments, or is “deregulated,” and converted to an auction system, success will come when a company that wants to sell electricity in a regulated market thinks that regulators will see nuclear as a reasonable choice for consumers, or, in a market system, when the price of production from the reactor is going to be lower than the average price on the wholesale market. The people who decide what kind of plant to build are driven by market signals.

It Won’t Work for Oil, Either

A corollary applies to the White House’s ambitions for oil. Just as leasing federal land to coal mines won't mean more coal production, opening up land to oil drilling won't increase oil production. That will depend on the price per barrel and the expectation of future prices compared to the price of production. Opening land could lower the cost of producing coal or oil. Firms must determine whether the projected decrease is enough to pursue investments and how those costs will measure against future fuel prices. If opening the land has any effect, it would not start for years to come.

There are too many moving parts in such a system; the government cannot possibly align all of the markets and incentives. This could only work if the government were both the buyer and seller of the entire coal supply chain from mines to purification to power generation to electricity consumption. Price signals matter. Price signals tell you the status of supply and demand without much delay or high costs of communication. Right now, price signals are telling us that coal is not valuable relative to other means of energy production.

In that context, Trump’s executive orders are a bit like building new runways that the airports will need on the day that pigs fly. One order seeks to open up more federal land for coal mining, but mines cost tens of millions or hundreds of millions of dollars to build, investments that would be stranded if the market for coal continues to decline. Another allows coal plants to delay retirement if their energy is needed, and it seems possible that a little bit of extra coal will be burned in the immediate future to meet an anticipated rise in demand. But there are cheaper ways to meet demand.

The big hurdle for coal, and for nuclear, is changes in natural gas. Fracking and combined-cycle gas plants have attacked the cost problem from both ends. Oddly, the growth of wind and solar on the grid has helped natural gas as well. Here’s how it works:

Fracking is a decades-old technology used in oil drilling, fracturing the rock to let the hydrocarbons flow to the well. But the natural gas industry began early in this century, using it on a type of rock that holds natural gas previously considered unrecoverable. And the industry has learned how to drill holes that twist from vertical to horizontal, to reach underground structures that hold gas. Annual gas production has more than doubled since then, to more than 40 trillion cubic feet, and increasing supplies have caused prices to collapse.

And a unit of natural gas goes farther than it used to at the power plant. Gas plants used to make electricity by using gas to boil water, and then running the steam through a turbine to create mechanical power; a generator turned that into electricity.

But now, utilities use a machine that resembles a jet engine, bolted to the ground, to create mechanical power. Then the exhaust from the jet is used to boil water, to make steam to make mechanical power. The older method made electricity with between 35 percent and 42 percent of the energy value of the natural gas. The newer method makes electricity with an energy value of up to 64 percent of the natural gas input. In other words, today, a utility can make about 50 percent more electricity with a unit of natural gas than it could with the kind of plant that was common in the 1970s and 80s.

Solar and wind helped natural gas because those generators are nimble. A grid with a lot of solar and wind, which can arrive and depart quickly, needs some dispatchable generation to compensate. Of the other assets on the grid, nuclear and coal move slowly. Batteries move faster, but their capacity is quite small relative to swings in production. Hydro, when available, can also help, but utilities have countered the volatility of wind and solar by relying on natural gas generators.

Don’t Argue with a Computer

But the White House does not appear to have accounted for how these factors translate into energy generation. A market system determines which power plants run at any given moment and how many hours a year each power plant will run. Working backward from that determination, utilities decide what to build and what to retire. Calling coal “beautiful” doesn’t change that. Neither does being attracted to nuclear because it is reliable, an aid to national security, or energy-dense.

In most of the United States, a computer decides, every hour or at shorter intervals, how much electricity is likely to be needed. And it ranks all the available generators, according to the bids they have submitted, from low to high. If a region needs 5,000 megawatts at 2 pm, the computer will go down the list until it has secured the five thousandth megawatt, and order those plants to run. The last megawatt ordered is the most expensive, and whatever that price is, that’s what all the generators are paid.

If that last increment costs $40 but the coal plant’s cost to operate is $45, the coal plant will not be ordered to run.

In fact, because the plant has limited operating flexibility, it may decide to bid in at $39 so that it is chosen, and to operate at a loss during some hours, because on average it will make money. Or the operators may shut the plant during mild weather, and only start it up during the cold or hot months, when demand is higher and they can make money. Letting the industry dig mines on additional federal land won’t change their calculus.

If plants are needed in continued operation, because of their location in a load pocket or because of the unreliability of natural gas supplies, the market will pay them a premium so that they remain profitable. That isn’t part of executive orders either.

In non-market areas, which are traditionally regulated, utilities will still dispatch their generators in price order, so that at any moment, the least costly combination of plants is serving load. If they see that coal plants are running very seldom and that their labor and maintenance costs are being spread over a very small number of hours of operation, they will apply to the public service commission to retire those plants.

How Nuclear Fits In

Most power reactors now in service have very limited flexibility. They will bid in at a low cost and, at some point, accept revenues that are below their operating cost, hoping to make money on average. When the price of natural gas was exceptionally low, several were retired, including Palisades in Michigan, Duane Arnold in Iowa, and Three Mile Island 1 in Pennsylvania, all of which the operators are now seeking to re-open.

For new nuclear, would-be builders face a complicated decision. If they think it will take ten years to get a plant online, the question is what the cost per megawatt-hour will be in, say, 2035, which is largely a function of construction cost, and what the average wholesale price of a megawatt-hour will be at that point, and in the early years of the plant’s lifetime. Market price will be important for the whole lifetime of the plant, which could be 60 or 80 years, but the first few years are far more important from a financial viewpoint.

For a plant in a traditionally regulated region, a builder has to convince the regulators that when the plant comes online, it is likely to be of more benefit to the consumers than an alternative would have been. The alternative could be gas, coal whose retirement was delayed, hydro, solar, and wind plus batteries, or money spent to reduce demand by improving end-use efficiency.

For a “merchant” plant, built in a market region, the builder needs to convince investors that the plant’s costs will be lower than market prices.

Clean energy tax credits, investment tax credits, and state mandates to reduce carbon emissions can all play a role. So can rationalizing regulation at the Nuclear Regulatory Commission, because it would reduce costs.

But the renaissance will dawn when a rational actor will calculate that a reactor—large, small or in between—is a better bet than any of the alternatives. That will move the technology beyond demonstration projects and back into the commercial mainstream.