Nuclear and Wind/solar

These pieces keep getting longer and more technical. This one breaks all the rules. Pity anyone trying to read this on a phone. There’s a PDF here.

Takeaways

1) Nuclear power and wind/solar are mutually exclusive.

2) A reliable grid based only on wind, sun and batteries is not possible.

3) A reliable grid based only on wind, sun and hydrogen may be possible, but it will be impoverishingly expensive.

4) Expensive nuclear is no where good enough.

5) Truly cheap nuclear power is our only salvation.

Nice dumb nukies; mean smart antis

While the strongest supporters of wind and solar often regard nuclear power as anathema, almost all nuclear proponents are careful to make clear that they are not against renewables. Most go further and say we need an all-of-the-above strategy in which nuclear is just part of the solution, often just a fill in for wind/solar's intermittency. The nuclear supporters are not being dishonest. They truly have nothing against wind and solar. In my case, I've been a sailor all my life. I love the wind. I helped my brother, Dave, install what at the time was one of the largest residential PV solar plants on an off-grid island in the Bahamas. Most pro-nukes would be happy with wind/solar if the economics were there.

But the anti-nukes understand a deeper truth. In order for nuclear to be truly successful in dealing with global warming, it must be cheap, as cheap as it was in the 1960's when the real cost was under 3 cents/kWh. And if nuclear is cheap, wind and solar will be relegated to a few globally unimportant niche markets. The grid will be all nuke plus a bit of fossil fuel for peaking and unplanned outages.

Here's the problem. Wind and solar are high CAPEX, effectively zero marginal cost, intermittent sources. Their intermittency requires nearly 100% back up with a dispatchable source. Therefore, in designing your grid, you must compare the fully built up cost of W/S with the fuel cost of whatever fuel it is replacing. From the point of the grid, the only thing you are saving by investing in wind/solar is the cost of the fuel that the W/S replaces, including the social cost of any CO2 that that fuel creates.

Nuclear is a high CAPEX, low marginal cost, dispatchable source. Nuclear's fuel cost is less than 0.5 cents/kWh with a very low CO2 intensity. Wind/solar adds almost no value to a grid in which the dispatchable source is nuclear. Once you've paid for the nuclear capacity, buying wind/solar capacity is a waste. So you don't.

Conversely, nuclear cannot compete with wind and solar on marginal cost. And high CAPEX nuclear cannot survive low capacity factors. If for whatever reason a country decides to invest in a lot of wind/solar, that will force the capacity factor of the dispatchable source down. Nuclear loses out to low capital cost/high marginal cost fossil fuel as the dispatchable power source. If we put enough wind and solar in place, nuclear is dead and fossil fuel lives. The anti-nukes know this well. They brag about ``death by capacity factor". Fossil fuel interests know this very well too. They just let their pawns in the anti-nuke movements do the public bragging.

Nuclear and wind/solar are not only not complementary. They are in direct competition. Anti-nukes recognize this. Pro-nukes apparently do not. Let's take Germany as an example.

Germany: A Case Study

The Gordian Knot Group (GKG) has created an electricity grid model which, given the user's assumptions, comes up with optimal combinations of market grid cost and CO2 emissions, for supplying a given hourly demand profile. The model is described in reference 1¹. The GKG model distinguishes between fixed CO2 (mainly the CO2 embedded in the plant material and construction) and variable CO2 (mainly the CO2 created by the production and consumption of fuel). Wind and solar CO2 is almost all fixed. This means as wind/solar capacity factors go down, the CO2 per marketed kWh go up.

We have exercised this model on Germany. Germany was chosen because excellent hourly load, wind, and solar data exist for this country. But the GKG model could be applied to any market for which this data is available.

For any given set of parameters, the model minimizes the sum of the wholesale cost of supplying the electricity --- the value of the resources consumed in generating the power --- plus the social cost of the resulting CO2 emissions, at a user supplied CO2 price. Ideally the latter should be the cost to society of emitting a ton of CO2 on the margin. No one knows what that is. So we run the model for CO2 prices, ranging from zero to $1600/ton CO2. This exposes the market cost versus CO2 trade off Germany faces. Given the ridiculously large uncertainty on the cost of nuclear, we also ran a range of nuclear overnight CAPEXes, running from what nuclear ideally should-cost, about $2000/kW, to prohibitively expensive ($32,000/kW).

This set of runs was for the 8 year period, 1993 to 2000 inclusive. We use the ENTSO hourly demand and wind and solar capacity factors. We assumed Germany was a greenfield. The path we followed took us from a pure wind and solar grid to an all of the above grid, progressively adding source and storage technologies as we went. Table 1 summarizes the results for our base case parameters.

How did the model come up with these numbers? It is worth going into the solutions in some detail. Let's concentrate on the far right column, where we assume Germany is willing to pay $1600 to reduce CO2 emissions by a ton. This is higher than any estimate of the social cost of CO2 that I have seen in the literature.

Only Wind, Sun, and Batteries

Figure 1 shows what happens if Germany allows herself only wind, sun and batteries. The grid cost is a crippling $574/MWh, and the CO2 intensity is over 150 grams per kWh. The problem is that batteries are so expensive that the least wasteful option is to accept a massive amount of curtailment in order to reduce the amount of battery capacity needed. Curtailed power is over three times marketed power. This drives the wind/sun capacity factors below 5% and the embedded wind/sun CO2 per kilowatt-hour up.

Figure 1. Only Wind, Sun and Batteries. Numbers above bars are capacity factors.

The installed wind/solar capacities required are preposterous. Despite about 30 hours of batteries, the total wind and solar nameplate capacity is 20 times the average demand. Onshore wind is 10 times the current installed capacity. New onshore wind additions in Germany have slowed to trickle due to local opposition to the massive turbines. Offshore wind is 15 times the current installed capacity. Germany simply does not have enough offshore to install this much wind, certainly not at the cost and capacity factors we have assumed. Even Germany cannot implement this grid. This is good. It would be both an economic and environmental disaster if she could.

Only Wind, Sun, and Green Hydrogen

Germany is blessed with ample salt domes. Germany can use excess wind/solar electricity to separate water into hydrogen and oxygen via electrolysis, and store this hydrogen in these domes. The hydrogen can be burned when it is needed. The round trip efficiency is a putrid 37%; but just about anything has to be better than batteries. Figure 2 shows the results for a grid made up of wind, sun and hydrogen in salt caverns, under optimistic assumptions about hydrogen production.

Figure 2. Only Wind, Sun and Hydrogen. Vertical axes different from Figure 1.

A big improvement. The grid cost is down to about 200 USD per MWh, and the carbon intensity is about 27 gCO2/kWh. However, the wholesale cost is four times that of a fossil fuel fired grid. The carbon intensity is not exactly net zero. Despite about 30 days of H2 storage, we still need an immense amount of wind and solar capacity, about 7 times average demand. The onshore wind required is about 6 times the current installed capacity of 57 GW. Germany is having difficulty increasing the current capacity by more that a GW or 2 a year, And we have not even started to electrify non-grid markets.

If you are a really wealthy country with a lot of salt domes willing to deindustrialize, green hydrogen might work. For the rest of us, it is a non-starter.

All sources, $8000/kW nuclear

Figure 3 shows the results if Germany allows herself all the options, but nuclear's overnight CAPEX is an exorbitant $8000/kW. The very high social cost of CO2 dissuades the model from using any coal.

Figure 3. All sources. Nuke overnight CAPEX, $8000/kW. Vertical axes changed again.

Almost all the power is from nuclear, which is doing some load following. The 77% capacity factor is considerably less than the 90% availability. We've gone from all wind/solar to almost all nuclear. However, nuclear is so expensive, it pays Germany to install some onshore wind and 21 GW of gas. There is no curtailment of the wind; but the gas is used only for a tiny bit of peaking. The grid cost is a very expensive $127/MWh. The CO2 intensity is an impressive 8 grams/kWh. Germany might be able to live with this for her current electricity. But electrification of non-grid markets is not going to happen, and the power is far too expensive for the developing world. Expensive nuclear is no where good enough.

All sources, $4000/kW nuclear

If nuclear CAPEX is $4000/kW, Figure 4, the program installs no wind or solar. It installs 22 GW's of gas which is used very sparingly. Surprisingly, the program opts for a tiny bit of batteries which are used over and on a diurnal basis. It even finds a niche for a bit of hydrogen. However, if the batteries and hydrogen were eliminated, the overall results would change hardly at all.

Figure 4. All sources. Nuke overnight CAPEX, $4000/kW

The CO2 intensity is down to a rock bottom 6 g/kWh; but grid cost is pricey at $82/MWh, Almost all of the CO2 is from nuclear not gas. If Germany outlaws gas in this situation, the CO2 intensity would go UP, since the marginal nuclear capacity factor would be around 2%. This is a scenario Germany could probably live with. It is far superior to her current situation. But the electricity is still too expensive for both the developing world and deep decarbonization. Expensive nuclear is not good enough.

All sources, $2000/kW nuclear

If nuclear approaches its should-cost of $2000/kW, the model comes up with Figure 5. The optimal solution is very similar to Figure 4. The extremely important difference is the grid cost is now down to an affordable $56/MWh. This is a solution that could scale to handling decarbonization; and, critically for developing countries, it is more than competitive with coal

Figure 5. All sources. Nuke overnight CAPEX, $2000/kW

It is interesting that the model continues to install quite a bit of gas, even though it uses it to produce only 0.1% of the electricity. This means we almost always have spare capacity to handle unplanned outages and unexpected demand spikes.

Purists will complain we are still burning some fossil fuel. But if Germany outlawed fossil fuel, the cost of the electricity would go up. The goal is not to reduce grid related CO2. The goal is to reduce all CO2. To do that we need to produce electricity as cheaply as we can.

Only cheap nuclear can solve the Gordian Knot

Figure 6 summarizes our results for a range of CO2 social costs. Only truly cheap nuclear offers humanity what it must have: both cheap electricity and low CO2 emissions. Expensive nuclear offers humanity the choice of impoverishment or global warming. As you move into the page in Figure 6, nuclear becomes more expensive, and the choice becomes more stark.

Figure 6. Best CO2 intensities and Grid LCOE's, all sources, Germany

The fact that the GKG would bother to run nuclear costs that differ by a factor of 16 is preposterous. If a nuclear overnight CAPEX of $2000/kW is possible, as the South Koreans, the Chinese, and the GKG claim, why in the world would you run $4000/kW, let alone anything higher?

The answer is we've seen nuclear CAPEXes of $8000/kW and higher. Vogtle 3/4 is above $10,000/kW. Flamanville 3 is in the same range. This cannot happen in a properly functioning, competitive market. In such a market, there is only one price, the best price. If nuclear cost is as critically important to the planet as Figure 6 claims, we must figure out what turns a $2000/kW plant into a $10,000/kW plant and eliminate it. The GKG publication Why Nuclear Power has been a Flop attempts to do just that.²

1. Devanney, J. The GKG Grid Model, downloadable from gordianknotbook.com. The paper , Low Carbon Electricity: the Options for Germany has a much better description of the model and a far more complete summary of the results for Germany

2. Devanney, J. Why Nuclear Power has been a Flop, downloadable from gordianknotbook.com.