The road to zero emissions must embrace nuclear power

Those unreliable renewables

A few weeks ago, I went sailing with a couple of friends. It was a beautiful day, winds were blowing steadily from the shore at about 20 knots (that’s about 23 mph or 37 kph for those who are unfamiliar with nautical measurements). The weather forecast was clear and sunny with a chance of thunderstorms in the late afternoon. Wind and solar under those conditions would have been generating a steady supply of electricity, wind at around 70% of its design capacity and solar close to 100%.

Around mid-day, there was an update to the weather forecast. A violent storm had formed and was approaching quickly, it would arrive in about 40 minutes. We were 30 minutes from the harbour, so we turned on the engine and with both diesel and sail power we headed back at maximum speed, stopping only to take down the sails as we approached the shore. The sky was black, and the wind had strengthened. A wind farm would have been at full design capacity, but solar generation would have been close to zero.

We made it to the harbour, but there was no way we could dock in those winds, so we had to motor away from the land and ride out the storm. A wind turbine would have rotated its blades to reduce the exposed area, and power generation would have been reduced. At the storm’s peak, the turbines would have shut off to protect the generating equipment.

The storm only lasted about 20 minutes, then the winds slowly dropped, and we were able to make our way back to shore. By the time we docked the wind had died to almost nothing. In the space of less than an hour a wind farm would have gone from a steady 70% production to full power, then to zero as the turbines were shut down, returning to full power as the storm started to subside, and then to zero once more in the lull behind the storm. A solar farm would have gone from full power to almost nothing as thick, black clouds covered the sky and then back to full power as the storm passed.

Storms like that are rare but wind conditions are always changing, and forecasts are rarely accurate. It is a challenge for a sailor, and it is also a challenge for the operators of our electricity supply.

The power grid is a complex beast, a mix of generators, distribution systems and users that must always be kept in balance. When you turn on your light, a power station somewhere on the grid must crank out a few extra watts to supply energy to your light bulb.

A grid with wind and solar power must react to unpredictable variations in supply as well as variations in demand, and as more renewables are added to the grid balancing the grid becomes more difficult and more costly.

The real cost of wind and solar

No matter how many wind turbines and solar panels are built there are always times when the wind isn’t blowing and the sun isn’t shining, so renewables must be backed up by an energy source that can be cranked up whenever the renewables fail.

To make sure that electricity is available when renewables aren’t producing, the grid operator pays power plants to be available even if they are not generating power. Without those “capacity fees”, the fossil fuel plants would go out of business and there would be no electricity when the wind isn’t blowing and the sun isn’t shining.

But even when they are operating, the fossil fuel plants may be forced to operate below their peak efficiency because they must constantly turn up and down to follow the vagaries of the wind and sun. A study of gas turbines in Ireland showed that combined cycle turbines following the wind achieved only 32% efficiency versus a design efficiency of 55%. That’s an extra 70% gas consumption per kWh. The result is increased cost and increased emissions.

Frequent stops and starts also shorten the life of the power plant increasing costs. Nitrous oxide (NOx) emissions are also more difficult to control when a power plant is operating at a reduced load, and gas plants struggle to stay below their NOx emission limits when operating at low power.

A grid with a high proportion of renewables must also have backup generators that can respond quickly to variations in supply. That means a higher proportion of less efficient simple cycle turbines and higher gas consumption. But for grid balancing, even simple cycle turbines cannot react fast enough from a cold start, so to provide reserve power, some turbines are kept on hot standby, a procedure known as “spinning”, which uses fuel without producing any power.

Natural gas is delivered to the power station on a “just in time” basis, there is no storage at the power station. The power station operator does not know ahead of time how much gas will be needed because he doesn’t know how much power will be provided by the wind, so most of the gas must be bought on the spot market. A cold snap with low wind forces the power station operator to buy gas at a time when demand from other users (home heating, for example) is at its highest, so we see enormous short-term spikes in gas prices in winter if the wind fails.

The capital, operation and maintenance cost of two power systems, the need to operate at lower efficiencies, the extra stop/starts, grid balancing, spinning, the use of less efficient turbines and the need to buy gas on the spot market at times of high demand all add costs to the operation of the grid. Those costs will increase as you add more intermittent power sources to the grid, but they are never shown in the published costs of wind and solar power.

Those extra costs are the reason why countries with high installed wind and solar capacity tend to have high electricity prices, and why your electricity bill will always go up when so-called “low cost” wind and solar are added to the grid. Higher installed wind and solar capacity correlate with higher prices, as shown in the chart below:

As more renewables are added to the grid, costs increase

As more wind turbines are added, capacity reaches the point where power must be curtailed on windy days. The chart below has been developed from 35 years of wind data from the UK, it is a simulation of wind power generation based on the UK’s current mix of onshore and offshore wind turbines, it shows what happens as more wind is added to the system. Real conditions are actually worse than the chart shows because I have not accounted for the need to curtail wind to prevent local overload of transmission systems, but for simplicity, let’s assume that all of the generated power can get to an end-user.

The x-axis is the installed nominal wind turbine capacity as a percentage of average demand.

The green line on the chart shows the percentage of power that can be generated by wind. It’s a straight line up to about 40%, but after 40% there are times when the generated power exceeds the demand, and the power must be curtailed – the orange line on the chart shows the curtailment.

As you add more turbines, the amount of curtailment goes up, and the useful capacity of each added turbine goes down – that’s the blue line on the chart.

In a system with an average demand of 1GW, the first GW of wind power gets the system to the point where about 40% of the generated energy can be provided by wind, the second GW only adds 27% and the third adds only 11%. The returns quickly diminish, and you can never reach “Net Zero”. The backup gas plants are always required and the more wind you add, the more expensive the system becomes.

Solar has a similar curve, when you reach the point where solar must be curtailed during the day, the costs start to rise.

Many countries, or states, have set targets of 60% or more renewable power generation from wind and solar by the end of this decade, without calculating the costs or having any concept of the practicalities of implementing such targets.

Some are aiming for 100% by 2050, hoping that some as yet unknown or untested technology will emerge to allow surplus energy to be stored when generated and re-used when needed.

How much storage is needed?

Energy storage sounds wonderful in theory, but scientists have been seeking a way to store electrical energy ever since the first power plant went into production. The most popular method is pumped hydro which is primarily used to compensate for differences in the night-time and daytime demand. Lithium-ion batteries are finding uses as short-term storage for grid balancing and peak load shifting. One day, batteries might be cheap and powerful enough to equalize the intra-day supply and demand from solar power systems. But neither of those solutions is a viable means of storage for a power grid that includes significant wind power.

The vagaries of wind mean that vastly more storage is needed to balance supply and demand. The daily fluctuations are almost irrelevant, and even the seasonal variations can be mitigated by finding the right mix of wind and solar.

Continuing with the UK as an example, the winds are normally stronger in winter when demand is at its peak, and weak in summer. Higher solar generation in summer can offset the lower wind generation and minimize the seasonal variations.

However, the storage requirements for backing up such a system are not fixed by the daily or seasonal imbalances. It is the multi-year variations in wind speeds that determine the required storage capacity. To evaluate storage requirements, using only one or two years of data will not provide the answer, it is necessary to examine wind records over a long period.

A publication by two German scientists, Quist and Runhau, evaluated the storage needs for the German grid running on 100% renewables and concluded that the optimum cost system would need enough storage to provide 36 Terawatt-hours of electricity. To put that into perspective, for readers who might think that solution lies with li-ion batteries, that amount of storage would absorb 250 years of the world’s total 2021 battery production, just for one country. Even if such large batteries could be built, they would not work because batteries slowly lose their charge, and the backup storage for wind and solar needs to remain in place for years, maybe decades. Quite clearly, batteries are not going to do the job, that’s why the referenced publication and my own analysis use hydrogen, stored in salt caverns, as the method of energy storage.

Making “green” hydrogen by electrolysis is an established technology, but it is not commonly used because it is more expensive than hydrogen made from natural gas. Proponents of renewable hydrogen claim that it can be cheap if the electricity for the process uses surplus renewables, but that argument is flawed. If a wind turbine must return 20 cents/kWh to be viable and half of its electricity goes to make hydrogen for free, then the people who are paying for the other half must pay 40 cents. The energy isn’t free, the cost is simply moved from one pocket to another.

The only way to properly determine the true cost of renewables is to look at the whole system as a single entity.

Hydrogen generation and storage add significant costs. Producing hydrogen from water, storing the hydrogen as a compressed gas, and regenerating electricity on demand from that hydrogen is going to have a round trip efficiency of about 35%, or possibly less. It will take about 3 kWh of input electricity to produce 1 kWh of output.

The hydrolyzers and all the associated equipment must operate using surplus renewables. I am not aware of any system that has tried to do that. There are hydrolyzer plants like this one at Bouin in France that make hydrogen from wind power. But that plant is a small load connected to a large wind farm which in turn is connected to the grid. Somewhere on that same grid, there are other generators (nuclear, coal or gas) that make up the grid shortfall when winds are light, so the hydrolyzer can operate without restrictions. A hydrogen plant using only surplus renewables will have frequent stops and starts and will operate most of the time at well below its design capacity and peak efficiency. This pushes up the cost.

Similarly, all the problems inherent in operating gas turbines when following the ups and downs of the wind are also present in whatever equipment is used to generate electricity from hydrogen, whether it’s a turbine or a fuel cell. All those inefficiencies must be considered in the system cost.

When determining the total system requirements, there is a balance between the installed generating capacity and the amount of storage that is needed. More generating capacity means lower storage needs and vice versa. Too much generating power adds unnecessary cost, as does too much storage. The cost curve is U-shaped, with a “sweet spot” where costs are minimized.

We can find the lowest cost system by calculating the cost of electricity under a range of scenarios and adjusting the mix of wind, solar and storage until we find the optimum solution. That’s what Qvist and Ruhnau did in their analysis of the German grid, and it’s what I have done in a simplified manner using the UK wind and solar data. My analysis is not a simulation of the UK grid, it does not have imports and exports or biomass. It is a simulation of a grid that uses wind, solar and hydrogen and has weather and power demand similar to the UK. It could apply to many countries that have similar wind and sunshine profiles.

The all-renewables case needs massive storage, the simulation shows a need for 31.5 Terawatt-hours, equivalent to 38 days of average demand. It is not the daily variations or the seasonal variations that determine the storage needs, it is the multi-year variations in the wind. Below is the chart of the stored capacity (the amount of electricity that can be generated from the stored hydrogen). The analysis used 35 years of weather data with the storage set to be equal at the beginning and end of the 35 year period.

Take a look at the period from 1998 to the end of 2008. It is a period when winds were generally higher than usual, and the storage caverns are filling. Then we hit the winter of 2009, winds are lighter than average, and the storage is drawn down. It recovers at the end of the year but in 2010 there is a wind drought in winter, at a time of high demand when there is very little solar. Winds are below average again in the winter of 2011 and by February the storage caverns are almost empty. A decade of hydrogen storage is gone in a little over two years of lighter than usual winds. This is not a reliable system, there is no guarantee that another year of light winds won’t happen when the storage caverns are empty.

Comparing the cost versus nuclear

The “all renewables” option is also a very expensive system. Using cost data from a range of sources including the IEA publication “Projected costs of generating electricity – 2020 edition”, and IRENA publication “Hydrogen - A Renewable Energy Perspective”, I have estimated the cost of generating “zero-emission” electricity under five different scenarios:

·         Wind and solar with hydrogen storage

·         Wind and solar with a baseload of nuclear power, hydrogen storage

·         Wind and solar with a nuclear equal to average demand, hydrogen storage

·         Nuclear with hydrogen storage to level out the demand

·         Nuclear only

A real grid would have other power sources, maybe some pumped hydro, biomass and some battery storage for load balancing, but the bulk of the costs will be the primary power sources. For simplicity, I have included only the primary sources.

The results are presented in the chart below:

The “all renewables” option is by far the most expensive. Users are basically paying for three complete generating systems. There are wind farms for when the sun isn’t shining, solar for when the wind isn’t blowing and hydrogen for when neither are producing. Adding further to the cost is an inefficient hydrogen production and storage system.

Including a nuclear baseload reduces costs by about one-third and adding more nuclear reduces costs further. The lowest cost option is a system that uses nuclear power with a relatively small amount of storage for load levelling. It is important to note that wind and solar serve no useful purpose in such a system, they only make the system more complex and make the nuclear plant less efficient and more costly.

Nuclear also has the smallest land footprint and does not depend so much on the supply of critical materials.

This analysis clearly comes out in favour of nuclear over renewables. However, it is based on one location (UK) where solar energy is weak and peak electricity demand happens in winter.

If you make the same analysis for Australia or the southwest USA where solar power achieves much higher capacity factors and peak demand happens in summer, there may be a case for including solar in the mix.

I have seen analyses that project much lower costs for renewables in 2050, based on declining costs over the past decade being projected over future decades.  I think that is a dangerous approach. Both wind and solar consume critical materials that are in limited supply. As we are seeing with both li-ion batteries and solar panels, the rising costs of raw materials may well offset any improvements in manufacturing methods or technology, forcing prices up instead of down. In any case, those same lower-cost arguments could equally be applied to nuclear, where technology and a more efficient supply chain have at least an equal chance of reducing final costs.

If a zero-emission power grid is an ultimate aim, most locations in the world would benefit from initiating a nuclear energy program as soon as possible, rather than spending time and money building renewables. Solar or wind generators that currently exist will be at the end of their useful working life in 25 years or less, they should be replaced by nuclear.

In a system with a mix of wind, solar and nuclear, the nuclear must be designed to provide 100% of the peak demand because the availability of wind and solar can never be guaranteed. In such a system, wind and solar serve no useful purpose.