How reliable are wind farms?

Modern society is fundamentally dependent on a reliable and on-demand supply of electricity. This electricity comes almost entirely from burning coal and natural gas, fissioning uranium or by large hydro-electric dams. On aggregate, these power plants can be relied on to supply electricity around the clock; a reliability that would seem miraculous to people living only a few centuries ago when light availability was completely dependent on whether the sun shone. Wind farms, however, cannot currently provide this reliability. In fact, on the scale of most countries aggregate wind farm output can be assumed to have almost zero reliability. In this sense, every wind farm must have a fossil fuel power plant sitting in wait for when the wind does not blow.

Let me begin by unpacking a banal statement: All power plants need back up. A century of technical innovation has resulted in electricity grids that are ultra-reliable by any reasonable standard, but power plants still cut off on occasion; they are vastly complex industrial machines and things will sometimes go wrong. When a power plant does goes off-line, others will respond by changing their output. In this sense, all power plants are backed up by each other.

Coal power plant outages, however, are always independent of each other. I live in Scotland, and the probability of Longannett power station – a short drive from where I live – going off-line at exactly the same time as Drax power station – a 3 hour train journey away – is close to zero. The same cannot be said for wind farms.

Anyone who regularly watches weather forecasts knows that wind speeds over large areas, e.g. the whole of Britain, are closely linked. If you are comforted by the claim that “if it is not windy in one place, it will be windy elsewhere”, I suggest you watch a weather forecast.

How much the aggregate output of countries’ wind farms varies can be discovered by looking through spreadsheets produced by grid operators And the lesson is clear: In every country aggregate wind farm output often goes close to zero. I will illustrate this for Britain and Germany.

Wind farms can reliably supply less than 1% of installed capacity

Britain is perhaps the windiest country in Europe; while Germany is more or less the least windy. In 2009, Boccard estimated that the average capacity factor of Germany’s wind farms was 18.3%, while in Britain it was 26.1%. In other words 10 GW (GW = billion watts) of installed capacity in Britain will deliver about 2.6 GW on average, but the figure will be 1.8 GW in Germany. Recent production data in Germany and Britain indicate that these are still reasonable estimates. (Britain’s government publishes annual figures here.)

How much does wind farm output vary in these countries? Let’s look at Germany first.  Last year the power output of Germany’s wind farms peaked at 26 GW at 6 pm on the 5th of December (see technical note for details of calculations). In contrast, minimum power output of Germany’s wind farms was 0.128 GW at 2 pm on the 4th of September. Minimum power output was therefore only 0.5% of maximum power output. Not quite zero, but not much higher either.

Overall, Britain has a much better wind regime than Germany, with higher average wind speeds and fewer lulls. However, it also sees periods of close to zero wind.

Britain’s total wind farm output peaked at 6 GW at midnight on the 21st of December. Its output reached a minimum of 0.025 GW at 11 pm on the 16th of June. The minimum was therefore only 0.4% of the maximum. Britain installed some new capacity between June and September. However, the lesson is reasonably clear; Britain and Germany’s aggregate wind farm output can be expected to go below 1% of total installed capacity with reasonable regularity.

The day of the peak in Britain is also notable for another reason; it shows how much wind farm output can vary in a single day.  By the end of that day wind farm output was 1% of what it was at the start of it. As the graph below shows output went from around 2.5 GW to almost 0 GW in a single day. This is a switch from average output to almost zero output in 24 hours.

 Wind farms should be viewed as fuel savers

The immediate consequence of this is that wind farms cannot be total replacements for fossil fuel or nuclear power plants. If we build wind farms, we need to acknowledge that we will also need conventional power plants to be ready to increase their output when wind farms produce almost no electricity. This will hopefully change with future innovation in energy storage or with the erection of continent sprawling super-grids, but it will likely remain the case for a while to come.

Wind farms, then, should be viewed as “fuel savers”. When they are generating electricity they save fuel, and CO2 emissions, because you need to ramp down a fossil fuel power plant. In other words, they displace electricity generation from fossil fuel power plants, but not the power plants themselves; the power plants will largely still be needed for when it is not windy.

This, of course, does not mean that we should not build wind farms. The benefits that result from the carbon dioxide emissions saved by wind farms are obvious. Similarly, wind farms are among the most economical ways of generating low carbon energy.  However, the role wind farms will play in an energy system should be acknowledged. Anyone advocating the large scale expansion of wind farms must recognise that they will have a large number of fossil fuel power plants on the side. Advocating an expansion of wind farms, while opposing almost all new gas power plants, as some environmentalists do, is either hypocritical or a display of ignorance of basic engineering realities. As the great physicist Richard Feynman said, “For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled.”

Technical notes

1. Wind farm output covered in the above graphs does not cover all wind farms. Some wind farms are not “visible” to the grid, and are not reported in these statistics. Because I am only interested in variation, not absolute numbers, the exclusion of some wind farms should not be material.

2. Data is taken from PF Bach’s website, who has aggregated the data from the German and British grid providers.

3. Calculations were performed using R and plotted using the package ggplot2.

This article is published in collaboration with The Energy Collective. Publication does not imply endorsement of views by the World Economic Forum.

To keep up with Forum:Agenda subscribe to our weekly newsletter.

Author: Robert Wilson is a PhD Student in Mathematical Ecology at the University of Strathclyde.

Image: An offshore wind farm stands in the water near the Danish island of Samso May 19, 2008. REUTERS/Bob Strong.