Summary and Policy Implications
The “capacity credit” of GB wind power has been quantified for the nine most recent winters in terms of the increase due to wind power of the Minimum Reserve Capacity. Wind power data for these nine past winters was scaled-up to 2016/17 levels, to give an indication of what might happen in the coming winter if there is an exact repeat of the weather of previous winters. Increases in Minimum Reserve Capacity due to wind tend to cluster around 1 GW, a small figure relative to the total installed capacity of around 15 GW (metered plus embedded wind).
In addition, the 2016/17 level wind power was scaled in steps between zero and four times, to show how the influence of wind power on Minimum Reserve Capacity might grow in the future. The rate of growth of Minimum Reserve Capacity from 2016/17 levels is found to be very low, typically a doubling of the 2016/17 wind capacity of 15 GW would only increase Minimum Reserve Capacity by of the order of 500 MW. This result reflects the fact that GB wind power has already achieved its “quick-win”, reducing cold breezy demand peaks to below the level of cold calm demand peaks, the latter are now responsible for the slow rate of growth of Minimum Reserve Capacity.
The results shown in this post establish that the debate about wind power on cold calm days is now well past its sell-by date. There is no doubt that the worst case demands on conventional generators in 2016/17 come from cold calm days, wind power reducing those from cold windy days to much lower levels. If the GB electricity system is stressed this winter, the only way to solve that problem would be to add more dispatchable capacity (and retain the existing capacity), adding more wind power would make very little difference.
Since firm dispactable power levels are falling faster than minimum wind reserve capacity is growing there is a danger that GB will soon be the first major economy to inflict upon itself power cuts on cold calm working days in winter.
Winter weather in the UK alternates between (relatively lengthy) mild and wet periods with westerly winds, and (relatively brief) cold and dry spells with Easterly/Northerly winds, the latter type of weather resulting in the highest demands for electricity, peaking in the early evening on working days. Blizzards, snow-drifts and wind-chill sometimes feature in weather reports, indicating cold days with strong winds, but sometimes cold weather has light winds, and many examples have been shown on this blog and elsewhere of very low GB wind power at peak demand time on cold winter days.
If your job was to provide on every day the peak electricity demands shown in red on the following figure, how much reliance would you place on the wind power shown in blue?
This post attempts to shed some light on that question. I have quantified wind power at peak demand times for the previous nine winters by computing the contribution of wind power to Minimum Reserve Capacity. Note that the wind power measured or estimated for previous winters is scaled up to 2016/17 levels, so that the results indicate what would happen in 2016/17 if there is an exact repeat of the weather of previous winters.
Minimum Reserve Capacity
Suppose that there is a steady amount of non-wind power available, greater than the highest demand, add the wind power, subtract the consumption and thereby obtain how much reserve capacity there is at peak demand times for all days between November 1st and the end of February. The quantity of interest for keeping the lights on every day is the MINIMUM reserve capacity, relative to that found without wind power, and this has been computed for a set of wind power capacities that vary between zero and four times its value in 2016/17:
In the figure above the wind power capacity is the total, the sum of metered and embedded wind, which has a value of around 15 GW in 2016/17. These nine samples of the random distribution of the contribution of wind power to reserve capacity has a cluster of cold winter values around 800 MW in 2016/17, with one cold winter sample as low as 600 MW.
Note that the contribution of wind power to Minimum Reserve Capacity should not be regarded as a percentage of the wind capacity, because the figure above shows that the contribution does NOT scale linearly with increasing wind capacity. Most of the profiles in the figure above have a relatively rapid rise from zero wind capacity, but then flatten-off to much lower slopes. What is happening is that most of the previous nine winters had demands on cold breezy days that were somewhat higher than the demands on cold calm days. Increasing wind power from zero initially reduces the cold breezy demands down to (and then below) the cold calm demands, at which point the gain in Minimum Reserve Capacity is determined by the much lower wind speeds of the cold calm days.
A striking feature of the figure above is the slow growth of Minimum Reserve Capacity, even doubling the 2016/17 wind capacity of 15 GW would typically increase reserve capacity by only several hundred MW. To get some insight into this feature the remainder of this post shows in detail the example of the winter of 2011/12, the most recent winter that had a proper cold spell.
Winter 2011/12 in detail
The following figure shows daily peak consumption (demodulated, see below for an explanation), and metered wind (measured) and embedded wind (estimated by National Grid), scaled-up to 2016/17 levels, for winter 2011/12:
The figure above shows a cold spell in February 2012, with considerably elevated consumption of electricity, and variable winds with a high degree of correlation between metered and embedded generators. Electricity “consumption” is defined here as the sum of “demand” and embedded wind, data for both having been downloaded from a National Grid website (see the NATIONAL GRID DATA page of this blog).
The critical cold days of interest have an annoying habit of sometimes falling on weekends and during the Christmas holiday period, giving much lower demands than if they had fallen on working days. To deal with that issue the consumption data has been “demodulated”, by temperature-weighted interpolation across weekends (including Fridays) and holidays. The following figure shows the original (red dots) and demodulated (solid red) consumption data, together with the mean of the Central England Temperature (HadCET), the variable used to perform the interpolation:
The Minimum Reserve Capacity with wind power is computed by assuming an arbitrary 62 GW of steady supply, adding the wind power, and subtracting the consumption. The same calculation is performed without any wind power, with the results of both calculations shown in the following figure:
The figure above shows that on most days the 2016/17 level wind power would give substantial rises in reserve capacity, but not on the cold day of 11th February 2012, a day of very light winds, as confirmed by the relative absence of isobars crossing GB on the following synoptic chart for that day:
The slow growth with wind power of Minimum Reserve Capacity for the winter weather of 2011/12 is dictated by the very low wind speeds of 11th February 2012, a fact that contradicts what appears to be a widespread belief that dispactable power can be retired and replaced by more wind power.
Conclusions: See the summary at the start of this post.