Is the growth of PV and wind making it more difficult to manage the UK electricity system and ensure that supply matches demand? Many people think so. In this article I look at one piece of contrary evidence that suggests that balancing the electricity grid was no more demanding this summer than last year, despite the huge growth in solar power.
This summer actually saw a sharp fall in the number of times coal and gas power plants had to sharply adjust their output to balance the varying output of intermittent renewables. If solar and wind growth had been causing problems balancing the electricity grid we would have expected the reverse. This is just one piece of data in a very complex area, but it is very good news for renewables.
The growth of PV in the last year (and, to a much lesser extent, unmeasured small scale wind power) has reduced the demand for electricity generation over the sunniest portion of the year. I’ve looked that the three month period from 9th May to 8th August and compared this year and last. All my data is taking from the Elexon portal.
This year, average electricity demand peaked at around 33.3 GW (33,300 MW) at 18.00 during this 92 day period, over two and a half GW, or about 8%, lower than in 2014. Some of that difference arises from the gradual fall in overall electricity use. Much comes from the striking jump in PV production this year.
The chart below shows the amount of electricity being produced by fossil fuel plants, big wind farms, biomass, hydro, imports and pumped storage every half hour (1-48) for the 3 month period from May 9th to August 8th. It excludes PV and small wind because these outputs are not measured centrally and are seen as net reductions in the electricity generation required by the major generators.
This chart demonstrates that – on average – the amount of variation in electricity production over the course of the day is actually lower than it was last year. The typical peak is just over 10 GW higher than the half hour of lowest demand whereas in 2014 the average daily variation was over 12 GW. Everything else being equal, this would make the UK electricity network easier to manage because the need to ramp up and down gas and coal plants will be less.
Looking specifically at fossil fuel plants, their electricity production was substantially lower in 2015 than last year. As we’d expect, fossil fuel plants have born the full reduction in demand for electricity.
Nothing surprising thus far. However two of the readers of this blog have written emails suggesting that National Grid has been finding it much harder to manage the stability of the UK network this summer, perhaps as a result of the unexpectedly large addition to PV capacity. Solar is, of course, highly variable during each 24 hour period and is also somewhat unpredictable (particularly for non-users of SolarForecast). National Grid has limited information on what is being produced in large solar farms and none at all about the production from your roof. To keep the UK network stable on a second by second basis requires National Grid to oblige fossil fuel plants (and pumped storage) to adjust their output very quickly, and with little warning.
So the question I asked was this: although on average across the 3 month period the amount of power produced by coal and gas plants was lower than last year, did it have to vary more rapidly during the average day to meet swings in the output of variable renewables? Are gas and coal power stations being asked to increase or cut their output by larger and larger amounts to deal with the intermittency of wind?
The answer seems to be ‘no’.
First, if we plot the average change in required gas and coal plant output in each half hour, the figures do not look very different between 2014 and 2015. As we’d expect, the rate of ramp up in the morning as the nation goes to work look slightly lower. This is because the sun has begun to shine more strongly on to PV panels, choking off the need for more power station output. Between point 14 (7am) and point 20 (10am), the blue line for 2015 is consistently below the 2014 line, but the differences are not great. At the end of the day, as the sun fades, the blue line conversely tends to be above the 2014 figure. Between point 32 (4pm) and point 40 (8pm) the brown line is roughly 100 MW below the blue.
These are averages for the 92 day period. And they show exactly what we’d expect. But the picture needs to be completed by looking at what happens on individual days. Is the degree of variability greater now? Once again the answer is no. The average movement between the required production from coal and gas power stations in each half hour has actually fallen slightly even as PV has surged. The average change - upward or downward - in fossil output from one half hour to the next has in fact fallen from about 530 MW to about 512 MW.
More importantly, perhaps, the number of times that the required output from fossil fuel plants has had to vary very sharply has also dipped, rather than risen.
Number of times output from fossil fuel power stations was required to RISE by 2 or 3 GW or more in a half hour period*
Over 3 GW 24 12
Over 2 GW 150 115
Number of times output from fossil fuel power stations was required to FALL by 2 or 3 GW or more in a half hour period
Over 3 GW 0 1
Over 2 GW 37 31
*Data taken from all half hour periods between 9th May and 8th August.
If anything, it looks as though the period at which power output in the summer** needs to be ramped up fastest – around 7am – now looks easier to manage. The sun is rising at the same time, helping reduce the extra demand by flooding power into the local grids.
The amount of time which the system is trundling along requiring roughly the same amount of power from fossil plants hasn’t changed. The number of half hours in which the upward or downward variation fell between +500 MW and -500 MW was stable at just under 3,000 periods (out of 4,400 or so).
This note has looked at the needs for variability in power output from gas and coal plants. The relatively optimistic finding – that, so far, the system seems to be coping well and is not experiencing problems adjusting output – should not obscure the separate point that renewables are forcing fossil fuel power stations to work fewer and fewer hours per year. Many plants are said to be not profitable. Those of us eager for renewables to grow as fast as possible need to work out how to provide the back up to wind and solar when gas and coal plants have closed because of falling demand.
By the way, I found the results of this analysis surprising. i was expecting evidence of increasing half hourly variability. Please don’t hesitate to write in if you think I’ve done something wrong, or missed a key point.
** This would be very different in the winter, when the fastest ramp up is around 4.30pm, just at the moment PV output - already low - zero.