How do solar panels influence UK power prices? Will Tesla’s batteries change this?

Over the last 4 years, the increase in solar power capacity in the UK has been considerable, going from 0.1GW in March 2011 to 6.6GW by April 2015. (See  UK government solar statistics.)

Feed in tariffs (FiT) guarantee the income paid to homeowners for the production of this electricity – and this makes it economically viable (BBC home solar article).  But looking at the bigger picture, what is the impact of solar power on the UK electricity price – the price paid by utilities for power – and hence your and my electricity bill?

Within day shape of power demand and supply

Taking data averaged over 2 weeks in the middle of May 2015 (May 11th – 24th inclusive), figure 1 shows the UK power demand, and how much is supplied by wind, by solar and by conventional plant (nuclear, coal, gas, oil).

The shape of the solar power supply is not surprising – the output peaks in the early afternoon when the sun’s rays are strongest. The supply of power from wind is much flatter across the day. The peak solar output is similar to the peak wind output, at around 4.1GW, for this period. In winter, the peak wind supply exceeds that of solar, whilst on a calm mid-summer’s day, solar supply exceeds that of wind.

Figure 1: mid May 2015 average supply and demand by hour of day

Comparing this with what the plot of demand from conventional plant would look like if there were no solar power, we see the curve is much flatter (figure 2).

Figure 2: within day UK electricity demand shape with and without solar power

Electricity price is influenced by many variables., one of which is how much additional power needs to come on to satisfy the extra demand this hour, and how long it will be on for (generating hours) since start-up costs can be spread over the duration of the running time: starting up a plant to run only for an hour is much more costly than starting up a plant and running it for the next 8 hours. To illustrate this, we look at the average hourly price in the day ahead electricity market for the 2 middle weeks in May, and plot the spot price as a function of how many hours the plant providing the last MW will be running for – with the results in figure 3 below. For plant running more than 17 hours, the fit is very good, whilst for plant running for less than 10 hours the fit is noisier, but generally increasing.

Figure 3: average mid May 2015 spot price as a function of generating hours of the last MW needed to satisfy demand

So spot price is (generally) less the more generating hours a plant has.

Figure 4 shows the distribution of generating hours with, and without, solar supply. With solar supply, around 30% of plant generates for at least half the day, whilst if there were no solar supply this would be close to 50%. So whilst the solar power makes for a greener supply of electricity, it also causes the conventional start-up cost to be spread over less hours, thereby making the conventional component of the electricity more expensive.

Figure 4: how the distribution of generating hours varies with and without solar supply.

Storing electricity

If we could store solar power during the early afternoon solar supply peak, then use it during the morning and evening demand peaks, conventional plant could have a flatter generation profile, thereby distributing start-up costs and lower the cost of the conventional component of the electricity.

Various options are being worked on, with Tesla’s home battery (BBC Tesla battery article) and a gas to hydrogen electolyzer (Guardian electolyzer article) two of the options recently in the news.

To work out the best case scenario for how much these storage solutions could save, let’s make some assumptions:

• The storage is 100% efficient.  In reality, it is currently lower, ~92%, but improving.
• The spot price vs generating hours relationship holds in the new generation plan.  This is a complex relationship and some changes would occur, including a gradual change in the amount of nuclear, coal, gas and oil if this more simple on/off profile existed day in, day out.
• We have perfect foresight as to the amount of solar produced.  In reality, forecasts contain errors, and so the schedule would likely “play it safe” rather than scheduling every last MW of solar energy.
• We aim for a solution as in figure 5 below.

Figure 5: How conventional demand can change with solar power storage

With these assumptions, storing solar and redistributing it according to the red line above would reduce the average spot price from £56/MWh to £47/MWh, a saving of ~15%.

At the current Tesla battery cost, it would take ~15 years to cover of this saving to cover the cost of battery required.

So, whilst the currently available options aren’t cost effective as yet, technology is definitely advancing in this field with wide reaching implications for the European energy markets.