What is best route to reduce the UK’s reliance on gas and fossil fuels for electricity generation?

This note provides an outline of one approach, using real data from the last year. The analysis shows that all the UK’s current electricity demand could be met by an expansion of onshore wind, offshore wind and solar PV to about 4.5 times current levels.

Crucially, half hourly matching of supply and demand across the year is carried out by hydrogen. When electricity is in surplus, electrolysers are used to make hydrogen. When in deficit, the hydrogen is burnt in gas turbines to generate electricity. The chart below illustrates this for 25th June this year when wind and sun would have been converted to hydrogen during the day but hydrogen would then have been needed in the night hours.

In this simple model fossil fuels are never needed to supplement supply. The inevitable energy losses in the conversion processes mean that more electricity is needed to cover total needs over the year but the excess production is less than 25% of the total. Complete electricity independence would thus come from having about 50-60 gigawatts of each of solar PV, onshore wind and offshore wind, up from about 11-14 gigawatts each today.

The note only looks at replacing fossil fuels in the current electricity supply. As decarbonisation proceeds, more activities – particularly transport and building heating – will be switched to electricity. This will increase the demand for power but this can also be handled by an expansion in renewables, as long as a storage medium such as hydrogen is used to balance supply and demand.

The analysis

1, The data used in the article was provided by Drax Electric Insights for the 365 days from 1st July 2020 to 30th June 2021. (Thank you particularly to Iain Staffell of Imperial College for giving me access to the database). Electricity demand and supply is logged for each half hour period. Electricity supply is split by type of generator, including fossil fuels, nuclear, renewables, storage and international connections.

2, I have added extra columns to the spreadsheet that allow me to multiply wind and solar output in each line by a multiple that can be varied. Thus if solar output is 5 GW in one half hour period and I use a multiple of 2, the column changes the output to 10 GW. I can use different multipliers for onshore, offshore and solar.

3, Sometimes the multiplied estimated output exceeds the electricity demand for that half hour. Sometimes, even after applying the multiple on all three sources of electricity, demand still exceeds supply.

4, In those periods when there is a surplus from renewables, the spreadsheet diverts that surplus into making hydrogen. The conversion efficiency is assumed to be 70%. That is, the energy value of the hydrogen that is generated by the electrolyser is 70% of the electricity input. 10 MWh in to the electrolyser creates 7 MWh of hydrogen (lower heating value). This is slightly higher than standard PEM or alkaline electrolysers today but will almost certainly be achieved within the next two years. The hydrogen is assumed to be stored.

5, In periods of deficit when renewables output doesn’t cover power demand, hydrogen is assumed to be taken from storage and burnt in a combined cycle gas turbine (CCGT), much as natural gas is today. Many of today’s large gas turbines can be converted to burning hydrogen, say the market leaders, GE, Siemens and Mitsubishi. The conversion efficiency is assumed to be 60%, roughly the performance of modern turbines. For every 10 MWh of hydrogen burnt by the gas turbine, 6 MWh of electricity is generated.

6, The process of generating hydrogen is 70% efficient and then turning back into electricity preserves of 60% of the energy value. This means that any period of deficit will require electricity generation that is only 42% efficient (70% times 60%). 10 MWh of electricity turned into hydrogen in a half hour of surplus will at some point be converted back into only 4.2 MWh of power at a time of deficit.

7, Simple calculations allowed me to work out how much electricity output from renewables needs to be expanded so that all electricity needs over the course of the July 2020 to June 2021 would be met by renewables with hydrogen storage. The UK’s electricity requirements could be met by many different combinations of offshore wind, onshore wind and solar. I found that the most ‘efficient’ mix of sources was to use approximately the same multiples of each of the three existing sources. By ‘efficient’ I mean the combination that requires the least amount of renewable electricity to be converted in hydrogen for storage. This mixture of sources therefore most closely matches the electricity required over the 17,520 half hour periods each year.

The results

8, A combination of 4.54 times expansion of onshore, offshore and solar produces a total of 678,913,290 MW, for half hour periods, or about 339.5 TWh. The actual demand in the year under study was 276.8 TWh, a lower figure than usual because of the impact of Covid. So the electricity system would have produced about 63 TWh more power than would have been generated in a fossil fuel network. This extra amount is what it takes to make the hydrogen in periods of surplus and use it in deficit half hours.

9, With this mixture of resources, the electricity system would have been in surplus for about 10,300 half hour periods and in deficit for about 7,220. The system would have required electrolyser capacity of about 75 gigawatts if the target were to ensure all electricity generated was used. In reality, if it is only to be used a few hours a year this would never be a sensible electrolyser capacity to install. In this case, just 30 GW of electrolyser capacity would have captured all but about 15 TWh of the surplus, or about 25% of the total surplus generated. It probably would be cheaper to overbuild the renewables rather than increase electrolyser capacity.

10, Other combinations of renewable sources usually require more electricity to be generated to cover the periods of deficit. In other cases, when one renewable is expanded more and another less, the match between demand and supply is less good, requiring more storage in the form of hydrogen to meet periods of deficit. However the differences are not enormous. For example, a system which multiplied onshore wind by 4, onshore wind by 2 and solar by 13.5 would have covered overall demand at the expense of generating about 352 TWh, rather than 339.5 mentioned in paragraph 8.

11, The only route of a reliance on fossil fuels is to base the energy system on renewables. The transition will be painful and expensive but will eventually result in lower costs and far greater less exposure to geopolitics. The current crisis of gas supply should oblige us to begin to take the difficult steps towards complete decarbonisation.

Appendix

What does a 4.54 times expansion mean?

1, Solar would move from 13.1 GW to 59.5 GW capacity

2, Onshore wind from 13.6 GW to 62.1 GW capacity

3, Offshore wind from 10.7 GW to 48.7 GW capacity.

The UK land and sea space is comfortably capable of accommodating this increase.

The issues with the analysis

1, The spreadsheet manipulation is overly simple. It assumes that all electricity supply comes from the three renewable sources. Other non-fossil sources (principally imports, nuclear and biomass) are excluded from the analysis. If they were included, the required multiple for renewables expansion would be lower.

2, It assumes constant capacity of solar and wind during the period. In fact, renewables capacity rose by just over 1% during the year. I use the figures for the end of the year. This will have marginally increased the multiple of new capacity needed.

3, I don’t take imports and exports into account, not least because shortages and surpluses are likely to be continent wide. When it is windy in the UK, it is very likely indeed to be windy in northern Germany and Denmark, for example.

4, My assumptions are also conservative in assuming that future renewables capacity is no more productive than at present. In fact offshore in wind in particular is improving in its capacity factors as the size and height of turbines increases.

5, I haven’t taken into account any storage costs for hydrogen. Nor have I modelled whether it would be more efficient, for example, to install multiple gigawatt hours of batteries to act as the first reserve in periods of excess or deficit. Hydrogen makes sense when storage is over longer periods but batteries work well as the storage medium for intra-day periods. (Storage losses with batteries are likely to be less than 10%, compared to almost 60% with hydrogen).

Follow on work

1, This programme needs to be costed. Because of the conversion losses from electricity to hydrogen and then back, more electricity needs to be generated than is apparently needed. Does this make the proposed route to self-sufficiency costlier than alternatives? I don’t think so but this topic needs more research. This note suggests that from an energy supply viewpoint it makes good sense to expand solar, offshore and onshore by roughly the same multiple. But does this make sense financially? Might it be better, for example, to focus on adding solar PV because the Levelised Cost of Energy (LCOE) is probably lower than offshore wind at the moment?

2, It would be best to do much more sophisticated modelling of the impact of existing and future nuclear power, hydro and pumped storage, biomass and imports. In addition, it would worth assessing the potential effect of demand shifting during the course of the day. But this could only reduce the total need for new wind and solar.

3, Critically we should also be assessing the impact of the future expansion of electricity demand due to heat pumps and transport electrification. This is complex; we would need to overlay temperatures to assess the need for electric heating. Much demand for electricity in transport can be deferred for hours, if not days. Modelling this would be very difficult indeed.

4, We also need to accurately model exactly what the capacity factor of new installations is likely to be, and exactly when the output will occur. Much solar PV is on household roofs not facing due south. But a very large percentage of new solar is likely to be in south facing solar fields so the total amount of power generated by 1 gigawatt of PV will be greater, but it will be concentrated in the hours around midday. We should also factor in the growth of large scale battery capacity.

Chris Goodall

chris@carboncommentary.com

September 23rd 2021