By Dave Elliott
The UK Energy Research Centre (UKERC) has produced an update to its 2006 report that had looked at the costs and impacts of using ‘intermittent’ electricity from renewables such as wind and solar. The 2006 study had only examined impacts with up to a 20% input, but the UKERC researchers now say that, even at the higher levels we are now expecting, it was still the case that the costs of balancing renewables could be low. However, they warned that, unless ‘urgent’ action was taken by the government to boost grid flexibility, the costs of adding renewables in future will be ‘much higher than they need to be’.
UKERC said the key to avoiding excessive cost was to invest in flexible systems such as demand-managing technologies, including demand-side response and storage, which can dramatically reduce the cost of grid balancing: ‘Ten years ago, penetration levels for renewables were small, and the costs of managing the grid to incorporate wind and solar were pretty trivial compared to the costs of building wind farms and solar panels themselves. Now, however, costs have fallen and renewables are close to being cost-competitive with fossil fuels. With higher penetration levels come higher system costs, and building flexibility into the system becomes much more important.’ If that’s done, the study puts the extra cost of using variable renewables at under £15/MWh for a 50% renewables input.
It explains that ‘In 2006 many of the impacts of and additional costs related to variable renewable generation were found to be negligible. The exceptions were 1) the cost of additional short-term reserves required to balance electricity supply and demand over the timescales of seconds to hours, and 2) the costs of the generating capacity required to ensure that a system can reliably meet peak demand. However, as penetration levels rise, other significant impacts come into play, including: curtailment (where variable renewable generation cannot be accepted onto the grid); transmission and distribution network reinforcement costs; the potential for reduced efficiency of the remaining thermal plant on a system; and the need to ensure that a system has sufficient mechanical inertia to maintain frequency stability. Depending on their design and operation, these impacts may also manifest themselves in the relative market value of output from both conventional and variable renewable generators.’
However, there can be interactions between these measures and the overall system cost. For example, adding more grid links or storage could reduce curtailment when there was excess green power and the need for more capacity for balancing when there were shortages. So the UKERC adopts a ‘whole system’ optimal approach but, even looking at individual components, it suggests that the added system balancing cost would be low.
‘At a 30% penetration level, where results from wind-based analyses dominate, most estimates are in a range between £4 and £7/MWh, with some outliers. All except two data points of the entire data set lie below the £15/MWh level, even as penetration levels rise to 50%. These findings are supported by the project team’s own calculations based on estimates of conventional plant cost (in this case CCGT) which suggest that costs will lie in the range between around £4 and £8/MWh at an assumed 20% capacity credit, between £9 and £11/MWh at a 10% capacity credit level, reaching a peak of less than £15/MWh even if the capacity credit of the variable renewable plant is assumed to be zero’.
That’s at the very most around 10% of the current cost of offshore wind and, more likely, much less. These results are based on a review of 200 studies, focusing on the medians. By contrast, in a recent post Energy Matters saw balancing as adding 36% to the cost of its 2050 renewables scenario – £37/MWh.
Some of the high outlier estimates that have been produced in recent years include what are sometimes called ‘profile costs’ – the cost imposed on other generators by having variable, but marginal-cost, renewables on the grid. In Germany, for example, marginal cost PV solar is pushing gas plants out of the lucrative day-time peak demand market – so they are losing money. The Potsdam Institute claimed that this sort of change meant that the total extra operational cost to the system of having renewables on the grid could therefore be double its estimate for just the simple balancing cost, taking the total to €65/MWh.
That may overstate both costs (IRENA’s estimates are much lower) but in any case, the German government is adamant that the generators must deal with meeting and balancing demand, including any profile costs – and has set the market rules accordingly. But some say that the suppliers should be compensated. That in part is what’s happening in the UK with the capacity market – generators are offered a subsidy to be available when needed. Although it is not seen as relating to profile cost, but rather as reflecting the cost of having plants available as a reserve for when there is no wind and no sun.
The UKERC study estimates what it calls the reserve cost at up to £5/MWh, if wind and solar supply 30% of electricity demand. At 50% penetration, this cost could rise to between £15 and £44/MWh, with the top end reflecting a very inflexible electricity system. That might be seen as being on top of the system integration cost. However, the UKERC warns that there may be some double counting here – you can’t just add the system integration and reserves cost. Some of the reserve capacity is used to deal with short term variability and peak demand. Moreover, since we don’t know what the supply mix or balancing mix will be, it’s hard to give a fair total balancing cost at this stage. The best guess, assuming most of the right flexible balancing measures are adopted, seems to be around £10/MWh, at a 30% penetration. That’s where we will be soon, in the 2020s, in the UK. Longer term it’s harder to say. Maybe £15/MWh? Given that, with more renewables on the grid, capacity credits reduce and, unless flexible balancing spreads, curtailment issues (due to surplus production at times) will become more important – though the UKERC says it is low at present in the EU generally.
However, if balancing includes the reduction of these surpluses (via demand management) or their use via storage or export, then curtailment can be avoided, and the balancing costs could be offset, partially or even entirely. That is what the recent Aurora study for the Solar Trade Association suggested. At worst, it said, integration costs would reach no more than £7/MWh, with 40GW of PV in 2030, and with storage, assuming battery costs continued to fall, much less. Indeed, there might be an overall system cost benefit, a point made earlier by Agora Energiewende in the German context.
Certainly for the UK, the government’s National Infrastructure Commission’s ‘Smart Power’ study claimed that, by 2030, the adoption of smart grid demand management, storage and wider grid integration could save £8bn p.a, since there would be less curtailment, lower peaks and less need for backup capacity, plus the potential to export surpluses. Clearly it will cost something to adopt these flexible systems, but as the UKERC report argues, they are vital if we want to avoid operational cost increases. And, as I argued in my IOP ebook ‘Balancing Green Power’ last year, they might give us a much better system, with supply and demand matched more efficiently.
A good overview: www.carbonbrief.org/in-depth-whole-system-costs-renewables
Interestingly, the 2016 report for DECC from Frontier Economics, only recently released after being sat on for some reason, came to similar conclusions to UKERC.
Though the GWPF was still not happy!