By Dave Elliott
Several academic studies have indicated that balancing variable renewables need not be expensive. An authoritative review of over 200 studies by UK Energy Research Centre in 2006 concluded: ‘Intermittency costs in Britain are of the order of £5 to £8/MWh, made up of £2 to £3/MWh from short-run balancing costs and £3 to £5/MWh from the cost of maintaining a higher system margin’.
Subsequently Imperial College London (ICL) has produced a series of influential studies developing this analysis, based on detailed scenarios. A recent one was ‘Whole-system cost of variable renewables in future GB electricity system’, carried out jointly with RWE Innogy, Renewable Energy Systems and ScottishPower Renewables.
It concludes that ‘system integration costs of variable renewables remain at a relatively low level even at penetration levels that are 3 times higher than today, provided there is only a modest improvement in system flexibility (such as through deployment of modest amount of energy storage and/or DSR)’. And it claims that ‘up to £4.7bn/year could be saved by improving system flexibility from today’s level’.
However, it will not all be plain sailing. The report warns that, since wind and solar PV are non-synchronous power sources, which tend not to contribute to system inertia, ‘the importance of frequency regulation in the future GB system is …expected to increase dramatically’. But it notes that ‘energy storage technologies and demand-side response could … significantly enhance system flexibility’ and the high value of flexibility is one of its main messages.
It focuses on system integration costs (SICs), the additional cost at the system level required to securely integrate a unit of generation technology. It notes that some of these costs are ‘not necessarily or wholly included in the capital or operating cost estimates’, e.g. increased balancing cost associated with increased requirements for system reserve due to higher uncertainty of variable renewable generation output, and increased requirements for fast frequency regulation (response) due to reduced system inertia as well as larger maximum generating unit size. There are also network reinforcements required in interconnection, transmission and distribution infrastructure (e.g. to connect remote wind resources) and increased backup capacity cost due to limited ability of e.g. variable renewable technologies to displace ‘firm’ generation capacity needed to ensure adequacy of supply. Finally, there are the costs of maintaining system carbon emissions, as the addition of some technologies may cause the overall emission performance of the system to deteriorate, requiring that additional low-carbon capacity is installed to maintain the same level of carbon emissions.
The ICL report says that ‘some of these components, such as increased balancing or network cost are already reflected to some extent in current charges imposed on generators, such as the balancing payments (BSUoS), or transmission and distribution charges (TNUoS and DUoS). Nevertheless, these charges are not fully cost-reflective. Some of the above components such as the backup capacity are not currently included in LCOE assessments in any form whilst other components such as constraint payments are completely imposed on distributed generators through the bilateral conditional grid connection agreements which include provision for uncompensated constraint’.
The report develops a variety of scenarios with different mixes of technology and flexibility, but the same overall level of emission reduction, and calculates SICS, using nuclear as a comparison baseline. It concludes that ‘a moderate improvement of system flexibility… brings the cost of the system down by £3.5bn/year, while at the same time reducing SIC of wind from more than £40/MWh down to around £11/MWh in the 2030 horizon. This level of SIC combined with the LCOE assumptions makes both offshore and onshore wind cost-effective compared to nuclear generation’.
Some SICs do increase with large renewable penetrations, for PV especially, due to local grid reinforcement needs, with the SIC for offshore and onshore wind in 2030 put at around £5-9/MWh across the medium to high flexible scenarios, while the SIC for solar PV in 2030 is put at £10-15/MWh with medium flexibility assumptions. But, it notes that ‘despite an increasing penetration of variable renewables between 2015 and 2030, SIC of wind and PV can be maintained at a relatively stable level (or even lower in some scenarios) provided that sufficient amount of flexible options is deployed’. And overall it says that ‘according to the LCOE assumptions adopted in the study, despite the positive SIC the whole-system cost of offshore and onshore wind and PV (i.e. the sum of their LCOE and SIC) still makes them more attractive than nuclear in the majority of 2030 scenarios with modest or high flexibility levels’.
E3G used this study for its own report, which not only claims that system integration costs would stay under £10/MWh for a wide range of system characteristics even at increased levels of renewables, but also says the overall cost of renewables will be low. Under current trends, onshore wind will, it says, be at least 22% cheaper than nuclear while offshore wind and solar PV will provide savings of over 4% & 8% respectively.
Evidence submitted by University College London (UCL) to the Lords’ inquiry on UK energy economics adopts similar views on renewables and their balancing costs (UCL Energy Institute & Institute for Sustainable Resources–Supplementary Written evidence (UEM0081): House of Lords Economic Affairs Committee’s Inquiry into The Economics of UK Energy Policy, see also the hearings sessions, including useful inputs from UCL Professor Mike Grubb. All at: https://www.parliament.uk/business/committees/committees-a-z/lords-select/economic-affairs-committee/inquiries/parliament-2015/uk-energy-policy/). It claims that backup costs ‘tend to be a relatively small penalty to the cost of renewables in terms of capacity backup (storage, which can arbitrage between these periods, has also seen radical cost reductions). ‘Dynamic’ backup – the ability to deal with the fluctuations of renewable output – is also relatively small’.
At the Inquiry hearings last year, Professor Mike Grubb from UCL suggested that the balancing cost for variable renewables, ‘is certainly less than £10 per megawatt hour, and it is probably of the order of 10% or less of the investment cost of renewables’. And he noted the new report for the Solar Trade Association from Aurora (see my earlier post), which put the cost of intermittency for the 11GW of solar currently on the system at £1.3/MWh.
UCL’s evidence says that an efficient and secure low carbon system requires an efficient mix of options to serve the residual load. But it would not need a lot. However, at least 20GW of peaking plant may be required to operate up to 15% of the year, for example on dark cold windless winter nights. This could be secured ‘by options that either already physically exist or otherwise involve low capital but high running costs: demand-side response, and new investment in e.g. reciprocating engines powered by diesel or gas.’ But even for new build, ‘the annualised capital +O&M cost of such peaking plant adds about one tenth to the corresponding capital+O&M (historic) cost of onshore wind’. Around 25GW may it says ‘be required to operate ‘mid merit’, e.g. for between 15 and 85% of the year. For this, in addition to any nuclear (which may be problematic for these conditions), the major options could include interconnection, biomass plants, and existing and new CCGTs. For new CCGTs, again, the cost adds about one tenth to the corresponding capital+O&M (historic) cost of offshore wind’.
So with generation costs falling and a rough upper backup costs estimate of ‘significantly less than £10/MWh’, there do not seem to be major problems. Government policies aside – as witnessed by the results of the last capacity market auction. Although storage did at least get a look in, including 500MW of batteries, it is still not focused on flexible options.
The UKERC has now produced a revised version of its 2006 review of renewable integration costs, which I will review in a forthcoming post. Its conclusions generally confirm the views above, but add a sense of urgency.