By Dave Elliott
Eric Martinot has been one of the key people behind the indispensable annual REN21 Renewable Energy Network publications reviewing the state of play with renewables globally: see The latest one says renewables now supply around 24% of global electricity. Martinot has also taken time out to look in detail at renewable integration issues, and this work has led to some excellent publications, including a useful non-technical overview ‘Grid Integration of Renewable Energy – Flexibility, Innovation, Experience’ for the Annual Review of Environment and Resources 2016 (February 2016). It focusses on the concept of flexibility, which it sees as vital for balancing variable renewables, including small continual changes, sudden large swings in availability, requiring rapid ramp-up of backup plants or other balancing measures, as well as for occasional longer periods of low input.
In this ARER paper Martinot says that ‘flexibility needs for system ramping have become a major concern in California, where the grid operator CAISO predicts a 13-gigawatt ramp occurring over a 3-hour period each afternoon by 2020, due to solar output declining as the sun sets. Meeting that 13-GW ramp is the equivalent of turning on thirteen nuclear power units over a 3-hour period, every afternoon. In Germany, projected ramps by 2022 reach an unprecedented 40 GW. However, Germany proved in 2015 that it is already able to handle a 13-GW ramp today, with little difficulty. This “demonstration” occurred during a mid-day solar eclipse, which caused a 6-GW down-ramp of solar (over 60 minutes) followed by a 13-GW up-ramp (over 75-minutes). Germany’s import and exports with neighboring countries, power market design allowing negative prices, and flexible coal plants together handled these ramps with no power outages’.
Making the system flexible to deal with this will add to the cost. The paper notes that ‘because each power grid is different, and consequently the measures needed to increase flexibility are different, and because analytical underpinnings are not well developed (including what counts as additional or incremental costs), the field is still relatively undeveloped, and controversy exists over how and what to count. For Germany, Agora gives integration costs of onshore wind and solar power, counting costs of “grid reinforcement” and “balancing” (ancillary services and forecast errors) as 0.5-1.3 eurocents/kilowatt-hour (kWh), which represent perhaps one-tenth to one-twentieth of the direct costs of renewable power. Agora also adds costs of 0.0-1.0 eurocents/kWh for costs imposed on the conventional generation fleet, in terms of “back-up capacity” and lost revenue, an even more controversial and difficult-to-quantify figure. For Europe as a whole, Pudjianto et al. similarly estimated integration costs of solar power at between 0.5-2.5 eurocents/kWh counting all costs, while also noting that integration costs decline when demand response or energy storage are present.’
As I point out in my new IoP book, ‘Balancing Green Power’, which covers some of the same areas, some put the costs higher, but these two studies otherwise pretty much agree, both of them reviewing the various balancing options, including flexible fossil generation, storage, smart grid demand management and supergrid imports and exports. Both studies agree that nuclear plants can’t help much, but Martinot sees coal plants as offering at least some flexibility. Martinot’s ARER paper goes on to look briefly at market design and policy support issues and overall is a very valuable contribution to the debate, as are his other more detailed new (in some cases joint) studies, which include one on California and one on US, Danish and German experiences for a Chinese study project: www.cpuc.ca.gov/WorkArea/DownloadAsset.aspx?id=9141 and http://www.martinot.info/Martinot_CVIG_2015_DE-DK-CA.pdf Also see: http://www.nrel.gov/docs/fy15osti/63366.pdf and http://www.martinot.info/Martinot_Kristov_Erickson_2015.pdf
What next? Martinot’s reviews of the field open up many new issues, some of which also emerged from my IoP book. For example, what are the least-cost approaches to balancing and least-cost combinations of measures? What are analytical frameworks and tools for determining least cost? Should incremental ‘integration costs’ be used, or should total system cost under different scenarios drive planning? How will market designs affect the quantity of flexible resources in the future? How can cross-border integration of markets facilitate greater interregional interconnection and balancing? What institutional changes are suggested, considering electricity, heating, and transport together, and considering the interests of electricity market participants? As Martinot concludes in his ARER paper, ‘the answers to these questions are not simple, and such inquiries will continue well into the coming decades’.
For the moment, the focus is on market adjustments to deal with backup requirements and curtailment issues. Martinot mentions the ramping market in California, designed to reduce stress on the system when renewable supply and/or consumer demand varies. Interestingly he notes that storage is not yet a major focus. For example, in Germany it is only seen as being used ‘after 2032, as Germany approaches a 50% share of renewables’, while ‘Denmark has no plans for electricity storage, relying instead on heat storage’, with CHP seen as a key balancing option. He adds that ‘California has mandated 1.3 GW of storage to be procured by its power companies by 2020, a relatively modest amount as California reaches 33% renewables’. Demand response was also mostly limited, with for example only around 1GW in Germany, although some variable ‘time-of-use’ tariffs were being considered in the US and elsewhere, along with smart grids and vehicle-to-grid charging systems.
Although he doesn’t cover it, the situation is similar in the UK – storage and smart grid demand response is still marginal. However, a recent report from the National Infrastructure Commission talked it all up, claiming that by 2030 smart power systems, including storage, demand response and new interconnectors, could save consumers up to £8bn p.a. And with concerns rising about supply shortfalls due to coal plant closures, an adjustment has recently been made to the capacity market to extend and bring forward new balancing/backup capacity, although the emphasis it seems will be on getting more gas-fired plant online, rather than developing new demand management or storage systems for balancing renewables. DECC said it would hold a new capacity auction for winter 2017/18 early next year, ahead of the capacity already agreed in the first two auction rounds: www.gov.uk/government/news/reforms-to-capacity-market-to-improve-energy-security-for-families-and-businesses Reaction: https://sandbag.org.uk/blog/2016/mar/1/capacity-mechanism-reform-final-hurrah-unabated-ga/
As renewables expand around the world, grid balancing will become more and more important. Building more fossil gas back-up plants as in the UK, or using coal plants more flexibly, as in Germany, may be cheap in the short-term but, given their emissions, other options are needed longer term. Demand response systems, along with smart grids and storage, and long distance supergrids, can manage locally variable supply and demand flexibly. As Martinot notes, the right mix will vary with location, and it is clear that fossil backup will persist for a while, but longer term, as I argue in my book, with power to grid conversion of surpluses and the other balancing measure fully developed, that may be be necessary. My new IOP book: http://iopscience.iop.org/book/978-0-7503-1230-1