It is also suggested that conventional plants may need to be shut down and restarted more often, again leading to additional costs through wear and tear and because the start-up and shutdown processes are inefficient.
A report, The Western Wind and Solar Integration Study Phase 2, published last year by the US National Renewable Energy Laboratory (NREL), however, demonstrates that the extra costs are very modest. With 33% solar and wind energy penetration, "cycling costs", as they are termed, account for $0.14-0.67/MWh. This is set alongside the fuel cost savings, which are around $28/MWh.
As a corollary of this finding, the study also concludes that the savings in carbon-dioxide emissions are roughly equal to the penetration level. So, with 33% of the electricity coming from wind, the carbon-dioxide emissions come down by 33%. This is more than some critics have previously argued.
The way in which the output from conventional plants need to be adjusted to accommodate the variability of wind can be gauged by looking at the way wind power and system demand vary on a typical electricity network. For example, in western Denmark over a two-day period at the end of 2013 (see below), there were very strong winds over most of Europe for a period of time. The graph shows how demand and wind-power production varied. For seven hours in the morning of 30 December, both wind and demand were increasing, therefore the wind-power production was actually easing the requirement placed on the thermal plant to increase production.
In practice, the output of the thermal plant was still increased during that period, probably because some additional power was sent over one of the interconnectors.
After 4pm, demand increased, and an increase in wind-power production again alleviated the excess demand on the thermal plant. Then from 7pm, demand decreased steadily. As the wind output was initially increasing, the thermal plant output was reduced and, possibly, more energy was dispatched over the interconnectors. From midnight until approximately 4am on 31 December, wind and demand were again falling together. For the rest of that day, there was a similar mixture of wind and demand sometimes moving together, sometimes in opposite direction.
Rare power swings
It is also worth noting that the magnitude of the hour-by-hour variations in wind output are similar to those of the demand. While there may occasionally be large wind-power swings, experience from a number of integration studies across the globe suggests that these are fairly rare.
The NREL study concludes that it is start-up and shutdown costs that dominate the cycling costs, rather than ramping costs. The modest size of the ramping costs may or may not be accurately reflected in the overall prices that generators charge for the provision of balancing services - contracts to increase or decrease output on demand. These are the mechanism that system operators use to ensure that their systems are kept in balance.
A recent review of balancing charges for wind energy, Integration Costs and the Value of Wind Power by the US Association for Energy Economics, published last year, reflected a wide range of prices, averaging about EUR2.5/MWh with 20% wind energy. The most recent studies mainly quote lower prices than many of the earlier studies, suggesting that increasing experience and better wind predictability are leading to a reduction in the costs of assimilating wind energy.