It is often assumed that renewable generation like wind power cannot be relied upon to provide any level of system security, because of its variable power output. We suspected that this was excessively pessimistic, and were interested in modelling the level of output that wind farm portfolios can guarantee to deliver or exceed over extended periods. If we accept that such guarantees inevitably come with some small, specified level of risk that it will be unable to deliver this power output at some point during the specified period, we can investigate how such guaranteed levels vary with the risk of failure and recent wind conditions. This work was conducted as part of our Distributed ReStart project with NGESO, quantifying the extent to which a group of wind farms in Scotland could contribute to restoration.
TNEI developed and utilised a cutting-edge statistical time series model (“SARIMA-GARCH”) to simulate wind power output trajectories, conditioned on recent conditions. This allowed the prediction of the MW level of reliable power output that could be delivered by different combinations of these wind farms, at any point in time, and with any chosen risk of failure.
Using this model, we were able to represent how the amount of output that wind could contribute during restoration depended on the level of risk that was considered acceptable, the duration of the period for which the output was required, and the nature of wind conditions immediately before the period starts.
We also revealed the degree to which some wind farm combinations are able to make a much more significant contribution to restoration than other others.