Identifying calendar-correlated day-ahead price profile clusters for enhanced energy storage scheduling

Optimising the scheduling of energy storage systems with respect to multiple revenue streams is crucial to the business case for installations in the UK and other countries with high electrical grid penetration. In this work we use hierarchical clustering for the first time to correlate groupings of UK day-ahead electricity price profiles with calendar period. We observe that there are three primary clusters in the 2017–2019 dataset, and hypothesise that these arise from the interplay of winter/summer variations in demand along with longer term variations in the wholesale gas price. Looking at finer detail, we find that in summer 2018 there is a clear split in weekday/ weekend price profiles, with the latter showing a significantly delayed price peak, and higher night time prices. These findings demonstrate the usefulness of the approach for revenue stacking, as the optimal bidding for ancillary services to fit around the performance of peak shaving will be influenced by the knowledge of such patterns, especially when the horizon for bidding is beyond the day ahead.