Using weather forecast uncertainty minimises electricity costs in low flexibility power systems

The rapid expansion of renewable generation capacities amplifies uncertainty in power system operation due to their inherent intermittency and challenges in accurate forecasting. Despite this, ensemble forecasts, which quantify weather forecast uncertainty, are not yet integrated into power system planning. This paper addresses the conditions under which utilizing this uncertainty is advantageous, focussing on how the available flexibility to balance forecast errors affects its value. To investigate this issue, a tool has been developed in which a day-ahead dispatch is cleared, and followed by a balancing which settles the discrepancy between dispatch and observation. A deterministic clearing is compared to a novel, stochastic clearing, in which the dispatch is optimised to minimise both day-ahead and anticipated balancing costs. The latter are estimated from a set of scenarios equivalent to ensemble forecast members. It was found that stochastic clearing outperforms deterministic clearing strongly in terms of electricity costs up to a threshold of upward flexible capacity. Due to grid constraints, this value could not be related to the distribution of forecast errors directly. This work has demonstrated how employing weather forecast uncertainty in power systems management may secure the supply, while relying not only on the addition of flexibility.