Operational load forecasts for dynamic flexibility procurement

Changes to the management of electricity distribution networks requires the development of essential system operation functions, which include refined forecasting capabilities. The work presented here identifies the most suitable approach to optimise and de-risk the integration of artificial intelligence (AI) into demand forecasting functions for operational decision making. As distribution network operators (DNOs) develop their distribution system operation (DSO) functions, a step change is required in their forecasting capability to optimise flexibility procurement. Annual, assumptions-driven models that forecast load at primary and secondary substations will not suffice; operational forecasts at the monthly, weekly and daily level are required to inform flexibility procurement and dispatch decisions that enable deferral of network reinforcement expenditure.