Impact of Forecasting Errors on Day-Ahead Scheduling of Price-Responsive Customers

Abstract

Scheduling the distributed energy resources in day-ahead scenarios requires the use of accurate forecasts of loads, generation, and energy prices. Forecasting errors may significantly influence the quality of the scheduling, resulting in overestimations or underestimations of the power required from the grid by a price-responsive customer. In this paper we provide an insight on the assessment of the impact of forecasting errors in a case study regarding an industrial customer who disposes of a photovoltaic system and of an electrical energy storage system. The day-ahead scheduling procedure, developed in previous works, is applied using forecasts obtained through random forests and through a benchmark model; the obtained results are compared with the ideal scenario, using the actual realizations of the unknown variables, in order to assess the responsiveness of the scheduling procedure to forecasting errors.