Optimal operation of energy storage to minimize wind spillage and mitigate wind power forecast errors

Wind spill occurs due to the non-correlation between load and wind profiles, and also wind power forecast errors. Scheduling energy storage units to reduce wind spillage gets complicated considering the difference between day-ahead wind power forecast range, hour-ahead wind power forecast, and actual wind power. This paper presents an algorithm that optimally schedules energy storage to address both applications - minimizing wind spillage and mitigating wind power forecast errors. First, energy storage is scheduled to minimize wind spillage by a mixed integer linear programming, given day-ahead wind power forecast range. Then, day-ahead energy storage operation is updated to mitigate the error between day-ahead wind power forecast and hour-ahead wind power forecast, using Discrete Wavelet Transform (DWT). Finally, energy storage operation is updated by DWT to mitigate the error between actual wind power and hour-ahead wind power forecast. Wind spillage and required back-up generation is calculated for different scenarios of day-ahead wind power forecast, energy storage technologies (e.g. compressed air energy storage, and sodium sulfur battery), and state of charge values. To showcase the applicability of the proposed approach, a case study based on the real world wind and load data obtained from the Bonneville Power Administration in 2013 is presented.