Sensitivity Analysis of Day-Ahead Energy Management Strategies under Variant Resolution Mission Profiles

Abstract

Energy management strategies directly influence the performance (e.g., cost) of residential PV systems. In state-of-the-art, most forecast-based energy management strategies adopt hour resolution to make decisions, which may have poor robustness and flexibility to fast-changing weather conditions. This paper compares the sensitivity of three typical operation strategies including maximum self-consumption, mixed integer linear programming, and particle swarm optimization under different kinds of resolutions and weather. The results show that the maximum self-consumption strategy has the best robustness and can utilize the battery most at the expense of total cost. In terms of cost, the mixed integer linear programming strategy performs best on the sunny day, and has the best scheduled result on the partly-cloudy day, but the particle swarm optimization strategy performs best in the real case.