Multi-time Scale Unit Commitment Optimization under Hybrid Uncertainties

Abstract

Recent years, the popularity of wind power and the widely use of diversified loads have increased the uncertainty of power systems in both supply and demand sides. This paper develops a multi-time scale unit commitment optimization model under wind power and future load uncertainties. First, dayahead wind power and electric load forecast is obtained by long short-term memory network, based on which the on/off status and first-period output of units are determined. Then rolling economic dispatch is applied when real time data is collected from the system. To solve the above unit commitment and economic dispatch model, an improved particle swarm optimization algorithm is proposed. Finally, several experiment were performed to demonstrate the effectiveness of this research.