Optimization Study of Fluctuation Prediction and Standby Unit Dispatch for Wind and Solar Power Generation

Abstract

With the growing use of renewable energy sources worldwide, the volatility problem of wind and solar power generation challenges the stability of power systems. In this paper, the problem of predicting the volatility of wind and solar power generation and optimizing the dispatch of standby units based on historical data is investigated. First, support vector regression (SVR) and random forest regression (RF) models are used for short-term volatility prediction of wind and solar power generation. The prediction accuracy of the models is improved by optimizing the thresholds, in which the RMSE of the SVR model is 7.49 and that of the RF model is 6.72, which are superior to the traditional methods. Next, a two-objective optimization model is proposed to optimize the standby unit scheduling with particle swarm optimization (PSO) algorithm to balance the stability of the power plant and the cost of the standby unit. The experimental results show that the optimized scheduling strategy can effectively reduce the start-stop frequency of the standby unit and improve the economy and reliability of the power system. The research in this paper provides an effective solution for the fluctuation prediction and scheduling optimization of wind and solar power generation, which has important application value.