Short period Wind Power Forecast Method Based on Maximum Correntropy Criterion

Abstract

Wind power forecast is an essential measure to increase the level of wind velocity consumption and also the basis of wind power dispatching operations. In the cause of increasing the precision of short-period wind velocity forecast, this srticle adopts a new evaluation criterion(MCC) in order to direct optimization of the parameters of the wind power model. The method first filters and normalizes the measured historical data and determines the optimal input variable dimension through a set of fixed parameters. Then, grid optimization and Particle Swarm Optimization (PSO) are used as evaluation criteria for parameter optimization. Finally, the obtained optimal parameters are used to forecast the short-term wind power, and four evaluation indexes are used to evaluate the forecasting results. The results show that the parameter optimization process using MCC as the evaluation criterion is more suitable for wind power output characteristics, and comparing with traditional optimisic methodthe the forecast accuracy is increased by 5 -10%.

Wind Power Interval Prediction Based on Robust Kernel Density Estimation

“…Based on the above situation, some scholars use statistical methods to study and construct the prediction interval of wind power generation load. The parameter estimation method is often used to fit the distribution of wind power generation load [6] , but the strong nondeterminacy of wind power generation load output is difficult to describe by a typical probability distribution accurately, so nonparametric statistical methods are used to mine the actual probability of wind power generation load output distribution information. Kernel Density Estimation(KDE) is a typical non-parametric estimation method widely used in wind power forecasting.…”