基于BP神经网络与非参数核密度估计的短期风电功率概率区间预测

Probability interval prediction of short-term wind power based on BP neural network and non-parametric kernel density estimation

考虑到风电功率确定性预测已不能满足电网调度和电网规划的要求,提出了一种基于BP神经网络与非参数核密度估计的短期风电功率概率区间预测方法。在利用神经网络进行功率点预测的基础上,将功率进行分区段误差统计,求取各区段核密度估计的概率密度函数,并在给定置信度下,预测风电功率波动区间。通过对我国依兰风电场数据进行仿真与分析,表明该方法具有良好的实用性,在置信度为90%下,各项评价预测指标均满足工程需求。

Considering that the deterministic prediction of wind power can no longer meet the requirements of power grid dispatching and grid planning,a short-term probability prediction method for wind power is proposed based on BP neural network and non-parametric kernel density estimation.Based on the prediction of power points using neural networks,the error statistics of the power is conducted by blocks,the probability density function of the core density estimation of each block is obtained,and the wind power fluctuation range is predicted under a given confidence.Through the simulation and analysis of the data of the Yilan wind farm in China,it is shown that the method has good practicability,and all the evaluation and prediction indicators meet the needs of the project with a confidence of 90%.