基于GWO-MLP的光伏系统输出功率短期预测模型

准确预测光伏系统的输出功率对微网系统的优化调度具有重要意义。为了减小光伏系统输出功率短期预测误差,文中采用多层感知器(Multi Layer Perceptron, MLP)神经网络作为主要的预测载体,将光照强度、温度、风速数据作为MLP的输入,光伏系统的输出功率作为MLP的输出,采用光伏电站的历史数据对MLP进行训练,并针对MLP在初始化权重和偏置量中存在的随机性问题,提出运用改进灰狼算法(Grey Wolf Optimizer, GWO)对MLP的初始权重和偏置量进行优化,减小MLP随机初始化的误差。仿真结果显示,文中提出的GWO-MLP在均方误差(Mean Square Error, MSE)、均方根误差(Root Mean Square Error, RMSE)、平均绝对误差(Mean Absolute Error, MAE)方面较MLP、Elman神经网络、支持向量机(Support Vector Machine, SVM)、极限学习机(Extreme Learning Machine, ELM)都有明显提高,表明所提方法可以准确预测光伏系统的输出功率。

Short term prediction model of output power of photovoltaic system based on GWO-MLP

Accurately predicting the output power of PV system is of great significance for optimal scheduling of microgrid systems. In order to reduce the short-time prediction error of photovoltaic system output power, the paper adopts a multi layer perceptron (MLP) neural network is used as the main solution, and the radiance, temperature and wind speed are taken as the input of the MLP, and the output power of the PV system is used as the output of the MLP. The historical data of PV plant is used to train MLP. The improved grey wolf optimizer (GWO) is used to optimize the initial weights and biases of MLP to reduce the error of random initialization of MLP. The simulation results show that the proposed IGWO-MLP is better than MLP, Elman-NN, SVM, ELM in terms of mean square error, root mean square error and mean absolute error, indicating that the proposed method can accurately predict the output power of PV systems.