基于小波神经网络的光伏系统发电量预测

Generation forecasting for photovoltaic system based on wavelet neural networks

针对光伏系统发电量的影响因素,建立具有超强泛化能力的小波神经网络短期发电量预测模型。以相同日类型条件下的光伏系统发电量、环境温度、光板温度、相对湿度的历史数据作为样本,对模型进行训练和发电量预测。通过小波神经网络模型和BP神经网络模型预测结果的对比分析表明:小波神经网络模型训练次数少,收敛速度快,预测精度高。

The various factors that affect the power generation output of photovoltaic（PV） system are studied,and a model of generation forecasting of PV system based on wavelet neural networks（WNN） is built.The historical data such as generation output,ambient temperature,solar panel temperature and relative humidity under similar weather conditions are taken together as samples to train the model and to forecast power generation output.In this paper,the WNN model and BP neural networks model are compared and analyzed,the results show： the WNN model has fewer training times,faster convergence rate and higher prediction accuracy comparing with BP neural networks model.