基于LSTM神经网络的风力发电量短期预测研究

Research on Short Term Prediction of Wind Power GenerationBased on LSTM Neural Network

基于LSTM神经网络的风力发电短期预测方法和传统的风力发电短期预测方法相比,其更贴近风力发电的实测值。文章首先采集需要的风力发电量数据,再对采集的风力发电量数据进行预处理,提高数据的可用性,在得到更高质量的风力发电量数据后,构建对风力发电量进行短期预测的LSTM短期预测模型。最后根据LSTM短期预测模型,生成相关的预测结果,完成基于LSTM神经网络的风力发电量短期预测研究,再进行基于LSTM神经网络的风力发电量短期预测试验。试验结果表明,基于LSTM神经网络的风力发电量短期预测精确度高。

Compared with traditional wind power short-term prediction methods,the LSTM neural networkbased wind power short-term prediction method is closer to the measured values of wind power generation.The article first collects the required wind power generation data,and then preprocesses the collected wind power generation data to improve data availability.After obtaining higher quality wind power generation data,an LSTM short-term prediction model is constructed for short-term prediction of wind power generation.Finally,based on the LSTM short-term prediction model,generate relevant prediction results,complete the research on short-term wind power generation prediction based on LSTM neural network,and then conduct short-term wind power generation prediction experiments based on LSTM neural network.The experimental results show that the short-term prediction accuracy of wind power generation based on LSTM neural network is high.