Research on Photovoltaic Power Generation Prediction Model Based on OOLSSA-LSTM
姚明宇1
扫码查看
点击上方二维码区域，可以放大扫码查看
作者信息
1. 华北水利水电大学,郑州 450046
折叠
摘要
预测精确度一直是负荷预测的重点研究方向.提出一种利用正交对立学习改进的麻雀搜索算法[An Improved Sparrow Search algorithm Based on orthogonal-opposition-based learning(OOLSSA)]与长短期记忆网络(Long Short Term Memory network)组成的OOLSSA-LSTM混合预测模型,利用正交对立学习对麻雀算法进行优化,对LSTM参数进行选择,避免人为选择造成的误差.通过实验对比单一的LSTM预测模型、SSA-LSTM预测模型及OOLSSA-LSTM预测模型的预测结果,得到的实验结果满足预期,证明此改进算法具有更好的寻优结果,为功率预测提供一个新方案.
Abstract
Forecasting accuracy is always the key research direction of load forecasting.The study proposes a hybrid prediction model composed of an Improved Sparrow Search algorithm Based on orthogonal-opposition-based learning(OOLSSA-LSTM)and Long Short Term Memory network uses orthogonal opposition learning,optimizes Sparrow algorithm by orthogonal opposition learning,and selects LSTM parameters to avoid errors caused by human selection.By comparing the prediction results of the single LSTM prediction model,SSA-LSTM prediction model and OOLSSA-LSTM prediction model,the experimental results meet the expectations,proving that the improved algorithm has better optimization results.This provides a new scheme for power prediction.
关键词
正交对立学习/麻雀搜索算法/OOLSSA-LSTM预测模型
Key words
Orthogonal opposition learning/Sparrow search algorithm/OOLSSA-LSTM prediction model
维普
万方数据