基于NGWO的光伏信息设备的状态预测

State Prediction of Photovoltaic Information Equipment Based on NGWO

光伏信息设备是光伏监控系统中的一个重要组成部分,为对其状态进行有效预测以更好地确保系统正常运转,提出了一种小生境灰狼优化(NGWO)的信息设备状态预测方法。该方法首先构建基于长短期记忆神经网络(LSTM)的光伏信息设备的状态预测模型,接着利用NGWO算法确定出LSTM的网络参数,NGWO算法具有较快的收敛速度和全局寻优的能力,保证了模型对信息设备状态的准确预测。实验结果表明,该方法能够很好地表征光伏信息设备状态的变化规律,且具有较好的泛化能力和预测精度。

Photovoltaic information equipment is an important part of the PV monitoring system.To effectively predict its state to better ensure the normal operation of the system,a niche grey wolf optimization(NGWO)information equipment state prediction method is proposed.The method first constructs a state prediction model of photovoltaic information equipment based on long-term and short-term memory neural network(LSTM),and then uses NGWO algorithm to determine the network parameters of LSTM.NGWO algorithm has faster convergence speed and global optimization ability,which guarantees the model accurately predicts the state of the information equipment.The experimental results show that the proposed method can well characterize the changing state of PV information equipment,and has good generalization ability and prediction accuracy.