用改进GA-BP神经网络的设备故障维修时间预测

Maintenance Time Prediction of Equipment Failure with Improved GA-BP Neural Network

光伏充电站中设备故障维修时间对运行效率有着重要影响。为提高光伏充电站设备维修时间的预测精度,考虑到神经网络算法中隐含层神经元数对算法预测精度的影响,提出了一种改进的GA-BP神经网络算法,并以光伏充电站60个设备维修时间为样本验证了改进算法的有效性。结果表明,GA-BP神经网络结构中隐含层神经元数取5时算法预测精度最高,且采用改进GA-BP神经网络算法预测时平均相对误差仅为6.1%,较灰色模型与BP神经网络算法分别降低了90.4%与57%。改进后的GA-BP神经网络的预测准确度远高于灰色模型和BP神经网络,得到的预测时间可为维修人员调度提供依据。

Maintenance time of equipment failure in photovoltaic charging stations has an important influence on operation efficiency.In order to improve the prediction accuracy of maintenance time,the effect of hidden layer neuron number on the prediction accuracy of the algorithm was considered and an improved GA-BP neural network was proposed.Taking 60 equipment maintenance time of a photovoltaic charging station as the specimens,the high prediction accuracy of improved GA-BP neural network has been proved.Results indicate that the GA-BP neural network has the highest prediction accuracy when the number of hidden layer neurons is 5.The average relative error of the improved GA-BP neural network between the predicted values and the expected values is 6.1%,decreasing 90.4% and 57% compared with gray model and BP neural network respectively.The prediction accuracy of the improved GA-BP neural network is much higher than that of gray model and BP neural network,and the predicted time can provide a basis for personnel scheduling.