摘要
针对现有矿山直流微电网设备线损测试故障识别准确率和分类准确率较低等问题,提出一种煤矿直流微电网设备线损故障测试方法。采用最大重叠离散小波变换(Maximal Overlap Discrete Wavelet Transform,MODWT)法提取煤矿直流微电网设备线损故障特征,并联合反向传播神经网络(Back Propagation Neural Network,BPNN)和自适应遗传算法(Adaptive Genetic Algorithm,AGA)构建GA-BP神经网络,提高BPNN的全局寻优能力。对训练后的GA-BP神经网络模型进行优化,以测试煤矿直流微电网设备线损故障情况。实验结果表明,所提方法的故障识别准确率和分类准确率较高。
Aiming at the problems of low accuracy in fault identification and classification of existing mine DC microgrid equipment line loss test,this paper puts forward a line loss fault test method for mine DC microgrid equipment.Maximum Overlapping Discrete Wavelet Transform(MODWT)method is used to extract the line loss fault characteristics of coal mine DC microgrid equipment,and GA-BP neural network is constructed by combining Back Propagation Neural Network(BPNN)and Adaptive Genetic Algorithm(AGA)to improve the global optimization ability of BPNN.The trained GA-BP neural network model is optimized to test the line loss fault of coal mine DC microgrid equipment.The experimental results show that the fault identification accuracy and classification accuracy of the proposed method are high.
作者
李瑞龙
LI Ruilong(Shaanxi Yongxin Mining Co.,Ltd.,Yulin 719407,China)
出处
《通信电源技术》
2023年第21期94-96,100,共4页
Telecom Power Technology