基于模糊RBF神经网络的高压断路器机构故障诊断

为了快速准确诊断高压断路器是否发生操作机构故障,文章提出了一种基于模糊RBF神经网络的高压断路器机构故障诊断方法。该方法首先分析高压断路器的分(合)闸线圈电流,提取时间和电流特征参数t1、t2、t3、t4、t5、I1、I2、I3,然后在RBF神经网络增加模糊化层,对特征参数进行相对模糊化运算,最后将模糊化后的特征参数输入到RBF神经网络进行故障识别、分类。该方法以ABBVD4高压断路器的88组实验数据为训练样本建立4种高压断路器操作机构的模糊RBF神经网络故障诊断模型,12组测试样本来验证其准确性,实验结果显示,模糊RBF神经网络高压断路器的故障诊断模型能够准确的诊断出故障类型,其准确率达到99%,具有良好的实用性。与基于模糊BP神经网络的故障诊断方法相比,该方法收敛速度快,训练时间短,均方差较小,为0.1073。

Fault Diagnosis of Operating Mechanism of High-voltage Circuit Breaker Based on Fuzzy RBF Neural Network

For rapid fault diagnosis of operating mechanism of high-voltage circuit breaker,a fault diagnosis method based on fuzzy RBF neural network was presented. In this method,the feature parameters t1,t2,t3,t4,t5,I1,I2 and I3 were extracted by analyzing the opening/closing coil current of high-voltage circuit breaker. Then,a fuzzification layer was added in the RBF neural network to fuzzify the characteristic parameters. Finally,these fuzzified characteristic parameters were input to the RBF neural network for fault identification and classification. Eighty-eight groups of experimental data of ABB VD4 circuit breaker were taken as the training samples to establish a fault diagnosis model of the circuit breaker's operating mechanism based on fuzzy RBF neural network for 4 kinds of fault types,and 12 groups of test samples were used to verify the validation of the proposed model. The results show that the model is valid and the accuracy of fault diagnosis reaches 99%. Compared with the fault diagnosis method based on fuzzy BP neural network,the proposed fault diagnosis method based on fuzzy RBF neural network has faster convergence speed,shorter training time,and smaller mean square deviation(0.107 3).