基于RST-SOM的高压断路器故障诊断

Fault Diagnosis of High-voltage Circuit Breaker Based on RST-SOM

由于高压断路器运行过程中产生的状态数据庞大,传统的基于人工神经网络的高压断路器故障诊断方法在针对这一问题时存在网络结构复杂、训练过程费时、诊断速率缓慢的缺点。由此,文中提出RST粗糙集结合SOM自组织特征映射网络的方法,通过RST理论对断路器故障数据中的各个属性进行评价并寻找最小属性集,以此消除特征信息中存在的冗余属性,得到约简决策表,并将新形成的故障特征数据作为输入结合自组织特征映射网络进行高压断路器故障诊断。经过验证,在确保整体准确率能够达到91%的情况下,缩短了训练时间,简化了网络结构,在工程实践应用中表现良好。

The traditional fault diagnosis method of high-voltage circuit breaker based on artificial neural network has a shortcoming that the complex of network structure because of the high-voltage circuit breaker fault feature information is too large,and in the previously proposed self-organizing feature map network to verify the fault diagnosis also exists when the training process is time-consuming,slow diagnosis of the shortcomings.So,a rough sets theory combined with self-organizing feature map network has proposed in this paper,firstly,the attributes of the fault data are evaluated by the rough set theory and the minimum attribute set is found,the redundant attributes in the feature information are eliminated,the reduction decision table is obtained,the new fault characteristic data is formed,the new fault characteristic data is also as the input of self-organizing feature map.The new method ensure that the overall accuracy rate achieved 91%,and simplifying the network structure,improve the running time and diagnostic rate after verification,so this algorithm can be effectively implemented in fault diagnosis.