基于径向基函数神经网络和模糊积分融合的电网分区故障诊断  

Divisional Fault Diagnosis of Power Grids Based on RBF Neural Network and Fuzzy Integral Fusion

作　　者：石东源[1] 熊国江[1] 陈金富[1] 李银红[1] 

机构地区：[1]强电磁工程与新技术国家重点实验室(华中科技大学),湖北省武汉市430074

出　　处：《中国电机工程学报》2014年第4期562-569,共8页Proceedings of the CSEE

基　　金：国家自然科学基金项目(50907024)~~

摘　　要：为有效解决分区故障诊断关于互连区域间联络线的诊断问题,提出了基于径向基函数神经网络和模糊积分融合的大电网故障诊断方法。该方法通过网络重叠分区将大电网划分为若干区域,故障发生后根据警报信息选择性触发警报信息所在区域对应的区域径向基函数神经网络诊断模块,然后利用模糊积分关联融合相连区域关于联络线的诊断输出,实现对联络线的故障诊断。该方法不仅可以诊断各区域内部发生的故障,而且能够有效地诊断区域间联络线发生的故障。算例仿真结果表明:该方法简单、有效,能弥补现有电网分区故障诊断方法在联络线故障诊断方面存在的不足,且能够处理各种复杂故障情况,具有良好的故障容错能力。This paper presented an effective method for fault diagnosis of large-scale power grids based on radial basis function（RBF） neural network and fuzzy integral fusion. The study aims at effectively solving the diagnosis problem about the tie lines connecting regional sub-grids in the divisional fault diagnosis scheme. An overlapping network division method was proposed to divide a large-scale power grid into several sub-grids. When faults occur, regional RBF neural network diagnostic modules corresponding to different sub-grids are selectively triggered according to local alarm information which implies the faults exist in the sub-grids. Then faults of tie lines can be diagnosed by applying fuzzy integral to fuse the diagnostic outputs of two connected sub-grids about the tie lines. The method can not only be efficient in diagnosing the faults within local regions, but also in diagnosing the faults of tie lines well. The simulation results show that the proposed method is simple, efficient and can make up for the shortcoming of existing divisional fault diagnosis methods in diagnosis of tie lines. Moreover, it can diagnose different complex faults with good fault tolerance capability.

关 键 词：大电网 电网分区 故障诊断 径向基函数神经网络 模糊积分 

分 类 号：TM71[电气工程—电力系统及自动化]