基于模糊神经网络算法的继电保护系统故障诊断技术研究  

Research on Fault Diagnosis Technology of Relay Protection System Based on Fuzzy Neural Network Algorithm

作　　者：王禹沣 侯世昌 WANG Yufeng;HOU Shichang(Yanshan University,Qinhuangdao 066000,China;State Grid Qinhuangdao Power Supply Company,Qinhuangdao 066000,China)

机构地区：[1]燕山大学,河北秦皇岛066000 [2]国网秦皇岛供电公司,河北秦皇岛066000

出　　处：《电工技术》2025年第1期133-135,共3页Electric Engineering

摘　　要：在复杂的电力系统中,系统结构的复杂性和故障类型的多样性,导致故障可能发生在任何环节,增加了故障源检测的难度,因此研究了基于模糊神经网络算法的继电保护系统故障诊断技术。根据继电系统故障与保护动作的期望机制关系,构建继电保护系统故障解析模型。以继电系统的技术参数为变量,将故障状态诊断问题转化为继电系统解析模型最拟合目标函数的求解问题。通过深度学习算法中模糊神经网络确定最拟合目标函数的最优解,以此得到继电保护系统故障诊断结果。实验结果表明,通过设计方法诊断得到的电流互感器电流状态与设置情况在时间维度和程度维度均具有较高的可靠性。In complex power systems,due to the complexity of the system structure and the diversity of fault types,faults can occur in any link,which increases the difficulty of fault source detection.Therefore,research is needed on fault diagnosis technology for relay protection systems based on fuzzy neural network algorithms.Construct a relay protection system fault analysis model based on the expected mechanism relationship between relay system faults and protection actions.Using the technical parameters of the relay system as variables,the problem of fault state diagnosis is transformed into a problem of solving the optimal fitting objective function of the relay system analytical model.By using fuzzy neural networks in deep learning algorithms to determine the optimal solution that best fits the objective function,the fault diagnosis results of the relay protection system can be obtained.The experimental test results show that the current state and setting of the current transformer diagnosed by the design method have high reliability in both time and degree dimensions.

关 键 词：深度学习 继电保护 故障诊断 解析模型 最拟合目标函数 模糊神经网络 

分 类 号：TM71[电气工程—电力系统及自动化]