基于改进免疫算法的工业通信网络控制器故障位置检测

Fault Location Detection of Industrial Communication Network Controller Based on Improved Immune Algorithm

以单一智能挖掘算法为主的控制器故障位置检测很容易陷入局部最优,导致检测难度增加,为此该文提出基于改进免疫算法的工业通信网络控制器故障位置检测方法。设计工业通信网络控制器故障告警关系二分图与关系边约束,建立控制器故障关联函数。采用小波包提取故障关联函数输出的高频数据特征,将其作为免疫系统的原始种群。通过改进免疫算法得出最优抗体,最优抗体所在位置即为故障位置,得到故障位置检测结果。实验结果表明,该方法可以有效缩短控制器故障位置检测耗时以及误检率,提升检测效果。

The detection of controller fault location,which is mainly based on a single intelligent mining algorithm,is prone to local optima and increases the difficulty of detection.To address this issue,a method for detecting the controller fault location in industrial communication networks based on an improved immune algorithm is proposed.The method involves designing a bipartite graph of the alarm relationship and relation edge constraints for the controller fault in the industrial communication network,and establishing the controller fault correlation function.The high-frequency data features outputted by the fault correlation function are extracted using wavelet packet decomposition and used as the original population of the immune system.By comparing and communicating the results obtained from both particle swarm optimization algorithm and immune algorithm,the optimal antibody is identified,and its position represents the fault location,thus obtaining the fault location detection result.Experimental results demonstrate that this method can effectively reduce the detection time and false alarm rate in controller fault location detection,and improve the detection performance.