基于改进细菌算法轨道电路室外设备故障分析  

Outdoor Equipment Fault Detection Based on Improved Bacterial Algorithm

作　　者：邢东峰 薛顺然 陈光武[1,2] 石建强 XING Dong-feng;XUE Shun-ran;CHENG Guang-wu;SHI Jian-qiang(Institute of Automatic Control,Lanzhou Jiaotong University,Lanzhou Gansu 730070,China;Key Laboratory of Gansu Province Plateau Traffic Information Engineering and Control,Lanzhou Gansu 730070,China)

机构地区：[1]兰州交通大学自动控制研究所,甘肃兰州730070 [2]甘肃省高原交通信息工程及控制重点实验室,甘肃兰州730070

出　　处：《计算机仿真》2021年第8期157-161,共5页Computer Simulation

摘　　要：故障检测是近年来的热门话题,大数据环境下,原有算法在计算量和效率上已不能满足当前的需求,智能算法应用越来越广泛,将其应用于故障检测有较好的效果,细菌算法是新兴的全局优化智能算法,但其原有算法有一定的局限性。将细菌算法结合自适应理论,优化了细菌算法中趋化步长和繁殖步骤,对比传统BFO最优值有7个数量级的提高,在收敛性方面提高了一倍,并且将其与贝叶斯网络结构学习相结合解决轨道电路故障诊断的问题。Fault detection is a hot topic in recent years.Under the big data environment,the original algorithm can not meet the current demand in terms of calculation amount and efficiency.The application of intelligent algorithms is more and more extensive,and it has a good effect on fault detection.The bacterial algorithm is an emerging global optimization intelligent algorithm,but its original algorithm has certain limitations.In this paper,the bacterial algorithm is combined with adaptive theory to optimize the chemotaxis step and breeding steps in the bacterial algorithm.Compared with the traditional BFO,the optimal value is improved by 7 orders of magnitude,and the convergence speed is doubled.Moreover,it is combined with Bayesian network structure learning to solve the problem of track circuit fault diagnosis.

关 键 词：改进细菌算法 贝叶斯网络 自适应理论 轨道电路 粒子群优化 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术] TP284.2[自动化与计算机技术—计算机科学与技术]