基于AdaBoost算法的新能源汽车电机异常故障检测  

Abnormal Fault Detection of New Energy Vehicle Motor Based on AdaBoost Algorithm

作　　者：倪龙飞 白倩 张治斌[2] NI Long-fei;BAI Qian;ZHANG Zhi-bin(Huanghe Jiaotong University,School of Intelligent Engineering,Jiaozuo Henan 454950,China;Henan Polytechnic University,Jiaozuo Henan 454950,China)

机构地区：[1]黄河交通学院智能工程学院,河南焦作454950 [2]河南理工大学,河南焦作454950

出　　处：《计算机仿真》2024年第4期97-101,共5页Computer Simulation

基　　金：河南省科技攻关项目资助(232102241028);河南省工程技术研究中心(266)。

摘　　要：新能源汽车的电机系统包含许多复杂的部件和子系统,部件之间的相互作用使得异常故障的检测变得复杂,而电机异常故障检测主要采用人工检测方式,即通过耳朵听声音,用眼睛观察,用手触摸找出故障位置,导致故障检测精度较低。因此,提出AdaBoost算法下新能源汽车电机异常故障检测方法。通过传感器采集电机信号,采用距离相似度、模糊隶属度函数提取信号特征,借助遗传算法的编码操作、交叉操作及其变异操作获取关键信号特征,运用自适应增强(Adaptive Boosting,AdaBoost)算法将信号特征分成正常信号和异常故障,以此实现对新能源汽车电机异常故障检测。实验结果表明,所提算法电机异常故障检测精度高,且耗时短。The motor system of new energy vehicles contains many complex components and subsystems.The interaction between these components makes the detection of abnormal faults complex.At present,the detection of motor abnormal faults mainly adopts manual detection methods,so the detection accuracy is relatively low.This paper presented a method of detecting abnormal faults of new energy vehicle motors based on AdaBoost algorithm.Firstly,we collected motor signals through sensors.Then,we extracted signal features by distance similarity and fuzzy membership functions.Meanwhile,we used encoding operation,crossover operation,and mutation operation of the genetic algorithm to obtain key signal features.Finally,we used the Adaptive Boosting(AdaBoost)algorithm to divide signal features into normal signals and abnormal faults.Thus,we realized the detection of abnormal faults in new energy vehicle motors.Experimental results prove that the proposed algorithm has high accuracy in detecting motor abnormal faults and takes less time.

关 键 词：弱分类器 强分类器 遗传算法 新能源汽车 电机异常故障检测 

分 类 号：TP399[自动化与计算机技术—计算机应用技术] TM307[自动化与计算机技术—计算机科学与技术]