基于DBSCAN与BP神经网络的轮速传感器故障诊断研究  

Study on fault diagnosis of intelligent wheel speed sensor based on DBSCAN and BP neural network

作　　者：郭箤 田锦明 张军 谢春旭 纪林海 何胜 田昊东 姚林燃 GUO Zu;TIAN Jinming;ZHANG Jun;XIE Chunxu;JI Linhai;HE Sheng;TIAN Haodong;YAO Linran(School of Electronic Engineering,Jiangsu Ocean University,Jiangsu Lianyungang 222005,China;Continental Automotive Electronics(Lianyungang)Co.,Ltd.,Jiangsu Lianyungang 222006,China)

机构地区：[1]江苏海洋大学电子工程学院,江苏连云港222005 [2]大陆汽车电子(连云港)有限公司,江苏连云港222006

出　　处：《工业仪表与自动化装置》2024年第6期99-103,共5页Industrial Instrumentation & Automation

摘　　要：针对工业生产时轮速传感器性能测试数据过多,难以对其故障类型进行识别的问题,设计了基于DBSCAN与BP神经网络的轮速传感器故障诊断方法。首先,根据轮速传感器工作原理以及实际测试数据分析了轮数传感器故障类型以及对应的故障数据。然后,利用DBSCAN对轮速传感器的性能测试数据进行异常值检测,同时建立BP神经网络进行训练及测试,用于对异常值对应的故障类型进行诊断与分类,并将BP神经网络与GRNN神经网络以及PNN神经网络的故障诊断速度以及准确率进行对比。实验结果显示:针对轮速传感器故障类型检测BP神经网络的速度及准确率有明显优势,该文设计的轮速传感器故障检测算法能够准确的从测试数据中提取故障数据并进行故障诊断。In view of the problem that the performance test data of the wheel speed sensor is difficult to identify the failure type in industrial production,the wheel speed sensor fault diagnosis method based on DBSCAN and BP neural network is designed.Firstly,according to the working principle of wheel speed sensor and the actual test data,the fault type of wheel number sensor and the corresponding fault data are analyzed.Then,DBSCAN is used to detect the performance test data of the wheel speed sensor,and to train and test the BP neural network,to diagnose and classify the fault types corresponding to the abnormal values,and to compare the fault diagnosis speed and accuracy of BP neural network with GRNN neural network and PNN neural network.The experimental results show that the wheel speed sensor fault type has obvious advantages.The wheel speed sensor fault detection algorithm designed in this paper can accurately extract the fault data from the test data and diagnose it.

关 键 词：轮速传感器 故障诊断 DBSCAN BP神经网络 

分 类 号：TP3063.4[自动化与计算机技术—计算机系统结构]