A comparative study of sensor fault diagnosis methods based on observer for ECAS system

作     者：Xu, Xing Wang, Wei Zou, Nannan Chen, Long Cui, Xiaoli 

作者机构：Jiangsu Univ Automot Engn Res Inst Zhenjiang 212013 Peoples R China Jiangsu Univ Sch Automot & Traff Engn Zhenjiang 212013 Peoples R China Hunan Inst Technol Sch Mech Engn Hengyang 412002 Peoples R China 

出 版 物：《MECHANICAL SYSTEMS AND SIGNAL PROCESSING》 (机械系统与信号处理)

年 卷 期：2017年第87卷第PartB期

页      面：169-183页

核心收录：

学科分类：08[工学] 0802[工学-机械工程] 

基　　金：Natural Science Foundation of Jiangsu Province, China [BK20131255] "Qing Lan" Project of Jiangsu Province National Natural Science Foundation of China Natural Science Foundation of Hunan Province, China [12JJ6036] 

主　　题：Electronically controlled air suspension (ECAS) Sensor faults Fault detection and isolation (FDI) Algorithms comparison 

摘      要：The performance and practicality of electronically controlled air suspension (ECAS) system are highly dependent on the state information supplied by kinds of sensors, but faults of sensors occur frequently. Based on a non-linearized 3-DOF 1/4 vehicle model, different methods of fault detection and isolation (FDI) are used to diagnose the sensor faults for ECAS system. The considered approaches include an extended Kalman filter (EKF) with concise algorithm, a strong tracking filter (STF) with robust tracking ability, and the cubature Kalman filter (CKF) with numerical precision. We propose three filters of EKF, STF, and CKF to design a state observer of ECAS system under typical sensor faults and noise. Results show that three approaches can successfully detect and isolate faults respectively despite of the existence of environmental noise, FDI time delay and fault sensitivity of different algorithms are different, meanwhile, compared with EKF and STF, CKF method has best performing FDI of sensor faults for ECAS system. (C) 2016 Elsevier Ltd. All rights reserved.