基于残差信息的汽车液压主动悬架故障诊断与隔离研究

为提高汽车液压主动悬架控制性能,提出一种针对液压作动器、车身垂向加速度传感器的故障诊断与隔离方法。建立四自由度汽车液压主动悬架动力学模型,并考虑液压作动器和传感器常见故障,建立悬架故障模型。采用极点配置方法设计故障检测滤波器,利用输出残差信息对作动器卡死和增益故障、传感器增益故障进行诊断。在此基础上,为准确诊断故障发生部位,利用相关系数和故障向量径长值对故障进行隔离。在MATLAB/Simulink环境下对故障悬架进行大量仿真计算,结果表明:残差波动超过一定阈值时可敏感地诊断出作动器或传感器故障;汽车液压主动悬架作动器故障采用最大相关系数准确隔离,车身垂向加速度传感器故障采用最大相关系数和最小故障向量径长值准确隔离。

Vehicle Hydraulic Active Suspension Fault Diagnosis and Isolation Based on Residual Error Information

To improve hydraulic vehicle active suspension system control performance,the hydraulic actuators and the car vertical acceleration sensor fault diagnoses and isolation method were put forward.The half-vehicle four degree-of-freedom hydraulic active suspension system dynamical model was firstly built,considering ordinary fault of hydraulic actuator and sensor,the suspension fault models were then built.A fault detection filter was designed by pole placement,and the actuators’ deadlocking and gain faults,the sensors’ gain fault were diagnosed through the residual error information.To fix the fault place accurately,the faults were isolated by the correlation coefficients and fault vector length values.Finally,under the MATLAB/Simulink environment,the considerable simulations were carried out to the fault suspension systems.The results show that the actuators or the sensors faults can be diagnosed sensitively when the residual error fluctuation exceeds a threshold;the vehicle hydraulic active suspension actuators’ faults can be isolated accurately by the maximal correlation coefficient,and the car vertical acceleration sensor fault can be isolated accurately by the maximal correlation coefficient and the minimal fault vector length value.