基于支持向量机回归算法的电子机械制动传感器系统故障诊断

Fault detection and diagnosis of EMB sensor system based on SVR

在介绍电子机械制动(EMB)系统总体结构的基础上,采用支持向量机回归(SVR)方法针对EMB传感器系统进行故障诊断研究。首先应用支持向量机回归算法建立EMB传感器故障预测模型,通过电流、压力及转速传感器在空间和时间上的冗余信息产生残差,利用残差阈值进行故障诊断,最后进行试验验证。试验结果表明,该算法可以对EMB系统的传感器进行有效的故障诊断,不需要考虑系统的精确模型,适用于EMB这种复杂的机电系统。

In this paper the overall structure of Electromechanieal Brake （EMB） is introduced; and the Support Vector Regression （SVR） is employed in the fault detection and diagnosis of the EMB sensors. First, the fault prediction model of the EMB sensors is built using SVR algorithm. Then, the residual sequences are generated from the redundancy information of the current sensor, force sensor and rotate speed sensor. Finally, the method is verified by experiments. Experiment results show that the proposed SVR algorithm can effectively detect the fault of the EMB sensor system that does consider the precise model of the system. The algorithm is applicable to complicated EMB systems.