动车组制动风管及接头早期泄漏的超声能量特征辨识研究

Research on Ultrasonic Energy Feature Identification on Early Leakage of Brake Ducts and Joints of EMUs

目前通过试风实验台检测动车组制动系统的气密性,能够实现制动风管及接头严重泄漏评估与判断,为了检测和判断其早期泄漏,提出一种制动风管早期泄漏的超声能量特征辨识方法。首先检测管道及接头因泄漏而产生的超声信号,接着利用带通滤波器组对超声信号进行分解,计算分解信号的能量及能量熵,利用能量熵识别制动风管及接头是否存在泄漏,进一步利用能量算子特征驱动支持向量机识别风管材质。建立泄漏试验台,采集钢管接头泄漏、软管接头泄漏、无泄漏的超声信号对提出的方法进行了试验验证,结果表明:能量熵能显著区分有无早期泄漏,制动风管泄漏识别准确率达到98%以上。现场实测数据表明:本文提出的方法能在车辆实际应用环境下实现泄漏的可靠辨识。文章所提算法运算简单、实时性高,具有一定的工程应用价值。

At present,serious leakage of brake ducts and joints can be assessed and detected by testing the air tightness of the braking system of EMUs with a test air test bench.In order to detect and judge its early leakage,this paper proposes an ultrasonic energy feature identification method for early leakage of brake ducts.The ultrasonic signal generated by the ducts and joints due to leaks is first detected,then the ultrasonic signal is decomposed using a band-pass filter set,and the energy and energy entropy of the decomposed signal are calculated.The energy entropy can be used to identify the presence of leaks in brake ducts and joints,and further energy operator feature drive is used to support vector machines to identify duct materials.A leakage test bench was established,and the proposed method was verified by collecting ultrasonic signals of steel pipe joint leaks,hose joint leaks,and no leaks.The results show that the energy entropy can significantly distinguish the presence or absence of early leaks,and the accuracy of braking duct leakage identification reaches more than 98%.The field measured data show that the proposed method can achieve reliable identification of leakage in the actual application environment of vehicles.The algorithm proposed in this paper is simple and real-time,and has certain engineering application value.