Neural Network Based Sensorless Diagnosis of Leakage Fault in a Hydraulic Pump
It is a crucial issue to perform effective condition monitoring and fault diagnosis of pump in a hydraulic system. A sensorless diagnosis of leakage fault in a submerged pump on the basis of neural network pattern recognition is proposed. In this study, the current of pump driving electrical motor is utilized as the diagnosis signal. The relationship between motor current and pump leakage is discussed first. Then the diagnosis strategy using intelligent pattern recognition technique, particularly neural network, is studied. Moreover, the experimental verifications are conducted to validate the proposed diagnosis method. Theoretical analysis and experimental verifications confirmed that the leakage fault of a pump can be detected in a cost-effective and noninvasive way by employing the current of the driving electrical motor of pump as the diagnosis signal.
Hydraulic pump neural network motor current
Shi Hongyan Hu Hong Xu Limei
Shenzhen Graduate School, Harbin Institute of Technology, Shenzhen, 518055, China
国际会议
北戴河
英文
103-105
2007-06-06（万方平台首次上网日期，不代表论文的发表时间）