基于D-S证据理论的直流电机故障诊断研究

多传感器、多手段检测的信息融合技术在电机故障诊断中的应用日益广泛。提出了一种基于D-S证据理论和BP算法的故障诊断方法,给出了基于BP神经网络和D-S证据理论的多传感器数据融合技术的直流电机故障诊断方法,首先通过分析直流电机的故障机理,得到电机在不同故障情况下表现的信号特征,然后利用多源信息间的冗余性和互补性,有效提取故障特征信息。将该方法应用于某型船直流电机的故障诊断,结果表明,该算法提高了故障诊断的可靠性和灵敏度。

Fault Diagnosis of DC Motor Based on D-S Evidence Theory

The Evidence Theory with multi-sensors and multi-means has been extensively applied to the fault diagnosis.A new algorithm for DC motor＇s fault diagnosis based on neural BP network and D-S Evidence Theory was proposed.Through an analysis of fault events in DC motor,the state features of the DC motor at different fault conditions（e.g.electronic circuit,rev and torque） were obtained.Then based on the features of multi-source signals in terms of redundancy and complementary,the diagnosis information was effectively extracted.The method was applied to a ship for DC motor＇s fault diagnosis.The results show that the approach improves the reliability and sensitivity of the fault diagnosis.