Research of acquisition and prediction method in early weak information of locomotive traction system

Abstract

Many key parts of mechanical equipment gradually enter failure period with service, and failures happen in some of them, it will lead to serious consequences with economic loss and crash of machine and death of human if weak failures cannot be identified in time, which might degrade to major failure. It would be of benefit to prevent the fault deterioration if faults could be identified in early period, with the states of key parts obtained. Locomotive gear was chosen as the research object here, a method of inhibition mode mixing and characteristic frequency acquisition based on dynamic cascade empirical mode decomposition is put forward for the weakness and aliasing out of uncertain factors. The fault features are much more obviously presented in different time scale by optimal intrinsic mode function (IMF) acquisition and the improvement of envelope solution method. The results of experiment show that weak faults can be obtained under much noise, and an effective method is provided for early fault prognosis and diagnosis.

Keywords

• locomotives
• feature extraction
• fault diagnosis
• gears
• different time scale
• fault features
• uncertain factors
• dynamic cascade empirical mode decomposition
• characteristic frequency acquisition
• inhibition mode mixing
• locomotive gear
• early period
• fault deterioration
• weak failures
• crash
• economic loss
• failure period
• mechanical equipment
• locomotive traction system
• early weak information
• prediction method
• early fault prognosis
• weak faults
• envelope solution method
• optimal intrinsic mode function acquisition