Study On Application Of Information Fusion In Fault Diagnosis System Of Traction Motor On Locomotive

Traction motors play a principal role in electric locomotive and its architecture is very complex, so its fault diagnosis is a hot topic and is very difficult in the fault diagnosis field of electric locomotive. Since the process of traction motors fault diagnosis has the same purpose and request with multi-sensor information fusion, the technique of information fusion apply in it is inevitable for the development of them.In this paper, a temporal-spatial architecture model is presented based on information fusion theory, which is deeply researched on the basis of former students linking to practical project and first applied in the fault diagnosis of traction motor successfully. In this model, the acquisition, processing and allocation of information, and the all-round diagnosis is included. Wavelet theory and time-domain theory are used in the processing of vibrant information for character information extraction.The paper has improved the contexture methodology of basic probability of D-S Reasoning when its application in information temporal-spatial fusion. The information that is monitored by a same sensor at different time is fused first by local diagnosis networks and then the data that is fused in time is fused in space order by whole decision-making network. After temporal-spatial information fusion, higher accurate and more reliable diagnosis result can be obtained and the loss of information is lesser than others.In the end, "Fault diagnosis system of traction motor on electric locomotive" based on the research of DJ₁ electric locomotive was designed and implemented. The system supplies effective support for preventive maintenance of the traction motor of DJ₁ electric locomotive according to practical application. At the same time, it is pre-researched and it must be modified and integrated before it reached the author's initial purpose.