A method of multi-information fusion for fault diagnosis of large gearboxes in mining

摘要

The method of multi-formation fusion and its application in monitoring and fault diagnosis of large mining gearboxes are studied in this paper. In this application the structural characteristics of the belief function allow for its simplification and development of application specific functions. Based on simple belief function and using statistical method to create evidence model, a new simplified combination algorithm is inferred, different from the traditional Dempster's rule. It has been proven that this new algorithm is effective and practical for oil analysis in the field of fault diagnosis.