Study On The Data Fusion Technology Based On Neural Network And D-S Inference In The Field Of The Mechanical Fault Diagnosis

This article is mainly about the mechanical equipment fault diagnostic method.Based on analysing the currently domestic and foreign sutdy,it gives a new diagnostic method based on neural net and D-S inference.The emulator experiment have done in the fault diagnosis of rotary system.Firstly,The definition and characteristic of mechanical fault are gived,the diagnostic complexity and necessity are discussed.Studying the traditional diagnostic methods and theirs advantages and shortcomings,the study and application of new method is urgent required in the further complex mechanical equipment.Secondly,started with neuron mode and transfer function,we discuss the networ struction and the function mapping,the characteristic of BP net,it shows that neural net can be used for the fault diagnosis.Eventually the momentum method of BP improved algorithm is selected for remedying the defect of the traditional BP algorithm.The diagnostic emulator experiment of single subnet is done by Matlab software.Beginning with the basic conception and merging rules of D-S inference,it shows that D-S inference can be used for information fusion.By analysing example,it shows D-S inference have accumulative effects of the evidential credence.Finally,we design the diagnostic method based on neural net and D-S inference.Started with the characteristics of the obtained information from the waiting diagnostic system,the three symptom field of fault characteristics are divided,the individual fault diagnosis is preliminarily done in the three subnet.the diagnostic result from three subnet are converted to basic probability assignment.Information fusion is token in time and spatial domain by D-S inference.By emulator experiment,the feasibility and availability of this diagnostic method are verified,it can increase the accuracy of fault diagnosis.