Aiming at the existing defects of evidence dempster-shafer theory (DST) in dealing with high conflict evidence, we proposed a new method to improve DST. By introducing concept of fuzzy consistent matrix, calculate the weights of factors, and put different sources of evidence into distinguish, and finally cast more than one vote to prevent the phenomenon, the average convergence of evidence. What’s more, the improved DST new method is applied to the rolling bearing fault diagnosis of CNC machine workbench .The test results show that the improved new synthetic formula increases the accuracy of fault diagnosis Ball, the conflict of evidence synthesis results better, to achieve better results.
Fault Diagnosis Based on Non-Negative Sparse Constrained Deep Neural Networks and Dempster–Shafer Theory
