Information correlation entropy based D-S evidence theory used in fault diagnosis

Abstract

Dempster-Shafer (D-S) evidence theory based multi-sensor information fusion (MSIF) plays an important role in fault diagnosis. Aiming to solve the problems via classical evidence theory, an improved D-S evidence theory through the introduction of information correlation entropy theory is reported in this paper. Then, the proposed method is employed to gearbox fault diagnosis. Experiment analysis results indicate that the new method is effective for MSIF.