@article { author = {Khalaj, Mehran and Khalaj, Fereshteh}, title = {Machine fault diagnosis through entropy weighting using evidential reasoning approach}, journal = {Journal of Industrial Engineering and Management Studies}, volume = {8}, number = {1}, pages = {114-134}, year = {2021}, publisher = {Iran Center for Management Studies}, issn = {2476-308X}, eissn = {2476-3098}, doi = {10.22116/jiems.2020.233227.1364}, abstract = {This paper presents an approach for the fault diagnosis in the state of fault in a machine by using a combination of the Dempster–Shafer (D-S) theory. At the first, feature extractions in each state have been combined based on evidential reasoning (ER) using kind of sensor information such as vibration, acoustic, pressure, and temperature, to detect and diagnose machine failure. Then, the main fusion will be obtained. In this process, the mass function assignment of any sensors to feature extraction, respectively, in every state of the machine is fused to indicate state quality. Within this framework, we propose a new way for main fusion to derive a consensus decision for fault diagnosis. In this paper, an approach developed to apply the evidential reasoning by defining adaptively weights into the improvement of the D–S evidence theory instead of the probability theory and the D–S evidence theory alone. Instead of using the evidential reasoning approach, this new approach applies entropy weighting in the D-S theory, in which all available data are used for making a decision. Entropy weighting can measure the uncertainty level of the fault decision and assist in obtaining a less uncertain fault decision. It is defined adaptively weights based on ambiguity measures associated with information obtained from each sensor. The ambiguity measure is defined by Shannon’s entropy. Many industries use old machines due to cost savings or lack of purchasing power. Maintenance policies in these factories are based on determining their fault experimentally and traditionally. Therefore, the main goal of this paper uses the improved evidence reasoning algorithm using a kind of sensor information to carry out fault diagnosis in these industrials. Then, a numerical example and a case study involving the ball mill machine in fault diagnosis are presented to show the rationality and efficiency of the proposed method.}, keywords = {evidential reasoning algorithm (ER algorithm),Information Fusion,Fault Diagnosis,multi-sensor fusion,Shannon Entropy}, url = {https://jiems.icms.ac.ir/article_133501.html}, eprint = {https://jiems.icms.ac.ir/article_133501_8731627b7b406706c9951ada1b5d4cd6.pdf} }