Efficient decision making technique for industrial machine condition monitoring

Estimating the industrial machine system reliability is a challenging and important problematic task for the manufacturing engineers. Reliability may be described as the probability that machine network will implement its proposed functions under the observing condition throughout a specified time period of running machine system network. Currently, typical single sensor method is common in use for fault diagnosis which based on the signature of single/multiple parameters, but due to uncertainty of modelling environment; it is difficult to diagnose the machine faults at early stages. In this situation, sometimes decision goes wrong due to uncertainty and may cause loss of throughput and significant financial losses. Fault diagnosis has to build a relation between the machine fault symptoms and estimating the severity of the fault. An artificial intelligence (AI) technique is proposed in this research for machine fault diagnosis by using the extracted features, where accurate mathematical method is difficult to build. This paper focuses on Dempster Shafer (D-S) evidence theory to diagnose the machine faults. The results reflect that the D-S theory is very effectively applied for the machine fault diagnosis and it also increase the reliability of machine operation process and decrease the uncertainty level in decision making.