Smart Fault Diagnosis of Rotating Machine Using Fuzzy Systems and Analysis the Effect of Different Membership Functions

In this paper, automatic fault diagnosis of rotating machines using vibrating data measured from different point of machines and an smart fuzzy knowledge-based systems is discussed. To this end, a new vibrations’ identification chart, recently published, is used. This vibration's identification chart contained frequency characteristics and phase angle and is represented for some usual defects such as unbalancy, misalignment, bent shaft and mechanical looseness. Designed fuzzy knowledge-based system has a very simple structure. It do not need any complicated training such as those are used for neural network training. To evaluate the performance of the designed fuzzy system in actual application, it is used for fault diagnosis of some rotating machines in Isfahan Steel Company such as Fans. The effect of different membership functions such as non-linear Gaussian, bell-shaped, sigmoid, s-shape and z-shape function for inputs and outputs of fuzzy rules database is investigated and.the results are compared with the results of some neural networks-based fault diagnosis systems. Results show the designed smart fuzzy system has acceptable performance in detecting fault.