Fault Diagnosis Method for Mechanical Rotor Systems using ANN

otating machinery is an integral part in majority of industries. In rotating machineries, faults are inevitable due to the errors in manufacturing, errors while assembling different parts of the system and due to different operating conditions such as heat generation, looseness, wear, etc. Hence, rotating machinery needs to be monitored continuously for identifying the faults. Any defect in the parts of the rotating machinery will affect its vibration behavior and nature of this effect is different for different faults. Hence condition monitoring based on vibration measurements can be used to identify those defects qualitatively. The current study mainly concentrated on comparing the performances of Standalone Artificial Neural Network and Genetic Algorithm based ANN in fault diagnosis of rotating machineries and developing a Fault Detection Program (FDP) for identifying different fault conditions. Vibration signals corresponding to each fault conditions were recorded from an experimental set up by means of a Lab view data acquisition system. The statistical features of vibration signals were extracted using a feature extraction code and it was given as the input data to train the ANN. From the study, it is concluded that GA based ANN is a better choice for fault diagnosis compared to Standalone ANN.