Identification of Cavitation phenomenon in centrifugal pump by Artificial Immune Network method

Reduce the cost of unscheduled shutdown and enhance the reliability of systems, is one of the important goals for various industries that could be achieved by condition monitoring. Cavitation is a common phenomenon in centrifugal pumps which causes the damage and its true identification in early stage is too important and help to increase the life of pump. In this paper cavitation is identified by use of Artificial Immune Net (AIN) that is modeled on the function of the human immune system. For this purpose, first data collection were done by a laboratory setup and various features extracted form vibration and current signals, in next step, feature selection and dimensions reduction were done by artificial immune method, then with AIN method, the system condition was identified. Finally, for comparing the results of this study with other methods, first feature selection and dimensions reduction were down by the PCA method then fault detection were down by nonlinear supportive vector machine (SVM), multi-layer artificial neural network (MLP), K-means and FCM methods with features of PCA and AIN methods. The results shown that AIN methods have more precise than other methods.