Automatic defect analysis of pumps using Adaptive Neuro- Fuzzy Inference System and Vibrational Features

Centrifugal pumps have a crucial role in many critical applications. Hence, their continuous availability is vital. Th is paper concentrates on vibrational-based condition monitoring and fault diagnosis of such pumps. Th e vibrational-based machine condition monitoring and fault diagnosis include several machinery fault detection and diagnostic techniques. Many machinery fault diagnostic techniques use automatic signal classification to increase accuracy and reduce errors made by human judgments. Th is paper presents an adaptive network fuzzy inference system (ANFIS) in an attempt to diagnose the pump fault type. Th e pump conditions in question included healthy, misalignment and three different bearing faults (Inner cage, outer cage and ball damage). Th ese features are taken from realistic vibrational signals by employing the wavelet transform technique. As input vectors, the features were put into an adaptive neuro-fuzzy inference system. Th e performance of the system was validated by applying the testing data set to the trained ANFIS model in order to detect different system conditions. With regard to the results, the total classification accuracy of the trained ANFIS is about 90.33 % for all faults, indicating that the system has a high potential to be used as an intelligent fault diagnosis system in real applications.