Fault diagnosis of main engine journal bearing based on vibration analysis using Fisher linear discriminant, K-nearest neighbor and support vector machine

Vibration technique in a machine condition monitoring provides useful reliable information, bringing significant cost benefits to industry. By comparing the signals of a machine running in normal and faulty conditions, detection of defected journal bearings is possible. This paper presents fault diagnosis of a journal bearing based on vibration analysis using three classifiers: Fisher Linear Discriminant (FLD), K-Nearest Neighbor (KNN) and Support Vector Machine (SVM). The frequency-domain vibration signals of an internal combustion engine with intact and defective main journal bearings were obtained. 30 features were extracted by using statistical and vibration parameters. These features were used as inputs to the classifiers. Two different solution methods - variable K value and RBF kernel width (σ) were applied for FLD, KNN and SVM, respectively, in order to achieve the best accuracy. Finally, performance of the three classifiers was calculated in journal bearing fault diagnosis. The results demonstrated that the performance of SVM was significantly better in comparison to FLD and KNN. Also the results confirmed the potential of this procedure in fault diagnosis of journal bearings.