Publication

Abstract : A machine’s operation can be improved by early detection of anomalies and reducing downtime, which can savemoney. During the operation of a system, live measurements are taken and processed to alert people to any irregularities that could point to a newly developing flawwith the help of residual analysis. It can be used to identify out-of-the-ordinary machine behavior by detecting deterioration by recreating a broken machineand adapting to model variation in tracking devices. Residual analysis can be used to detect abnormal behavior in a system while building models of a broken machine can help to detect decay. Monitoring changes in the system through adaptive online model parameters can also help to identify potential faults before they become severe. In this paper the focus is given to predict the occurrence of the fault using SVM and Residual Analysis method. Comparative analysis between these two techniques is also done using MATLAB software.