Publication

Abstract : Vibration analysis in gearbox condition monitoring is vital for ensuring the dependability, safety, and optimal performance of rotating machinery. Vibration analysis is one of the major techniques in the monitoring of mechanical components. In this study, the vibration data were taken for healthy and defective conditions (dry & wet conditions) using vibration sensors positioned at various locations. These datasets were imported to MATLAB for feature extraction and analysis. The gear fault-related features in the vibration signals are extracted by Empirical Mode Decomposition (EMD). Further, features like RMS, kurtosis, and skewness were extracted. The extracted features were classified in MATLAB classification algorithms like Support Vector Machine (SVM). Amongst different SVM algorithm for Dry condition the most accurate algorithm was medium Gaussian SVM with accuracy of 92.1% and weighted KNN with accuracy of 81.6%. For wet condition most appropriate algorithm were Medium Gaussian SVM with an accuracy of 94.4% and weighted KNN with accuracy of 89%. The confusion matrix plot will be used to distinguish between the classifiers.