Fault Classification in Cylinders Using Multilayer Perceptrons, Support Vector Machines and Guassian Mixture Models

TL;DR

This paper compares Gaussian mixture models, support vector machines, and multilayer perceptrons for fault classification in cylindrical shells, demonstrating that GMM achieves the highest accuracy among the three methods.

Contribution

The study introduces and evaluates GMM, SVM, and MLP for fault classification in cylindrical shells, highlighting the superior performance of GMM.

Findings

GMM achieves 98% accuracy

SVM achieves 94% accuracy

MLP achieves 88% accuracy

Abstract

Gaussian mixture models (GMM) and support vector machines (SVM) are introduced to classify faults in a population of cylindrical shells. The proposed procedures are tested on a population of 20 cylindrical shells and their performance is compared to the procedure, which uses multi-layer perceptrons (MLP). The modal properties extracted from vibration data are used to train the GMM, SVM and MLP. It is observed that the GMM produces 98%, SVM produces 94% classification accuracy while the MLP produces 88% classification rates.