Fault Classification of Pump Using Support Vector Machine (SVM) Method

Anindita Adikaputri Vinaya
Engineering Management Department
Universitas Internasional Semen Indonesia
Gresik, Indonesia
anindita.vinaya@uisi.ac.id

Abstract

A machine is a mechanical or electrical equipment using the working principle of converting energy to assist human activities or produce certain products. The condition of a machine should be maintained and monitored in a good condition. Therefore, the condition of the machine needs to be detected before serious damage occurs. This study aims to detect the type of pump damage with a machine learning approach. The object of this study was water pump with Panasonic GP-129 type. The goal in this study is to classify three fault of pump conditions, those are misalignment, unbalance, and bearing fault. Based on the results obtained, the classification of pump fault using SVM methods had average accuracy of 98.35% on the Linear SVM and Cubic SVM models, and average accuracy of 100% on the Quadratic SVM model

Keywords: SVM, Pump, Vector

Full Paper: Download Full Paper

Plagiarism Check: Download Check Plagiarism

Peer Review: Download Peer Review