Making of Web-Based Software for Data Operation Analysis of a Diesel Engine Using Principal Component Method Case Study : Diesel Engine SWD 16TM410 in PLTD Tarakan

• Main Author: Khairul Arifin, Fariz
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/33822
• Institution: Institut Teknologi Bandung

Good maintenance can prevent the failure of diesel engine which can cause the industrial processes cease. Abnormal symptoms of the diesel engine can occur before a variable passes its limit value that issued by the manufacturer. This condition must be hard for the maintenance operator to detect these abnormalities. In this study, a software developed for detecting these abnormalities in the operation data using Principal Component Analysis (PCA) method. The software in this study is using the Moving Window PCA (PCA) algorithm to analyze 43 operation variables consist of fuel rack position, exhaust temperature, turbocharger, lube oil, and cooling water. In addition, this study also discussed about the differences between the two types of PCA models. First model is the model with genuine data from the actual measurement and second model is the model with data from the actual measurement with modification in several variables which detected as abnormal variables in previous measurement. The results obtained from these two models show that these two models can detect the abnormalities of operation data at some point of maintenance activities. However, the two model cannot predict the failure early, so the model needs further development.