FAILURE PREDICTION OF ENGINEERING PROBLEMS USING INTERACTIVE COMPUTING NOTEBOOK ENVIRONMENT

• Main Author: FLORINA LING, CASTELO
• Format: Final Year Project Report
• Language: English; English
• Published: Universiti Malaysia Sarawak, (UNIMAS) 2020
• Subjects: QA Mathematics
• Online Access: http://ir.unimas.my/id/eprint/34055/1/Florina%20Ling%20Anak%20Castelo%20-%2024%20pgs.pdf; http://ir.unimas.my/id/eprint/34055/4/Florina%20Ling%20Castelo%20ft.pdf; http://ir.unimas.my/id/eprint/34055/

In recent years, research has proposed several machine learning (ML) approaches to predict remaining useful life (RUL) in engineering field which involved computer science skills. This paper proposed to predict turbofan engine remaining useful life (RUL) based on the engine historical degradation data provided by NASA C-MAPPS. CMAPPS is a tool stands for ‘Commercial Modular Aero- Propulsion System Simulation’ to simulate realistic large commercial turbofan engine data. Prediction model built based on regression problem using dimensionality reduction method and regression algorithms. Dimensionality reduction would extract only important features for more accurate prediction. Model performance is dramatically affected by the algorithm robustness which are the basis of this thesis. The efficiency of the model is evaluated using Pearson correlation coefficient. Results showed regression model could give a satisfactory prediction result based on the test data provided by CMAPPS. The effectiveness of the methodology for early prediction provides alert in machine degradation before it reaches failure. This efficient procedure could prevent severe failure occurrence and maintenance costs.