Session: 01-19 Aero-engine Control and Diagnostics II

Paper Number: 127544

127544 - Machine Learning Regression Models for Turbofan Engines: A Comparative Study on Remaining Useful Life Prediction 

In the field of Prognostics Health Management (PHM) in aviation, there's a pressing need to accurately predicting the Remaining Useful Life (RUL) of turbofan engines, especially at critical stages like the 'knee point' where degradation accelerates. While numerous regression models exist, their direct relevance, advantages, and challenges for this application remain under-investigated. This gives rise to a central research inquiry: Among the regression models, from linear and kernel-based to high-dimensional structures and Artificial Neural Networks, which are most effective for forecasting the RUL of turbofan engines, and how can their interpretability be optimized to ensure stakeholder trust? This question highlights a notable gap in research: a thorough comparative study focused on turbofan RUL prediction in aviation, particularly emphasizing model clarity and reliability. To address this, the study proposes a methodical examination of various regression models, gauging their fit, hurdles, and adaptability for RUL estimation. The methodology involves selecting pertinent regression models, conducting systematic evaluations using a standardized dataset of turbofan degradation patterns, enhancing interpretability with tools like SHAP and LIME, and validating findings against different datasets generated from Commercial Modular Aero-Propulsion System Simulation (C-MAPSS) with different scenarios. Through this structured approach, the paper aims to contribute to both academic literature and practical applications in aviation PHM.

Presenting Author: Zahi M. Omer Technology Innovation Institute

Presenting Author Biography: Dr. Zahi M. Omer is a Researcher in Predictive Analytics at the Technology Innovation Institute. He holds a Ph.D. in Electrical Engineering and has cultivated a research interest spanning Artificial Intelligence, Photovoltaics, Control, and Prognostics and Health Management (PHM). Among his many accomplishments, Dr. Omer has been distinguished with the Chancellor Innovation Award from UAE University and the Sharjah Sustainability Award in AI Applications. His noteworthy publications encompass advancements in ensemble machine learning for maximum power point tracking, rechargeable batteries, renewable energy systems, and optimization techniques.

Authors:

Zahi M. Omer Technology Innovation Institute
Abdulla Al Seiari Technology Innovation Institute
Zhaohui Cen Technology Innovation Institute
Francisco Bilendo Technology Innovation Institute