Optimizing Gas Turbine Engine Thrust Predictions for UAV Propulsion System Using Regularized Regression Techniques
Authors :- U Patel, A Kumar, J Jyoti Rath, S Tripathi
Publication :- Journal of Aerospace Engineering, ASCE, 2025
Precise thrust prediction is essential for optimizing and controlling gas turbine engines in unmanned aerial vehicle (UAV) propulsion systems, where efficiency and reliability are critical. The nonlinear and complex interactions between operating parameters and thrust output present significant challenges for traditional predictive models. This study investigates advanced regularization techniques, specifically the least absolute shrinkage and selection operator (LASSO), ridge regression, and elastic net, to model the thrust characteristics of a standard turbine engine. Utilizing an in-house experimental test rig at the Institute of Infrastructure Technology Research and Management, Ahmedabad, data were collected across a wide operational range of 40,000 to 90,000 RPM. The data set was systematically divided into training and validation sets, and each model was rigorously evaluated for its ability to predict thrust under varying input conditions. Among these techniques, the ridge regression model demonstrated superior predictive accuracy and robustness compared to both LASSO and elastic net by effectively managing model regularization. These findings underscore the potential of ridge regression as an efficacious tool for thrust prediction in turbine-based propulsion systems, offering a data-driven solution that can significantly enhance UAV design and performance optimization under dynamic operating conditions.