Propulsive Modeling for TF-30 and JT9D-3 Turbofan Engines of Transport Aircraft by Using Genetic Algorithms and Neural Networks

Abstract

Not only for aircraft performance calculations and trajectory predictions but also for air traffic management simulation tools and operational equipment, accomplishment of an accurate propulsive model for transport aircraft has a crucial priority and is a remarkable topic for aircraft industry. In the literature, there are very few propulsive modeling studies; furthermore, the demand for an accurate thrust model still remains unfulfilled. In this study, a new turbofan engine propulsive model determining the relationship between thrust, flight altitude, and Mach number was developed by using genetic algorithms (GAs) method and multilayer feed-forward neural networks (FNNs) utilizing Levenberg-Marquardt (LM), delta-bar-delta (DBD), and conjugate gradient (CG) learning algorithms. Estimated thrust values by the derived models showed a good fitting with actual thrust data for both models, which validated each model’s accuracy.