Predictive control method for aero-engine life extension considering combustion chamber outlet temperature distribution

Nonuniform combustor-exit temperature distribution can induce local hot spots on turbine blades, increase thermal gradients, and accelerate thermomechanical fatigue damage. Conventional aero-engine control optimization mainly focuses on thrust and fuel consumption, with limited consideration of component service life. To address this limitation, this paper proposes a life-extending model predictive control method for turbofan engines that accounts for combustor-exit temperature distribution. CFD simulations are first conducted to characterize the outlet temperature distribution under different operating conditions. A prediction model that combines a deep neural network with a state-variable model is constructed to meet the requirements of online optimization. The resulting nonlinear optimization problem is then reformulated as a quadratic programming problem. Validation results show that the maximum dynamic prediction error is 4.35 %, indicating sufficient accuracy for online predictive control. For the PLA transition from 20° to 70°, the proposed method reduces the peak turbine-inlet hot-spot temperature by 32 K and increases the predicted thermomechanical-fatigue safe life of the high-pressure turbine guide vane by 19.3 % compared with the conventional LEC method.