Multiobjective Evolutionary Optimization of Turbofan Engine Min-Max Fuel Control Algorithm

In this paper, modeling of Min-Max controller and evolutionary multiobjective optimization for gain tuning controller of turbofan engine are presented. To achieve this purpose, first a turbofan engine is modeledin GSP software. Then engine parameters model, by using extracted GSP simulation data and based onNARX structure of neural network is developed. For model validation a test fuel signal is produced and model performance is assessed by means of it. Next, turbofan engines control requirements and constraints are described and in accordance with it a fuel controller based on Min-Max strategy is designed and diverse control loops in controller is described. Each of theseloopshas aproportionalcontroller that are knownascontrol gains of the min-max controller. Then, for determining the gains of the controller, gain tuning process is formulated as a Genetic Algorithm Optimization problem in order to GA algorithm finds the best solution by its evolutionary generations. In this optimization problem, the settling time during acceleration and deceleration, engine fuel consumption and the amount of engine emission are considered as objective functions to be minimized. The obtained results from simulation of optimized controller and engine show, the final controller not only optimizes objective functions but also satisfies all control modes of engine during acceleration and deceleration modes.