Aeroengine Performance Optimization Control Method Based on Model-Free Adaptive Prediction Algorithm

In order to overcome the problems of complex dynamic characteristics and significant model uncertainties of aero-engine control, this paper proposes an optimal control strategy for aero-engine based on model free adaptive predictive control (MFAPC) method. First, a data-driven compact form dynamic linearization (CFDL) model of the engine is established based on the input and output data, as a replacement of its nonlinear model and the estimation algorithm of the Pseudo Jacobi Matrix(PJM) is designed to construct the model-free adaptive control; based on this, a prediction algorithm of PJM is added, and future input and output information is introduced to make the algorithm have a better control effect and stronger robustness;; considering that the parameter adjustment space of the MFAPC controller is large and the control effect of the MFAPC method is very much dependent on the setting of its controller parameters, a parameter optimization method based on arctic puffin optimization (APO) is designed to realize the self-tuning of the controller parameters. The results of the simulation demonstrate that, in comparison with the basic MFAC controller and the PID controller, the proposed APO-MFAPC controller exhibits superior robustness and improved control performance. It is further demonstrated that the proposed algorithm has strong adaptive adjustment ability with reasonable computational assumption.