Research On Predictive Control Method Of Turbo-shaft Engine Based On Nonlinear Model

In view of the strongly nonlinear turbo-shaft engine model,there are some limits of the parameters in the process of engine control.The traditional closed-loop controller composed of feed-forward compensation plus PI has a poor effect on the transient state control.In order to meet the requirements of the limits in the process of engine transient control,this paper does some research on the nonlinear model predictive control of the turbo-shaft engine.In order to verify the effectiveness of the control law,the steady-state control law of a certain type of civil turbo-shaft engine was explored.The cascade PI controller was used to analyze the variation law of the parameter margin of the turbo-shaft engine in the full flight envelope and under different control targets to provide a basis for predictive control optimization algorithms.Because the LSSVM prediction model used in the nonlinear predictive control method requires a large amount of sample data for training,a sequential quadratic programming(SQP)optimization algorithm with constraints is used to collect the transition state parameters change data of each point within the envelope as a sample data.Sparse Least Squares Support Vector Machine(LSSVM)is used to train and test sample data.Because the trained LSSVM model contains dynamic data,a transition state control method composed of LSSVM as feed-forward compensation and PI closed-loop circuit is designed to realize the transition state control of turbo-shaft engine.By simulating the component-level model of a civil turbo-shaft engine,the results show that the controller can effectively improve the transient state control effect of the turbo-shaft engine.In order to achieve the stable and reliable control of the turbo-shaft engine during the working process and meet the parameter limit requirements,a model predictive control method based on a least squares support vector machine model is designed.The closed-loop controller composed of feed-forward compensation plus PI are compared and the simulation results show that the nonlinear model predictive control method has a good effect on the power turbine speed control.