Research On Nonlinear Control For Aeroengines Based On Equilibrium Manifold Expansion Model

Modeling limitations make the aeroengine control problem quite different to those of most other nonlinear plants. The nonlinear componential level (NCL) models developed according to engines'aero thermodynamic principles are much too complex and not suitable for control design uses. The present aeroengine control applications widely use linear controllers and gain scheduling technique to design nonlinear controller. However, this technique lacks a theoretical basis to guarantee the stability and performance of the closed-loop system. Therefore developing engine models with simple structure and good physical interpretability is significant for both the research on theoretically guaranteed nonlinear control and future industry application. Equilibrium manifold expansion (EME) model exhibits some advantage in its application to aeroengines, but more investigation is still required. Motivated by the need of simple but reliable control oriented aeroengine model and control techniques, this dissertation is devoted to a systematical research on the EME model.Firstly, application problems of linear parameter-varying (LPV) control for aeroengines were studied. The reason that various control techniques in research had no real application was tried to be explained from a model point of view. The significance of LPV control in improving the widely used gain scheduling control for aeroengines was raised. Both local stability analysis of the closed-loop system and a turbofan control experiment revealed the problem of linearization LPV model of aeroengines. The application issues of an improved linearization LPV model, the velocity based LPV model was also discussed. It was pointed out that gain scheduling control for aeroengines could not achieve any breakthrough without a reliable LPV modeling method.Secondly, property and identification features of the EME model were systematically studied. A totally new definition of this model was given in advance to facilitate model analysis and further research, which clearly differentiated it from the currently well-known linear model, linearization family and LPV model. With this definition, the model property was studied, including its relation to the plant, constraint of model parameters and the mapping design that may affect the modeling result. Identification procedures were exemplified by a turbofan, assisted by which features and rationality of the so-called two-step modeling solution were revealed. With this solution, a more advantageous linear modeling method of aeroengines was proposed and then validated. Thirdly, the EME model based approximate feedback linearization control was studied. Features of EME model based control were studied through an analysis of the EME model of closed-loop system. The EME model based controller could guarantee local stability and performance of the real system. With this feature, the failure of the linearization LPV model based turbofan control in the earlier chapter was explained, and the EME model based approximate input-output feedback linearization control for aeroengines was proposed, of which the robustness was validated through plenty of turbofan simulations.Finally, EME model based gain scheduling control was studied. The value of local gain scheduling control techniques for aeroengines was raised. The EME description of gain scheduling controllers was derived from an analysis of the closed-loop gain scheduling system. Features of controller scheduling in EME form were analyzed. Issues related to design of gain scheduling controller in EME form were discussed, including simplified controller of this type and its existence, and requirement of further simplification to LPV form. Its application to a turbofan engine was investigated, in which effect of the mapping design and adaptation to large transient control is evaluated.