An Improved Engine Model Adaption Method Based on Search Interval Optimization

Aeroengine performance model is an important tool to reflect the engine working condition and evaluate the engine performance. The engine model adaption method based on nonlinear component-level performance model is an important technical support to achieve accurate assessment of engine performance. However, when traditional optimization algorithm-based model adaption method is applied to the actual issues, especially when the known information for model adaption is insufficient, the accuracy of the model adaption is easily affected by the search interval of the independent variables, and problems such as smearing effects and local optimality can easily occur. In this paper, an improved engine model adaption method based on initial search interval optimization and interval adaptive adjustment is proposed. This method can effectively use the initial deviation information of model and target parameters and the feedback from the optimization algorithm in the adaption process, realizing the function of “optimizing the search interval before adaption and dynamically adjusting the search interval during adaption.” By applying these two methods to genetic algorithm (GA), the robustness of the adaption method based on nonlinear component-level performance model can be effectively improved, and the accurate evaluation of engine performance under multiple information conditions can be realized. The effectiveness of the method was verified by using ground state data of a mixed flow turbofan.