A hybrid design method of steady-state throttling control schedules for high-flow variable cycle engine

Abstract

It is critical for variable cycle engine to design throttling control schedules that exploit its full steady-state performance potential. However, the increasing amount of variable geometry components makes it difficult to design the optimal control schedules, and the control schedules cannot adapt to the engine performance degradation. Therefore, a novel hybrid design method of steady-state throttling control schedules is proposed in this study by combining an offline optimization phase and an online compensation phase. In the offline optimization phase, a novel multi-objective differential evolution algorithm with neighborhood-based hierarchical mutation and cluster-based environment selection is first proposed to optimize the control schedules. The stratified selection method is then proposed to select the optimal and smooth control schedules. In the online compensation phase, the sensitivity calculation is first employed to identify the variable geometry component most sensitive to engine performance, and the compensator based on the extreme gradient boosting method is then designed to adjust the control schedules online. The hybrid design method is applied to the high-flow variable cycle engine. By comparing with the traditional design method, the results demonstrate that the proposed method can optimize the engine performance under the premise of ensuring good thrust linearity and that the control schedules can adapt to performance degradation during online use.