基于复合推进系统动态模型-状态变量模型的航空发动机直接推力预测控制

Direct Thrust Predictive Control of Aeroengine Based on Compact Propulsion System Dynamic Model-State Variable Model

关键词:  航空发动机  预测控制  复合推进系统动态模型  状态变量模型  直接推力控制

Key words:  Aero-engine  Predictive control  Compact propulsion system dynamic model  State variable model  Direct thrust control

摘要:

直接推力控制可以有效改善推力控制的品质，针对航空发动机直接推力控制问题，进行了模型预测控制（Model Predictive Control， MPC）研究。为了提升航空发动机推力控制的精度，提出了基于复合推进系统动态模型-状态变量模型（Compact Propulsion System Dynamic Model-State Variable Model-State Variable Model，CPSDM-SVM）的航空发动机直接推力预测控制方法。CPSDM实时估计出不可测参数（推力、喘振裕度等）的基准值，SVM则根据未来输入实时预测发动机未来响应。由于CPSDM将发动机分为进气道、核心机、喷管、喘振裕度、推力等进行建模，在兼顾精度的同时，提高机载模型的实时性。CPSDM-SVM作为MPC算法中的预测模型，具有较高的精度和实时性。仿真结果表明，在与基于分段线化模型的传统模型预测控制方法实时性基本相同的情况下，所提出方法控制效果有明显的提升，调节时间缩短了1.17s。所提出方法稳态控制精度为0.08%，传统方法稳态精度为2.58%。因此，所提出方法在保证实时性的条件下，提升了控制精度和控制效果。

Abstract:

Direct thrust control can effectively improve the quality of thrust control. A direct thrust controller was designed by the model predictive control（MPC） method. In order to improve the precision of thrust control of aero-engine， a direct thrust predictive control method based on Compact Propulsion System Dynamic Model-State Variable Model （CPSDM-SVM） was proposed. CPSDM estimated the reference value of unmeasurable parameters （thrust， surge margin， etc.） in real time， and SVM predicted the future engine response in real time based on future input. As CPSDM divides the engine into inlet， core machine， nozzle， surge margin， thrust， etc. for modeling， the accuracy is taken into account while improving the real-time performance of the airborne model. CPSDM-SVM， as a prediction model of MPC mothed， has high accuracy and real-time performance. Simulation results show that the control effect of the proposed method is significantly improved and the adjustment time is reduced by 1.17s under the condition that the real-time performance of the proposed method is basically the same as that of the traditional model predictive control method based on piecewise linear model. The steady-state control accuracy of the proposed method is 0.08%， while that of the traditional method is 2.58%. Therefore， the proposed method can improve the control precision and effect while ensuring the real-time performance.