涡扇发动机最低红外辐射特征实时优化控制机理研究

Minimum infrared radiation characteristic of engine is one of the key factors to realize stealth of aircraft. Real-time optimal control of engine is a potential technical way to reduce infrared radiation characteristic effectively. However, the optimal control mechanism for minimum infrared radiation characteristics has not yet been explored, and the corresponding on-board adaptive model is also lacking, which is a frontier problem to be discussed and solved urgently. A method of turbofan engine infrared feature prediction considering both accuracy and real-time performance is proposed in this project. The component level model of turbofan engine with infrared prediction function is developed. For solving dimension disaster, an on-board adaptive composite model modeling method based on data-driven and physical mechanism is proposed. Meanwhile, an on-board adaptive turbofan engine model is established in broad flight envelope. The optimal control of minimum infrared radiation characteristics is studied and its control mechanism is explored. Moreover, a robust optimization algorithm based on soft constraints is proposed to solve the problem that thrust is difficult to be maintained and sensitive to the accuracy of on-board model. Finally, digital simulations and hardware-in-loop experiments are carried out to verify the effectiveness of the proposed method. An on-board adaptive model modeling method which is suitable for large envelope, multi-variables and multi-states is established in the project. In addition, the universal control mechanism of turbofan engine infrared radiation characteristics will be revealed.

发动机低红外辐射特征是实现飞机隐身的关键因素之一，发动机实时优化控制是有效降低红外辐射特征的潜在技术途径。然而，以最低红外辐射特征为目标的优化控制机理尚未被探明，也缺乏相应的机载自适应模型，这是亟待深入探讨和解决的前沿问题。本项目拟针对涡扇发动机，提出一种兼顾精度和实时性的红外特征预测方法，发展具有红外预测功能的发动机部件级模型；针对维数灾难问题，提出基于数据驱动和物理机理的机载自适应复合模型建模方法，建立适应大包线与部件性能退化状态的发动机机载自适应模型；开展最低红外辐射特征优化控制研究，探究其控制机理，针对推力难以保持及对机载模型精度敏感问题，提出基于软约束的鲁棒优化算法。最后进行数字仿真和硬件在回路仿真实验，验证本项目提出方法的有效性。本项目成果形成适用于大包线、多变量、多状态的机载自适应模型建模方法，揭示涡扇发动机红外辐射特征的普适性控制机理。

本项目首先开展了部件级模型（仿真平台）建模方法研究，提出了一种兼顾精度和实时性的红外特征预测方法，探究了发动机红外辐射强度的影响规律，发展了具有红外预测功能的发动机部件级模型。然后，进行机载自适应模型建模方法研究，针对维数灾难问题，提出基于深度学习和物理机理相融合的机载自适应复合模型建模方法，探索适用于稳态机载模型建模的相似准则，建立了大包线、多变量、多状态的发动机机载自适应模型。开展了最低红外辐射特征优化控制的机理研究，设计了最低红外辐射特性在线结构，为了提高优化算法的鲁棒性，提出了区间自适应线性规划算法线性规划。最后在典型爬升点、亚音速巡航点及超声速巡航点开展了数字仿真实验和硬件在回路仿真实验，结果表明，机载模型具有较高的精度和实时性，所提出方法可在不增加发动机重量的前提下，红外辐射特征下降6%-10%之间，验证本项目研究方法的有效性。本项目成果形成了适用于大包线、多变量、多状态的机载自适应模型建模方法，揭示了涡扇发动机红外辐射特征的普适性控制机理。