基于神经网络逆控制的TBCC发动机多变量限制管理  

Multi-variable limit management of TBCC engine based on neural network inverse control

作　　者：于兵强 张永亮 聂聆聪[2] 黄金泉[1] 鲁峰[1] YU Bingqiang;ZHANG Yongliang;NIE Lingcong;HUANG Jinquan;LU Feng(College of Energy and Power,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China;Beijing Power Machinery Institute,Beijing 100074,China)

机构地区：[1]南京航空航天大学能源与动力学院,江苏南京210016 [2]北京动力机械研究所,北京100074

出　　处：《推进技术》2024年第12期74-84,共11页Journal of Propulsion Technology

基　　金：航空发动机及燃气轮机基础科学中心项目(P2022-B-V-002-001)。

摘　　要：涡轮基组合循环(TBCC)发动机的控制系统既需要对执行机构协同控制以充分发挥每个工作模态的性能优势,又需要实现限制管理功能以保证发动机在安全条件下工作。本文通过分析串联式TBCC发动机流路计算过程,建立其性能动态模型,提出了一种基于神经网络预测反馈与逆控制的TBCC发动机多变量主控回路,其在单一模式阶跃响应超调小于3%,模态转换推力流量波动小于4%。在多变量控制架构中引入了限制管理策略,通过对比分析基于模型预测控制的多变量约束方法,仿真表明本文提出方法在考虑多变量耦合基础上,在过渡态和模态转换过程中满足超限幅度小于0.2%和0.07%,能有效实现限制管理,且结构简单,易于实现。The control system of the turbine-based combined cycle engine not only needs to coordinately control the actuators to give full play to the performance advantages of each operating mode,but also needs to im⁃plement limit management functions to ensure that the engine operates under safe conditions.This paper analyzes the flow path calculation process of the series TBCC engine,establishes its performance dynamic model,and pro⁃poses a multi-variable main control loop for the TBCC engine based on neural network predictive feedback and in⁃verse control.Its overshoot in the single-mode step response is less than 3%,and the thrust and flow rate fluctua⁃tion during the mode transition is less than 4%.A limit management strategy is introduced into the multivariable control architecture.By comparing and analyzing the multivariable constraint method based on model predictive control,the simulation shows that the method proposed in this paper,on the basis of considering multivariable coupling,can effectively realize limit management with the over-limit amplitude less than 0.2%and 0.07%dur⁃ing the transition state and mode transition process.Moreover,it has a simple structure and is easy to implement.

关 键 词：组合发动机 限制保护 Min-Max切换 模型预测控制 神经网络逆控制 

分 类 号：V231.1[航空宇航科学与技术—航空宇航推进理论与工程]