基于显式预测控制的涡扇发动机控制器设计

Design of Turbofan Engine Controller Based on Explicit Predictive Control

关键词:  涡扇发动机  限制管理  显式模型预测控制  无扰切换  硬件在环试验

Key words:  Turbofan engine  Limit management  Explicit model predictive control  Bumpless switch  Hardware-in-loop

摘要:

模型预测控制因其能够较好地处理涡扇发动机约束问题且实现多变量控制而具有较大的应用潜力。为了解决传统模型预测控制算法在实际应用中存在运算量大、实时性较差的问题，基于显式模型预测设计了航空发动机多变量控制器。在控制结构上采用多速率双闭环系统，保证了控制精度；在计算上通过多参数规划将在线优化问题转化为线性函数计算问题，大幅度减少了计算量；在发动机从开环切闭环的过程提出一种增量式切换方法，实现无扰切换。数值仿真和硬件在环仿真结果表明，转速和压比稳态误差分别不超过±0.25%和±1%，控制器在25ms控制周期内能完成计算，满足嵌入式系统实时性要求。

Abstract:

Model predictive control （MPC） has great application potential because it can deal with the constraints of turbofan engine and realize multivariable control. A multivariable controller for aero-engine based on explicit model predictive control is proposed so as to solve the problems of large amount of calculation and poor real-time performance of traditional model predictive control algorithm in practical application. The control accuracy can be guaranteed through the multi-rate double closed-loop system in the control structure. The calculation amount has been dramatically reduced by transforming the online optimization problem into the linear function calculation problem with the multi-parametric programming method. In the process of engine switching from open loop to closed loop， an incremental switching method is proposed to realize undisturbed switching. The results of numerical simulation and hardware-in-loop simulation show that the steady-state errors of rotational speed and pressure ratio are less than ±0.25% and ±1%， respectively and the controller can complete the calculation during the 25ms control period， which meets the real-time requirements of embedded system.