基于射影算子的变循环发动机鲁棒自适应控制器设计

Robust Adaptive Controller Design of Variable CycleEngine Based on Projection

关键词:  变循环航空发动机  多变量控制  鲁棒自适应控制  射影算子  动态跟踪

Key words:  Variable cycle engine  Multi-variable control  Robust adaptive control  Projecttion  Dynamic tracking

摘要:

针对变循环航空发动机大范围工作包线下的复杂外部扰动现象引起的多变量鲁棒自适应控制器设计问题，给出基于射影算子的自适应律设计方法并将其应用至模型参考自适应控制器设计中。通过理论分析与变循环航空发动机某一状态点的仿真分析，阐明系统存在外部干扰时传统自适应律的设计方法对于指令跟踪不足等问题，进而在射影算子的理论框架基础上设计自适应律，将存在干扰时的状态误差动态方程的解限制在设定的凸集范围内，从而实现系统状态误差动态方程不大于零的稳定性要求，最终在LQR基准控制器框架上，实现基于射影算子的多变量鲁棒自适应控制器设计，以改善自适应控制系统的鲁棒性。基于该方法对变循环航空发动机进行控制器设计和仿真。结果表明:基于射影算子的自适应律设计方法改善了传统自适应律设计方法的鲁棒性问题，实现了控制系统对外部干扰的有效抑制，当系统各控制回路加入不同外部随机噪声信号时，均达到了期望的控制效果；变循环航空发动机各个状态点均能够跟踪期望的闭环参考指令，各状态点稳态控制误差均小于0.5%，系统超调量小于1%，调节时间小于1.2s，动态跟踪误差不大于0.5%，符合发动机控制系统技术要求。

Abstract:

In order to solve the problem of multivariable robust adaptive controller design with external disturbances which caused by the variable cycle engine’s wide range of flight envelope，an adaptive law based on projection was given and applied to the model reference adaptive controller design. First，the traditional design method of adaptive law with problems of order tracking when system existed external disturbances were clarified by the theoretical analysis and simulation analysis of a variable cycle engine’s state point. Then adaptive law on the basis of the theoretical framework of projection was designed. And the solutions of the state error’s dynamic equations were limited within the setting range of convex sets，so as to realize the stability requirements that state error’s dynamic equations was not greater than zero，finally the projection multi-variable robust adaptive control design based on LQR controller was finished to improve the robustness of adaptive control system. The simulation results of variable cycle engine have shown that: the design method of adaptive law based on projection improves the robustness problem of traditional method，effectively suppresses external disturbances，achieves the expected control effect when joined different external random noise signals. Meanwhile，at the different state points，the steady control errors are less than 0.5%，the overshoot values are less than 1%，the adjust times are less than 1.2s，the errors of dynamic tracking are less than 0.5%，the results conform to the aero engine control system technology’s requirements.