【作者】 王前宇；

【导师】 孙健国；

【作者基本信息】 南京航空航天大学 ， 航空宇航科学与技术， 2010， 硕士

【摘要】 多变量鲁棒控制是航空发动机控制发展的趋势,随着航空发动机控制量和被控制量的增多,如何从发动机的众多输出参数中选择合适的被控制量以及设计合适的控制结构成为航空发动机多变量控制中必须解决的问题。本文围绕航空发动机多变量鲁棒控制这一主题,开展了航空发动机建模、输出选择和控制结构设计、多变量鲁棒控制器设计及工程化实现研究。首先研究建立航空动机非线性部件级模型方法,在原有模型基础上,添加了导叶角对发动机影响的功能模块,完善了航空发动机非线性部件级模型;并研究了航空发动机自适应模型的建立方法,提出了基于误差反馈控制的航空发动机自适应模型建立方法,为今后进一步研究基于模型的航空发动机智能控制奠定了一定的基础。然后在此模型基础上,开展了航空发动机输出量选择和和控制结构设计研究,采用基于鲁棒稳定性和条件数（RSCN）方法分别对该型发动机进行了三变量、四变量输出选择,再采用相对矩阵增益（RGA）、块相对增益（BRG）方法和基于相互作用分解（RID）方法分别对它们进行了控制结构分析,确定了它们的控制结构;研究了增广LQR（ALQR）控制器设计方法与基于线性矩阵不等式（LMI）的航空发动机H₂/H_∞解耦控制器设计方法;采用遗传算法优化基于线性矩阵不等式的H₂/H_∞控制器的动态性能;通过将被控对象回路之间的耦合看成扰动信号并转化成H_∞性能指标,提出了基于线性矩阵不等式的H₂/H_∞解耦控制器设计方法;进行仿真验证了控制器设计方法的有效性;研究了航空发动机控制器工程化实现方法,包括抗积分饱和方法、加力扰动抑制方法和控制器整形化方法;更多还原

【Abstract】 Multivariable robust control is the developing trend of aero-engine control. Along with the increasing of the number of engine’s adjustable variables and controlled variables, how to select the better controlled variables and how to design the control configuaration of the engine have become crucial problems in engine’s control. This dissertation decomposed the aero-engine control system designing into three steps: controlled variables selection and control configuration design, multivariable robust controller design, and engineering implement of the controller. And this dissertation also studied the method of building aero-engine’s adaptive model.The research of this dissertation begins with the component-level modeling of areo-engine. In the basis of the existing component-level model, this dissertation added a new module which can reflect the influence of guide vane to the aero-engine. And adaptive modeling method also studied in this dissertation. A new adaptive modeling method which modeling the adaptive model based on err-feedback control was presented.And then, the controlled variabes selection and control structure design was studied based on the aero-engine model. The controlled variables of aero-engine’s three-variable control systes and four-variable control systes were selected by controlled variables selection method which is called RSCN （Robust Stability & Conditon Number）. And then design the control structure by two methods: Relative Gain Array （RGA） & Block Relative Gain （BRG） and Relative Interaction Decomposition （RID）.In chaper 4, Augmented LQR （ALQR） controller design method and H₂/H_∞controller design methed based on Linear Matrix Inequality （LMI） were studied; The dynamic performance of H₂/H_∞controller based on LMI was optimized by genetic algorithm. By converting the loop coupling of the controlled plant to H_∞performance index, a new controller design method,H₂/H_∞decoupling controller design methed based on LMI, was presented. And then, engineering implement of the controller was researched, includes anti-windup, anti-afterburner disturbance and transform the float controller to integer controller. And simulations were conducted to verify these methods.更多还原