【作者】 李业波；

【导师】 李秋红；

【作者基本信息】 南京航空航天大学 ， 航空宇航推进理论与工程， 2011， 硕士

【摘要】 本文是为飞/推综合控制系统仿真平台提供航空发动机组件化模型。旨在利用软件组件化的思想,将COM组件技术应用到航空发动机建模中,建立一个可重用、通用性强以及精度高的组件化模型,并在此模型基础之上,研究航空发动机性能参数估计的方法。首先将航空发动机按照结构划分为10个主要部件,分别建立各部件的数学模型并将其封装成COM组件。各组件之间以及与模型与仿真平台之间的数据传递通过接口实现。此外,还创建了航空发动机的一些功能组件,用于封装功能函数等。其次,进行航空发动机仿真平台的设计,包括界面设计和内部运行管理系统设计。仿真平台的界面是在MFC单文档框架的基础上创建的,利用树形控件和绘图技术等设计了组件栏和模型仿真区;在内部运行管理系统中,提供了牛顿-拉夫森法和Broyden秩1拟牛顿法解模型动态方程。本文还提出了将支持向量回归机与遗传算法相结合进行航空发动机参数估计的方法。针对简约最小二乘支持向量回归机的简约集确定问题,利用遗传算法具有很好的全局寻优能力来选择支持向量,构成简约集,以实现对航空发动机参数的准确估计。为了验证此估计方法的实用性,论文还设计了ALQR推力控制器,以推力估计器的推力估计值为反馈量,形成闭环控制系统。仿真结果表明,所设计的推力估计器能准确对推力进行估计,将其作为反馈用于直接推力控制系统,实现了发动机在飞行包线内的快速响应和平稳过渡。更多还原

【Abstract】 This paper is designed to provide componentization model of aero-engine for integrated flight/propulsion control system simulation platform. In this paper, the software componentization ideology is adopted and the COM technology is applied to aero-engine modeling in order to establish a reusable, universal, and high accuracy aero-engine model. On the basis of this model, the aero-engine performance parameter estimation method is researched.Firstly, the aero-engine is divided into ten main components from the perspective of its structure. Mathematics model of every component is established and then every mathematics model is packaged as a COM component. Data transfer among all components, as well as between model and simulation platform, is realized by way of the interfaces. In addition, some components are built to package the important functions.Secondly, the design of the aero-engine simulation platform is carried out, which includes the design of interface and inner function. The interface of the simulation platform is established on the basis of MFC single-document framework. The tree control and drawing technique is adopted to design the component column and the model simulation area. In the inner working system, both Newton-Raphsion method and Broyden quasi-Newton method are introduced for the purpose of solving the dynamic balance equations of the model.In this paper, a method is proposed for aero-engine performance parameters estimation, based on the combination of support vector regression (SVR) and genetic algorithm (GA). As for confirming the reduced set of reduced least squares support vector regression (RLSSVR), genetic algorithm (GA) is used because of its global performance seeking capability to select support vector and compromise the reduced set in order to realize accurate parameters estimation of the aero-engine performance. To verify the practicality of this aero-engine performance parameters estimation method, the ALQR thrust controller is designed and the estimate thrust is used as feedback signal to form a close loop control system. The simulations demonstrate that the thrust estimator can estimate thrust accurately and that the estimate thrust as feedback is applied to the thrust control system, so that the engine can get rapid response and realize smooth transition in the fly envelope.更多还原