大信号扰动下基于切换序列的阻抗源网络动态建模与优化控制研究

项目名称： 大信号扰动下基于切换序列的阻抗源网络动态建模与优化控制研究

项目编号： No.51307009

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 电工技术

项目作者： 汤赐

作者单位： 长沙理工大学

项目金额： 26万元

中文摘要： 阻抗源网络的强非线性、时变性及运行模态复杂多变性的特点，导致现有基于小信号线性化的建模与控制方法，无法精准描述大信号扰动下阻抗源网络的动态特性，难以优化控制大信号扰动下阻抗源网络的动态行为。本项目研究(1)引入切换序列刻画大信号扰动下阻抗源网络的状态轨迹，表征整体上离散位置跳跃迁移、局部上连续运行渐进变化的运动特征；(2)借助切换序列决定阻抗源网络各个线性子系统的作用顺序和持续时间，构建阻抗源网络周期性切换线性系统模型，精准描述大信号扰动下的动态特性；(3)借鉴状态空间的概念，着眼于稳定性，状态能控性、能达性、能观性，输出能控性等基本控制特性，创建大信号扰动下阻抗源网络基于切换序列的理论体系；(4)构造大信号扰动下基于范数、收敛路径、收敛距离、收敛方向、综合性能的优化切换序列，通过状态轨迹的收敛过程，优化控制阻抗源网络的动态行为。本项目的研究有望解决大信号扰动下阻抗源网络的建模与控制难题。

中文关键词： 阻抗源网络；大信号扰动；切换序列；动态建模；优化控制

英文摘要： The characteristics of strong nonlinearity, time-variation, and complicated operating modes, which inherented in the impedance network, result in small-signal-linearization-based modeling and control methods not appropriate to describe accurrately dynamic features and to control optimally dynamic behaviors for the impedance network underwent large-signal perturbations. For solveing such a problem, switched sequences are lead into the dipicting of state trajectories of the impedance network, thus the movements property with hop shifting for discrete positions on the whole while gradual varying for continous motions locally is successfully characterized. With the help of swithced sequences, every linear subsystems of the impedance network are decided their orders and durations. After that, a periodic switched linear model for the impedance network is found so that dynamic features under large-signal perturbations are achived to describe accurately. Refering to the concept of the state space, a basic theory system for the switched-sequences-based impedance netowork under large-signal perturbations is set up in terms of stability, controllability, reachability, observability, and output controllability. Five optimal switched sequences suitable for large-signal perturbations, such as norms, convergence paths, converg

英文关键词： impedance network；large-signal perturbations；switched sequences；dynamic modeling；optimal control