Bavafa-Toosi Yazdan, Ohmori Hiromitsu, Labibi Batool
School of Integrated Design Engineering, Graduate School of Science and Technology, Keio University, Hiyoshi, Kohoku-Ku, Yokohama, Kanagawa, Japan.
ISA Trans. 2005 Oct;44(4):501-13. doi: 10.1016/s0019-0578(07)60057-3.
Most of the existing results on the decentralized stabilization of large-scale systems consider systems in the input-output-decentralized form, i.e., systems whose input and output matrices are block diagonal. The controller synthesis for such systems is less involved than that for input-output-centralized ones. In this paper, a transformation is proposed for the input-output decentralization of the generic large-scale system. Using the transformed system, a sufficient condition for failure-tolerant performance stabilization of the original large-scale system in a desirable performance region under decentralized linear output feedback is established. The problem is then reformulated as a constrained nonlinear optimization problem. The proposed methodology results in the optimal reconciliation of failure-tolerant performance stabilization of the overall system, and reliability and low actuator gains of the isolated subsystems. The effectiveness of the proposed approach is demonstrated by an example.
现有的关于大规模系统分散镇定的大多数结果都考虑输入-输出分散形式的系统，即输入矩阵和输出矩阵为块对角阵的系统。此类系统的控制器综合比输入-输出集中式系统的控制器综合要简单。本文针对一般大规模系统的输入-输出分散提出了一种变换。利用变换后的系统，建立了在分散线性输出反馈下，原大规模系统在期望性能区域内实现容错性能镇定的充分条件。然后将该问题重新表述为一个约束非线性优化问题。所提出的方法实现了整个系统容错性能镇定与孤立子系统的可靠性和低执行器增益之间的最优协调。通过一个例子说明了所提方法的有效性。
