Dynamic output compensation between selected channels in power systems. Part 1. The framework

Abstract

This paper presents a dynamic compensation framework for the design of a set of low order dynamic MISO controllers between selected input and output channels at each control site, or control agent, in power systems. The first step is to pick the best set of available measurements to feed in individual controllers at each control agent. The choice of measurements is made in conjunction with the control objectives in multiple time scales focusing on selected system modes. The measurements serve as controller inputs and are used for characterizing the system with low order models whose transfer matrices are good approximations of the full system. This step involves Hankel singular values analysis and balanced system representations. The second step is to design the controls. Two approaches are described. The idea of one method is to mimic the relevant dynamics of linear state feedback with a controller that takes into account structural constraints by separately designing each control input based on the best measurements available at each agent. Instead of retaining modes and mode shapes, as with projective control, this design method approximates the controller optimizing a time-domain performance index. While not as systematic as H ∞ design or as optimization techniques, the approach yields practical designs, superior to state feedback and projective control in terms of robustness. Furthermore, iterative μ-synthesis can be used as an alternate design if robustness is not sufficient. In Part 2 of the paper, examples are carried out. Measurements selection, agent model building, and application of robust control, of projective control and dynamic compensation in the design framework presented in Part 1 are illustrated.