Robust Decentralized Controller Design for Power Systems using Convex Optimization involving LMIs

This paper deals with the application of robust decentralized controller design for power systems using linear matrix inequality (LMI) techniques. In the design, the desired stability of the system is guaranteed while at the same time the tolerable bounds in the uncertainties due to structural changes, nonlinearities and load variations, are maximized. The approach allows the inclusion of additional design constraints such as the size and structure of the gain matrices. The paper also presents a decomposition procedure using the clustering technique of the states, inputs and outputs structure information to compute directly the appropriate diagonal structures of the output gain matrix for practical implementation. The algorithms were implemented on a test system and simulation results for power system stabilizer (PSS) design are presented.