Design of optimal variable structure controllers: Applications to power system dynamics

In this thesis, optimal design of variable structure controllers applied to different power system dynamic problems is explored. Iterative heuristics algorithms namely: Particle Swarm optimization, Tabu Search algorithm, and Genetic Algorithms are utilized in the optimal design of the controllers. Two important power system problems (Load frequency control and Power system dynamic stability) are studied in this thesis. The proposed design of VSC was applied to single area, two areas, and three areas Load frequency control systems both with and without nonlinearities and to a single machine infinite bus-bar system.; The variable structure controller parameters are tuned optimally to minimize certain performance indices that reflect the objectives of the design. These objectives include both improved dynamic behavior and reduced chattering in the control signal. Scaling of the feedback gains in the VSC on convergence of the performance index or the square of the error in the signal further reduced the chattering while preserving the improved dynamic behavior of the system. Comparison with conventional design methods for both problems using the new design procedure is also included in the study. The study shows that the proposed optimal design procedure provides a simple systematic way of arriving at the settings of the VSC with improved dynamic behavior and reduced chattering. This method allows with ease the inclusion of nonlinearities into the studied systems to realize an optimally designed controller. The robustness of the proposed controller against system parameters variations had been studied. It has been found that the proposed controllers are very little sensitive to these variations.