| Power systems are under increasing stress as deregulation introduces several new economic objectives for operation. Since power systems are being operated close to their limits, weak connections, unexpected events, hidden failures in protection system, human errors, and a host of other factors may cause a system to lose stability and even lead to catastrophic failure. Therefore, the need for a systematic study and design of a comprehensive system control strategy is gaining more attention. Among these control methods, controlled system islanding is deemed as the final resort to save the system from a blackout.; In the literature, many approaches have been proposed to undertake this task. However, some of these approaches only take static power flow into consideration; others require a great deal of computational effort. It has been observed that following large disturbances, groups of generators tend to swing together. Attention has thus been drawn to the stability of inter-area oscillations between groups of machines. The slow-coherency based generator grouping, which has been widely studied in the literature, provides a potential method for capturing the movement of generators between groups under disturbance. The issue becomes on how to take advantage of slow coherency generator grouping and island the system by finding the set of lines to be tripped. Furthermore, through various simulations and analysis, it has been found that generator grouping indeed changes with respect to large changes in system load conditions.; In this dissertation, a comprehensive approach has been proposed to conduct slow coherency based controlled power system islanding using the minimal cutset technique from graph theory with the transition from calculating real power imbalance within the island to calculating the net flow through the cutset. Furthermore, a novel approach has been developed to trace the loci of the coherency indices of the slow modes in the system with respect to variation in system conditions to obtain the updated coherency information between generators using continuation method. Finally, the approach has been applied to a 10 generator 39 bus New England system, and a 29 generator 179 bus model of the WECC system. |
