A knowledge-based system for generation capability dispatch during bulk power system restoration

During restoration, system operators need to consider special characteristics of power system elements. For example, the startup of some generators can be initiated within their plants, but other generators require cranking power from neighboring stations. Some generators may require manual controls in the beginning of restoration and have characteristics different from those of normal operations. The emergency ratings of generators allow faster ramping to speed up the restoration process. The emergency ratings of transmission lines will relax the limit of the power they can carry. System operators' burden can be relieved if there is a computer tool which can take care of these critical constraints and parameters and provide guidance throughout the whole restoration process. With such a tool, operators can deal with unusual operating conditions and avoid mistakes which may lead to failures of restoration.; This dissertation presents the results of development of a knowledge-based system (KBS) for bulk power system restoration. This KBS considers an extensive list of critical constraints and parameters for restoration of the generation and transmission parts of a power system. Based on these constraints and parameters, a generic goal-driven strategy is developed. Part of the strategy is a procedure to determine the generator starting times, which helps maximize the MW generation capability over a restoration period. This procedure incorporates heuristics related to critical constraints, searches and temporal reasoning. In addition to the generation within a power system, this KBS also considers the utilization of tie lines from neighboring utilities. A generic strategy to maximize the system generation capability with tie line import, export, and wheeling is developed. This KBS can be used to investigate how the whole restoration process can benefit from tie lines. The KBS is implemented in the Knowledge Engineering System (KES) expert system shell. Simulations of the KBS are performed with data from the Philadelphia Electric Company (PECo) system.; A continued effort is to enhance this KBS so that it will be versatile enough to satisfy different goals at different restoration stages for different utilities. A set of generic restoration actions (GRAs) have been identified. This set of GRAs can be used as basic building blocks to construct restoration plans for different utility requirements. (Abstract shortened by UMI.)...