Application of computational intelligence to power system vulnerability assessment and adaptive protection using high-speed communication

The purpose of vulnerability assessment is to determine when a disruption of service is likely to occur and to take steps to reduce the associated risk. With the growth of power systems, increases in their complexity, and the trend toward deregulation, vulnerability assessment has come to be a demanding process. In this new environment, accurate vulnerability assessment is especially vital and, above all, real-time assessment is greatly needed.; This dissertation investigates the potential of computational intelligence techniques, such as the use of neural networks in conjunction with the Particle Swarm Optimization (PSO) algorithm, as applied to the problem of power system vulnerability assessment.; First, a new feature-extraction technique is presented. A neural network structure, called the “auto-encoder,” is used to reduce input space dimensionality while preserving as much as possible of the original feature information. Power system data might be unavailable because of measurement device or communication channel failure. I propose a combination, to be called “missing feature restoration,” of an auto-encoder with a PSO algorithm as a new way of dealing with missing features. Next, two vulnerability assessment methods are proposed. One is a vulnerability index based on the distance of the current operating point from the vulnerability border of the system. The other is an index based on the anticipated loss of load. Next, a neural network is developed to estimate system vulnerability, bring together all of the techniques proposed. Finally, a preventive control algorithm is proposed. This new algorithm should provide valuable guidance for steering a system away from vulnerable operating regions, should a given operating state be judged vulnerable. I also include a review of “adaptive protection.” “Coordinated Wide-Area Protection and Control” scheme is presented. Some examples are developed, thereby showing the practical use of the new vulnerability assessment techniques.; Each of proposals is tested on simulations of realistic-sized systems, using models such as the IEEE 50-Genrator System and the WSCC 179-Bus System.