RAPID TRANSMISSION CAPACITY MARGIN DETERMINATION FOR DYNAMIC SECURITYASSESSMENT USING ARTIFICIAL NEURAL NETWORKS

This paper presents a method for rapidly computing transmission capaci ty margin for security problems. The method uses an artificial neural network (ANN) to model a delineating surface between the region of sec urity and the region of insecurity in the space of most critical param eters. If the number of critical parameters is two, the delineating 's urface' is simply the familiar nomogram line. In general, however, the delineating surface is multidimensional since the number of critical parameters for most security problems is greater than three. The margi n to security limitations is computed as the solution to a nonlinear e quality-constrained optimization problem, where the equality constrain t is the operational boundary surface function, and the objective func tion is the shortest 'distance' between an operating point and the del ineating surface.