A GA-based method for optimizing topological observability index in electric power networks

Abstract

This paper proposes a topological observability index for power system static state estimation. Topology observability is equivalent to the existence of full-rank spanning trees in a network representing meter allocation. The proposed index may be expressed as the total number of the trees. In order to enhance topological observability, it is necessary to optimize the index with a given set of measurements. Since the problem of the index optimization may be described as one of integer programming with a nonlinear discontinuous complicated cost function, the conventional methods do not allow us to provide the solution. A genetic algorithm (GA) is one of promising technologies for complicated optimization problems. In this paper, GA is used to solve the problem. >