Statistical method for identification of weak nodes in power system based on voltage magnitude deviation

Abstract

Identification of weak buses in an electric power system is important for operational and planning purposes. In this paper, a framework that uses the variance of the magnitude of voltage at load buses when reactive power of all loads are varied concurrently is used to identify a set of weak buses. In the framework, reactive power for all load buses are generated concurrently and independently based on the probability distribution function of load demands at each node. Three IEEE test systems are used to test the framework in Pypower in conjunction with Python programming language. The result shows that weak buses can be identified with the framework. It was observed that the first three set of most critical nodes in a network can be identified within a variation of 1% of reactive power at load nodes in case of ordinary power flow (pf), while the same set of nodes can be identified with at least 5% of variation of reactive power at nodes based on optimal power flow (opf). One area of application of this work could be in the use of the variance of voltage magnitude measurements to quickly identify a subset of weak nodes in a network.