Maximum likelihood identification of power system dynamic equivalents

Abstract

This paper presents a technique for identifying unknown parameters for a dynamic equivalent of a portion of an electric power system from on-line measurements made in another portion of the system. The resulting model is suitable for use in a standard transient stability program such as those used by system analysts for evaluating the security of a proposed or existing system. The method is based on stochastic system identification using the maximum likelihood technique and relies on natural system fluctuation. This paper describes the mathematical development of a maximum likelihood identifier for a dynamic model of a power system. The resulting identification program was tested using a simulated power system tuned to match measurements made on a real system.