Non invasive parameter identification of power plant characteristics based on recorded network transient data

Abstract

Synchronous generators are the most widely used machines in power generation. Identifying their parameters in a non invasive way is very challenging due to the inherent nonlinearity of power plant performance. This thesis proposes a parameter identification method using particle swarm optimisation (PSO) for the identification of synchronous machine, excitation system and turbine parameters. The PSO allows a generator model output to be used as the objective function to give a new, more efficient method of parameter identification. This thesis highlights the effectiveness of the proposed method for the identification of power plant parameters, using both simulation and real recorded transient data. The thesis also considers the effectiveness of the method as the number of parameters to be identified is increased, and the effect of using differing forms of disturbances on parameter identification.