Konferensartikel

The power industry heavily relies on power system modeling to understand system operations, perform system planning studies, and identify and correct problems that arise within the system. By minimizing the error between the models and actual physical system, it can be ensured that the models provide representation of both the existing and future the power system. Many of these models in the system are user-defined, i.e. they are specialized representations of a specific system component in the system. It is important that these customized models produce an accurate response. However, maintaining such models is costly, so it is of value to determine if those models can be replaced with a generic model. Phasor measurement data can be used to calibrate model parameters and reduce error. In this paper, this process is automated for a generator in the Itaipu power plant using RaPId, a MATLAB toolbox that integrates measurements, models using Modelica/ FMI standards, and optimization routines. This is achieved by using a combination of particle swarm optimization (PSO) algorithm and classical gradient optimization routines to calibrate the model parameters. In this paper, the calibration of a generic model of a synchronous generator, automatic voltage regulator, and power system stabilizer are estimated and compared to the user-defined models for an automatic voltage regulator, power system stabilizer, and turbine governor.

Modelica, FMI, synchronous generator parameter estimation, PSO, system identification