Effective estimation of angular speed of synchronous generator based on stator voltage measurement

Estimation of the parameters of the power generators are important part of a power system modeling and simulation. A reliable estimation of a synchronous generator parameters can be done using waveforms measured directly on the terminals of the machine. The waveforms are recorded during a carefully chosen disturbance tests of the generating unit. One of the quantities necessary to determine the generator parameters is the deviation of the generator angular speed. However, because of the safety reasons, it is very difficult to get access to the direct measurement of this quantity on a operating generating unit. Angular speed estimation based on a generator stator voltage seems to be an alternative solution to the problem. The aim of the paper is to find a computationally efficient and accurate method of indirect angular speed estimation of a synchronous generator which is based on the recording of its stator voltage during a load rejection test. The algorithms evaluated in the paper are based on Prony's estimator, zero-crossing detection, least-square data fitting and layer recurrent neural network. Paper contains literature survey, mathematical model of the considered generator, analysis of its stator voltage recorded in a power plant, description of the selected algorithms and results of the algorithms estimation using a signal with known frequency variation. Performed algorithms evaluation allows to choose the best solution in terms of accuracy and computational effort.