Naslov
Computing the sampling time for induction machine parameter estimation using complex exponential series estimation

Autori
Benšić, Tin

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
25th Young statisticians meeting YSM25, Programme - Abstracts - Participants / Friedlt, Herwig - Graz : Institute of Statistics, Graz University of Technology, 2021, 9-9

Skup
25th Young statisticians meeting (YSM)

Mjesto i datum
Vorau, Austrija, 15.10.2021. - 17.10.2021

Vrsta sudjelovanja
Pozvano predavanje

Vrsta recenzije
Nije recenziran

Ključne riječi
induction machine parameter estimation, sampling time, data reduction, complex exponential series estimation

Sažetak
Induction motor is modeled by nonlinear state space model for which an offline and online parameter estimation techniques need to be developed. The offline techniques permit design of experiments and are generally well researched. However, most techniques choose how fast the sampled points of data are measured by experience. Through application of regression estimation of complex exponential series, the method for selecting the sampling interval is derived. The initial data set is manipulated through optimization procedure to accurately approximate the data obtained from a nonlinear system response with the complex exponential series. The estimated parameters of the series, namely the discrete system poles, are then used to compute the sampling time for the application of parameter estimation procedures for the nonlinear induction machine model. It is shown that the amount of data for offline parameter estimation of induction motor model can be significantly reduced. The data reduction permits the use of parameter estimators based on metaheuristic optimizers with realistic execution times.

Izvorni jezik
Engleski

Znanstvena područja
Matematika, Elektrotehnika

POVEZANOST RADA