The empirical identification of synchronous machine parameters
The empirical identification of synchronous machine parameters
A methodology is presented for estimating the parameters of detailed electrical machine models. Advanced computational methods are derived for the analysis of observed machine behaviour using data from both well established and more recent tests. Because of their great economic and technological importance, emphasis is placed on parameter estimation for large turboalternators; however, the derived methods are applicable to arty electrical machine. The least-squares technique is used to calculate the parameters of a Laplace transform model best fitting an observed test response and two algorithms are presented. The first is a refinement of the method due to Sanathanan and Koerner and has been found to be superior to a comparable method based on a step-by-step solution. The machine test response is input to these routines and can be measured directly in the frequency domain or estimated by numerical transformation of time-domain measurements; techniques are devised to overcome the problems of aliasing and leakage which determine the accuracy of this transformation. Detailed examination is given to three machine tests - the sudden short-circuit test and two standstill tests, one in each of the time and frequency domains - and the conventional theory of electrical machines is extended to derive equivalent circuit models from the test results. The problems of extracting direct and quadrature axis information from short-circuit tests results are investigated and computational techniques devised to extract this information using only armature current data. Use of the parameter estimation methodology will enable more accurate models of machine behaviour to be made and, therefore, more meaningful predictions of performance. The results of its application on both simulated and actual machine tests is-reported, including a series of the three principal tests on a 66014W turboalternator. To realise the methodology, a large body of software has been written for both interactive and batch processing environments and can be used for the analysis of test results, to test new methods or as a specification for future systems. Although developed for electrical machine identification, the software is perfectly general and can be applied to any continuous linear system.
University of Southampton
1979
Ross, Andrew Alexander
(1979)
The empirical identification of synchronous machine parameters.
University of Southampton, Doctoral Thesis.
Record type:
Thesis
(Doctoral)
Abstract
A methodology is presented for estimating the parameters of detailed electrical machine models. Advanced computational methods are derived for the analysis of observed machine behaviour using data from both well established and more recent tests. Because of their great economic and technological importance, emphasis is placed on parameter estimation for large turboalternators; however, the derived methods are applicable to arty electrical machine. The least-squares technique is used to calculate the parameters of a Laplace transform model best fitting an observed test response and two algorithms are presented. The first is a refinement of the method due to Sanathanan and Koerner and has been found to be superior to a comparable method based on a step-by-step solution. The machine test response is input to these routines and can be measured directly in the frequency domain or estimated by numerical transformation of time-domain measurements; techniques are devised to overcome the problems of aliasing and leakage which determine the accuracy of this transformation. Detailed examination is given to three machine tests - the sudden short-circuit test and two standstill tests, one in each of the time and frequency domains - and the conventional theory of electrical machines is extended to derive equivalent circuit models from the test results. The problems of extracting direct and quadrature axis information from short-circuit tests results are investigated and computational techniques devised to extract this information using only armature current data. Use of the parameter estimation methodology will enable more accurate models of machine behaviour to be made and, therefore, more meaningful predictions of performance. The results of its application on both simulated and actual machine tests is-reported, including a series of the three principal tests on a 66014W turboalternator. To realise the methodology, a large body of software has been written for both interactive and batch processing environments and can be used for the analysis of test results, to test new methods or as a specification for future systems. Although developed for electrical machine identification, the software is perfectly general and can be applied to any continuous linear system.
This record has no associated files available for download.
More information
Published date: 1979
Identifiers
Local EPrints ID: 462693
URI: http://eprints.soton.ac.uk/id/eprint/462693
PURE UUID: e6b5cf4c-db4d-4781-8443-87e94531dad3
Catalogue record
Date deposited: 04 Jul 2022 19:41
Last modified: 04 Jul 2022 19:41
Export record
Contributors
Author:
Andrew Alexander Ross
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics