Markov Chain Monte Carlo estimation of non-stationary GSC-TARMA models

Application to random vibration modelling and analysis for an operating wind turbine

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearch

Abstract

Generalized Stochastic Constraint Time-dependent ARMA (GSC-TARMA) models utilize stochastic representations for the time-evolution of their parameters and constitute effective representations for many non-stationary signals. This is especially so for signals obtained from systems characterized by stochastic evolution of their underlying dynamics. Yet, the estimation of GSC models via Maximum Likelihood type methods is generally cumbersome and complicated. In this study a Bayesian method employing Markov Chain Monte Carlo schemes is postulated. The method is applied to the modelling of wind turbine non-stationary random vibration under normal operating conditions. The use of GSC models for this application is motivated by the random fluctuations in the turbine's speed, as well as the stochastic non-stationary nature of the aerodynamic phenomena. The method is shown to yield good modelling accuracy and flexibility.
Original languageEnglish
Title of host publicationProceedings of ISMA 2016 - International Conference on Noise and Vibration Engineering and USD2016 - International Conference on Uncertainty in Structural Dynamics
PublisherISMA
Publication date2016
Pages4153-4167
Publication statusPublished - 2016
Externally publishedYes
Event27th International Conference on Noise and Vibration Engineering, ISMA 2016 and International Conference on Uncertainty in Structural Dynamics, USD2016 - Leuven, Belgium
Duration: 19. Sep 201621. Sep 2019

Conference

Conference27th International Conference on Noise and Vibration Engineering, ISMA 2016 and International Conference on Uncertainty in Structural Dynamics, USD2016
CountryBelgium
CityLeuven
Period19/09/201621/09/2019

Fingerprint

Wind turbines
Markov processes
Stochastic models
Maximum likelihood
Aerodynamics
Turbines

Cite this

Avendaño-Valencia, L. D., & Fassois, S. D. (2016). Markov Chain Monte Carlo estimation of non-stationary GSC-TARMA models: Application to random vibration modelling and analysis for an operating wind turbine. In Proceedings of ISMA 2016 - International Conference on Noise and Vibration Engineering and USD2016 - International Conference on Uncertainty in Structural Dynamics (pp. 4153-4167). ISMA.
Avendaño-Valencia, L.D. ; Fassois, S.D. / Markov Chain Monte Carlo estimation of non-stationary GSC-TARMA models : Application to random vibration modelling and analysis for an operating wind turbine. Proceedings of ISMA 2016 - International Conference on Noise and Vibration Engineering and USD2016 - International Conference on Uncertainty in Structural Dynamics. ISMA, 2016. pp. 4153-4167
@inproceedings{07623b6009e04690a68650aab0ee4f66,
title = "Markov Chain Monte Carlo estimation of non-stationary GSC-TARMA models: Application to random vibration modelling and analysis for an operating wind turbine",
abstract = "Generalized Stochastic Constraint Time-dependent ARMA (GSC-TARMA) models utilize stochastic representations for the time-evolution of their parameters and constitute effective representations for many non-stationary signals. This is especially so for signals obtained from systems characterized by stochastic evolution of their underlying dynamics. Yet, the estimation of GSC models via Maximum Likelihood type methods is generally cumbersome and complicated. In this study a Bayesian method employing Markov Chain Monte Carlo schemes is postulated. The method is applied to the modelling of wind turbine non-stationary random vibration under normal operating conditions. The use of GSC models for this application is motivated by the random fluctuations in the turbine's speed, as well as the stochastic non-stationary nature of the aerodynamic phenomena. The method is shown to yield good modelling accuracy and flexibility.",
author = "L.D. Avenda{\~n}o-Valencia and S.D. Fassois",
year = "2016",
language = "English",
pages = "4153--4167",
booktitle = "Proceedings of ISMA 2016 - International Conference on Noise and Vibration Engineering and USD2016 - International Conference on Uncertainty in Structural Dynamics",
publisher = "ISMA",

}

Avendaño-Valencia, LD & Fassois, SD 2016, Markov Chain Monte Carlo estimation of non-stationary GSC-TARMA models: Application to random vibration modelling and analysis for an operating wind turbine. in Proceedings of ISMA 2016 - International Conference on Noise and Vibration Engineering and USD2016 - International Conference on Uncertainty in Structural Dynamics. ISMA, pp. 4153-4167, 27th International Conference on Noise and Vibration Engineering, ISMA 2016 and International Conference on Uncertainty in Structural Dynamics, USD2016, Leuven, Belgium, 19/09/2016.

Markov Chain Monte Carlo estimation of non-stationary GSC-TARMA models : Application to random vibration modelling and analysis for an operating wind turbine. / Avendaño-Valencia, L.D.; Fassois, S.D.

Proceedings of ISMA 2016 - International Conference on Noise and Vibration Engineering and USD2016 - International Conference on Uncertainty in Structural Dynamics. ISMA, 2016. p. 4153-4167.

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearch

TY - GEN

T1 - Markov Chain Monte Carlo estimation of non-stationary GSC-TARMA models

T2 - Application to random vibration modelling and analysis for an operating wind turbine

AU - Avendaño-Valencia, L.D.

AU - Fassois, S.D.

PY - 2016

Y1 - 2016

N2 - Generalized Stochastic Constraint Time-dependent ARMA (GSC-TARMA) models utilize stochastic representations for the time-evolution of their parameters and constitute effective representations for many non-stationary signals. This is especially so for signals obtained from systems characterized by stochastic evolution of their underlying dynamics. Yet, the estimation of GSC models via Maximum Likelihood type methods is generally cumbersome and complicated. In this study a Bayesian method employing Markov Chain Monte Carlo schemes is postulated. The method is applied to the modelling of wind turbine non-stationary random vibration under normal operating conditions. The use of GSC models for this application is motivated by the random fluctuations in the turbine's speed, as well as the stochastic non-stationary nature of the aerodynamic phenomena. The method is shown to yield good modelling accuracy and flexibility.

AB - Generalized Stochastic Constraint Time-dependent ARMA (GSC-TARMA) models utilize stochastic representations for the time-evolution of their parameters and constitute effective representations for many non-stationary signals. This is especially so for signals obtained from systems characterized by stochastic evolution of their underlying dynamics. Yet, the estimation of GSC models via Maximum Likelihood type methods is generally cumbersome and complicated. In this study a Bayesian method employing Markov Chain Monte Carlo schemes is postulated. The method is applied to the modelling of wind turbine non-stationary random vibration under normal operating conditions. The use of GSC models for this application is motivated by the random fluctuations in the turbine's speed, as well as the stochastic non-stationary nature of the aerodynamic phenomena. The method is shown to yield good modelling accuracy and flexibility.

UR - http://past.isma-isaac.be/isma2016/proceedings/toc/

M3 - Article in proceedings

SP - 4153

EP - 4167

BT - Proceedings of ISMA 2016 - International Conference on Noise and Vibration Engineering and USD2016 - International Conference on Uncertainty in Structural Dynamics

PB - ISMA

ER -

Avendaño-Valencia LD, Fassois SD. Markov Chain Monte Carlo estimation of non-stationary GSC-TARMA models: Application to random vibration modelling and analysis for an operating wind turbine. In Proceedings of ISMA 2016 - International Conference on Noise and Vibration Engineering and USD2016 - International Conference on Uncertainty in Structural Dynamics. ISMA. 2016. p. 4153-4167