Details
| Original language | English |
|---|---|
| Title of host publication | COMPDYN 2019 |
| Subtitle of host publication | 7th International Conference on Computational Methods in Structural Dynamics and Earthquake Engineering, Proceedings |
| Editors | Manolis Papadrakakis, Michalis Fragiadakis |
| Pages | 592-599 |
| Number of pages | 8 |
| ISBN (electronic) | 9786188284463 |
| Publication status | Published - 2019 |
| Event | 7th International Conference on Computational Methods in Structural Dynamics and Earthquake Engineering, COMPDYN 2019 - Crete, Greece Duration: 24 Jun 2019 → 26 Jun 2019 Conference number: 7 |
Abstract
Modern approaches to solve dynamic problems where random vibration is of significance will in most of cases rely upon the fundamental concept of the power spectrum as a core model for excitation and response process representation. This is partly due to the practicality of spectral models for frequency domain analysis, as well as their ease of use for generating compatible time domain samples. Such samples may be utilised for numerical performance evaluation of structures, those represented by complex non-linear models. While development of spectral estimation methods that utilise ensemble statistics to produce a single or finite number of deterministic spectral estimate(s), result in a familiar spectral model that can be directly understood and applied in structural analyses, significant information pertaining to the non-ergodic characteristics of the process are still lost. In this work, an approach for a stochastic load representation framework that captures epistemic model uncertainties by encompassing inherent statistical differences that exist across real data sets is used. The new developed stochastic load representation is utilising imprecise probabilities to capture these epistemic uncertainties and represent this information effectively. In some cases, there will be sufficient source data available to identify that a relaxed power spectral estimate is likely to better represent the process, but not enough data to establish a probabilistic description of a relaxed model. In these cases, an interval approach will be employed to capture the epistemic uncertainty in the spectral density of the process. Combined with the stochastic nature of the process itself, this leads to an imprecise probabilistic model. Since data is limited in this case, a parametric approach is utilised. The most likely power spectrum of the ensemble is identified and the model is relaxed by implementing interval parameters such that the resulting bounds form an envelope for all estimated spectral powers.
Keywords
- Fuzzy Methods, Imprecise Probabilities, Power Spectrum Estimation, Random Vibrations, Relaxed Power Spectra Model, Uncertainty Quantification
ASJC Scopus subject areas
- Earth and Planetary Sciences(all)
- Computers in Earth Sciences
- Earth and Planetary Sciences(all)
- Geotechnical Engineering and Engineering Geology
- Mathematics(all)
- Computational Mathematics
- Engineering(all)
- Civil and Structural Engineering
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
COMPDYN 2019: 7th International Conference on Computational Methods in Structural Dynamics and Earthquake Engineering, Proceedings. ed. / Manolis Papadrakakis; Michalis Fragiadakis. 2019. p. 592-599.
Research output: Chapter in book/report/conference proceeding › Conference contribution › Research › peer review
}
TY - GEN
T1 - Relaxed stationary power spectrum model using imprecise probabilities
AU - Behrendt, Marco
AU - Comerford, Liam
AU - Beer, Michael
N1 - Conference code: 7
PY - 2019
Y1 - 2019
N2 - Modern approaches to solve dynamic problems where random vibration is of significance will in most of cases rely upon the fundamental concept of the power spectrum as a core model for excitation and response process representation. This is partly due to the practicality of spectral models for frequency domain analysis, as well as their ease of use for generating compatible time domain samples. Such samples may be utilised for numerical performance evaluation of structures, those represented by complex non-linear models. While development of spectral estimation methods that utilise ensemble statistics to produce a single or finite number of deterministic spectral estimate(s), result in a familiar spectral model that can be directly understood and applied in structural analyses, significant information pertaining to the non-ergodic characteristics of the process are still lost. In this work, an approach for a stochastic load representation framework that captures epistemic model uncertainties by encompassing inherent statistical differences that exist across real data sets is used. The new developed stochastic load representation is utilising imprecise probabilities to capture these epistemic uncertainties and represent this information effectively. In some cases, there will be sufficient source data available to identify that a relaxed power spectral estimate is likely to better represent the process, but not enough data to establish a probabilistic description of a relaxed model. In these cases, an interval approach will be employed to capture the epistemic uncertainty in the spectral density of the process. Combined with the stochastic nature of the process itself, this leads to an imprecise probabilistic model. Since data is limited in this case, a parametric approach is utilised. The most likely power spectrum of the ensemble is identified and the model is relaxed by implementing interval parameters such that the resulting bounds form an envelope for all estimated spectral powers.
AB - Modern approaches to solve dynamic problems where random vibration is of significance will in most of cases rely upon the fundamental concept of the power spectrum as a core model for excitation and response process representation. This is partly due to the practicality of spectral models for frequency domain analysis, as well as their ease of use for generating compatible time domain samples. Such samples may be utilised for numerical performance evaluation of structures, those represented by complex non-linear models. While development of spectral estimation methods that utilise ensemble statistics to produce a single or finite number of deterministic spectral estimate(s), result in a familiar spectral model that can be directly understood and applied in structural analyses, significant information pertaining to the non-ergodic characteristics of the process are still lost. In this work, an approach for a stochastic load representation framework that captures epistemic model uncertainties by encompassing inherent statistical differences that exist across real data sets is used. The new developed stochastic load representation is utilising imprecise probabilities to capture these epistemic uncertainties and represent this information effectively. In some cases, there will be sufficient source data available to identify that a relaxed power spectral estimate is likely to better represent the process, but not enough data to establish a probabilistic description of a relaxed model. In these cases, an interval approach will be employed to capture the epistemic uncertainty in the spectral density of the process. Combined with the stochastic nature of the process itself, this leads to an imprecise probabilistic model. Since data is limited in this case, a parametric approach is utilised. The most likely power spectrum of the ensemble is identified and the model is relaxed by implementing interval parameters such that the resulting bounds form an envelope for all estimated spectral powers.
KW - Fuzzy Methods
KW - Imprecise Probabilities
KW - Power Spectrum Estimation
KW - Random Vibrations
KW - Relaxed Power Spectra Model
KW - Uncertainty Quantification
UR - http://www.scopus.com/inward/record.url?scp=85079089659&partnerID=8YFLogxK
U2 - 10.7712/120119.6941.19045
DO - 10.7712/120119.6941.19045
M3 - Conference contribution
AN - SCOPUS:85079089659
SP - 592
EP - 599
BT - COMPDYN 2019
A2 - Papadrakakis, Manolis
A2 - Fragiadakis, Michalis
T2 - 7th International Conference on Computational Methods in Structural Dynamics and Earthquake Engineering, COMPDYN 2019
Y2 - 24 June 2019 through 26 June 2019
ER -