Details
| Original language | English |
|---|---|
| Title of host publication | Proceedings of ISMA 2020 - International Conference on Noise and Vibration Engineering and USD 2020 - International Conference on Uncertainty in Structural Dynamics |
| Editors | W. Desmet, B. Pluymers, D. Moens, S. Vandemaele |
| Pages | 3719-3733 |
| Number of pages | 15 |
| ISBN (electronic) | 9789082893113 |
| Publication status | Published - 2020 |
| Event | 2020 International Conference on Noise and Vibration Engineering, ISMA 2020 and 2020 International Conference on Uncertainty in Structural Dynamics, USD 2020 - online (Leuven), Belgium Duration: 7 Sept 2020 → 9 Sept 2020 |
Abstract
For the identification of the most influential input variables in a simulation model, global sensitivity analyses can be applied. For deterministic simulation models, i.e. models that yield the same outputs when being evaluated twice with the same set of inputs, sensitivity analyses are widely known. However, some simulation models are non-deterministic (i.e. stochastic) and include uncontrollable variables (also called seed or stochastic inputs). For stochastic models, each model evaluation leads to at least slightly different results. As a consequence, total sensitivity indices - measuring the influence of an input including all its interactions - cannot be calculated exactly, since uncontrollable variables cannot be fixed. So far, approximations are either based on an averaging approach or quasi total indices are used. Depending on the model and the considered input, any of the two approaches can yield better results. That is why, here, a method to determine the preferable approach is proposed.
ASJC Scopus subject areas
- Engineering(all)
- Mechanical Engineering
- Engineering(all)
- Mechanics of Materials
- Physics and Astronomy(all)
- Acoustics and Ultrasonics
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
Proceedings of ISMA 2020 - International Conference on Noise and Vibration Engineering and USD 2020 - International Conference on Uncertainty in Structural Dynamics. ed. / W. Desmet; B. Pluymers; D. Moens; S. Vandemaele. 2020. p. 3719-3733.
Research output: Chapter in book/report/conference proceeding › Conference contribution › Research
}
TY - GEN
T1 - Stochastic sensitivity analysis: determination of the best approximation of Sobol’ sensitivity indices
AU - Hübler, C.
N1 - Funding Information: The financial support of the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) for the ENERGIZE project (436547100) is gratefully acknowledged.
PY - 2020
Y1 - 2020
N2 - For the identification of the most influential input variables in a simulation model, global sensitivity analyses can be applied. For deterministic simulation models, i.e. models that yield the same outputs when being evaluated twice with the same set of inputs, sensitivity analyses are widely known. However, some simulation models are non-deterministic (i.e. stochastic) and include uncontrollable variables (also called seed or stochastic inputs). For stochastic models, each model evaluation leads to at least slightly different results. As a consequence, total sensitivity indices - measuring the influence of an input including all its interactions - cannot be calculated exactly, since uncontrollable variables cannot be fixed. So far, approximations are either based on an averaging approach or quasi total indices are used. Depending on the model and the considered input, any of the two approaches can yield better results. That is why, here, a method to determine the preferable approach is proposed.
AB - For the identification of the most influential input variables in a simulation model, global sensitivity analyses can be applied. For deterministic simulation models, i.e. models that yield the same outputs when being evaluated twice with the same set of inputs, sensitivity analyses are widely known. However, some simulation models are non-deterministic (i.e. stochastic) and include uncontrollable variables (also called seed or stochastic inputs). For stochastic models, each model evaluation leads to at least slightly different results. As a consequence, total sensitivity indices - measuring the influence of an input including all its interactions - cannot be calculated exactly, since uncontrollable variables cannot be fixed. So far, approximations are either based on an averaging approach or quasi total indices are used. Depending on the model and the considered input, any of the two approaches can yield better results. That is why, here, a method to determine the preferable approach is proposed.
UR - http://www.scopus.com/inward/record.url?scp=85105801340&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85105801340
SP - 3719
EP - 3733
BT - Proceedings of ISMA 2020 - International Conference on Noise and Vibration Engineering and USD 2020 - International Conference on Uncertainty in Structural Dynamics
A2 - Desmet, W.
A2 - Pluymers, B.
A2 - Moens, D.
A2 - Vandemaele, S.
T2 - 2020 International Conference on Noise and Vibration Engineering, ISMA 2020 and 2020 International Conference on Uncertainty in Structural Dynamics, USD 2020
Y2 - 7 September 2020 through 9 September 2020
ER -