Multiphase flow applications of nonintrusive reduced-order models with Gaussian process emulation
Reduced-order models (ROMs) are computationally inexpensive simplifications of high-fidelity complex ones. Such models can be found in computational fluid dynamics where they can be used to predict the characteristics of multiphase flows. In previous work, we presented a ROM analysis framework that...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Cambridge University Press
2022-01-01
|
| Series: | Data-Centric Engineering |
| Subjects: | |
| Online Access: | https://www.cambridge.org/core/product/identifier/S2632673622000193/type/journal_article |
| _version_ | 1811156433483333632 |
|---|---|
| author | Themistoklis Botsas Indranil Pan Lachlan R. Mason Omar K. Matar |
| author_facet | Themistoklis Botsas Indranil Pan Lachlan R. Mason Omar K. Matar |
| author_sort | Themistoklis Botsas |
| collection | DOAJ |
| description | Reduced-order models (ROMs) are computationally inexpensive simplifications of high-fidelity complex ones. Such models can be found in computational fluid dynamics where they can be used to predict the characteristics of multiphase flows. In previous work, we presented a ROM analysis framework that coupled compression techniques, such as autoencoders, with Gaussian process regression in the latent space. This pairing has significant advantages over the standard encoding–decoding routine, such as the ability to interpolate or extrapolate in the initial conditions’ space, which can provide predictions even when simulation data are not available. In this work, we focus on this major advantage and show its effectiveness by performing the pipeline on three multiphase flow applications. We also extend the methodology by using deep Gaussian processes as the interpolation algorithm and compare the performance of our two variations, as well as another variation from the literature that uses long short-term memory networks, for the interpolation. |
| first_indexed | 2024-04-10T04:51:30Z |
| format | Article |
| id | doaj.art-6755315767f544b5b90ff590625396c1 |
| institution | Directory Open Access Journal |
| issn | 2632-6736 |
| language | English |
| last_indexed | 2024-04-10T04:51:30Z |
| publishDate | 2022-01-01 |
| publisher | Cambridge University Press |
| record_format | Article |
| series | Data-Centric Engineering |
| spelling | doaj.art-6755315767f544b5b90ff590625396c12023-03-09T12:31:51ZengCambridge University PressData-Centric Engineering2632-67362022-01-01310.1017/dce.2022.19Multiphase flow applications of nonintrusive reduced-order models with Gaussian process emulationThemistoklis Botsas0https://orcid.org/0000-0001-5195-673XIndranil Pan1https://orcid.org/0000-0002-9624-5146Lachlan R. Mason2Omar K. Matar3Data Centric Engineering, The Alan Turing Institute, 96 Euston Rd, London NW1 2DB, United KingdomData Centric Engineering, The Alan Turing Institute, 96 Euston Rd, London NW1 2DB, United Kingdom Department of Chemical Engineering, Imperial College London, Exhibition Rd, South Kensington, London SW7 2BX, United Kingdom School of Mathematics, Statistics & Physics, Newcastle University, Newcastle upon Tyne NE1 7RU, United KingdomData Centric Engineering, The Alan Turing Institute, 96 Euston Rd, London NW1 2DB, United Kingdom Department of Chemical Engineering, Imperial College London, Exhibition Rd, South Kensington, London SW7 2BX, United KingdomData Centric Engineering, The Alan Turing Institute, 96 Euston Rd, London NW1 2DB, United Kingdom Department of Chemical Engineering, Imperial College London, Exhibition Rd, South Kensington, London SW7 2BX, United KingdomReduced-order models (ROMs) are computationally inexpensive simplifications of high-fidelity complex ones. Such models can be found in computational fluid dynamics where they can be used to predict the characteristics of multiphase flows. In previous work, we presented a ROM analysis framework that coupled compression techniques, such as autoencoders, with Gaussian process regression in the latent space. This pairing has significant advantages over the standard encoding–decoding routine, such as the ability to interpolate or extrapolate in the initial conditions’ space, which can provide predictions even when simulation data are not available. In this work, we focus on this major advantage and show its effectiveness by performing the pipeline on three multiphase flow applications. We also extend the methodology by using deep Gaussian processes as the interpolation algorithm and compare the performance of our two variations, as well as another variation from the literature that uses long short-term memory networks, for the interpolation.https://www.cambridge.org/core/product/identifier/S2632673622000193/type/journal_articleAutoencodersdeep Gaussian processGaussian processreduced-order models |
| spellingShingle | Themistoklis Botsas Indranil Pan Lachlan R. Mason Omar K. Matar Multiphase flow applications of nonintrusive reduced-order models with Gaussian process emulation Data-Centric Engineering Autoencoders deep Gaussian process Gaussian process reduced-order models |
| title | Multiphase flow applications of nonintrusive reduced-order models with Gaussian process emulation |
| title_full | Multiphase flow applications of nonintrusive reduced-order models with Gaussian process emulation |
| title_fullStr | Multiphase flow applications of nonintrusive reduced-order models with Gaussian process emulation |
| title_full_unstemmed | Multiphase flow applications of nonintrusive reduced-order models with Gaussian process emulation |
| title_short | Multiphase flow applications of nonintrusive reduced-order models with Gaussian process emulation |
| title_sort | multiphase flow applications of nonintrusive reduced order models with gaussian process emulation |
| topic | Autoencoders deep Gaussian process Gaussian process reduced-order models |
| url | https://www.cambridge.org/core/product/identifier/S2632673622000193/type/journal_article |
| work_keys_str_mv | AT themistoklisbotsas multiphaseflowapplicationsofnonintrusivereducedordermodelswithgaussianprocessemulation AT indranilpan multiphaseflowapplicationsofnonintrusivereducedordermodelswithgaussianprocessemulation AT lachlanrmason multiphaseflowapplicationsofnonintrusivereducedordermodelswithgaussianprocessemulation AT omarkmatar multiphaseflowapplicationsofnonintrusivereducedordermodelswithgaussianprocessemulation |