Predictive large-eddy-simulation wall modeling via physics-informed neural networks. Yang, X., I., A., Zafar, S., Wang, J., & Xiao, H. Physical Review Fluids, 4(3):34602, American Physical Society, 4, 2019. ![Predictive large-eddy-simulation wall modeling via physics-informed neural networks [link]](https://bibbase.org/img/filetypes/link.svg "link")
Website doi abstract bibtex 5 downloads While data-based approaches were found to be useful for subgrid scale (SGS) modeling in Reynolds-averaged Navier-Stokes (RANS) simulations, there have not been many attempts at using machine learning techniques for wall modeling in large-eddy simulations (LESs). Large-eddy simulation differs from RANS simulation in many aspects. For one thing, LES is scale resolving. For another, LES is in and of itself a high-fidelity tool. Because data sets of higher fidelity are in general not frequently accessible or available, this poses additional challenges to data-based modeling in LES. Further, SGS modeling usually needs flow information at only large scales, in contrast with wall modeling, which needs to account for both near-wall small scales and large scales above the wall. In this work we discuss how the above-noted challenges may be addressed when taking a data-based approach for wall modeling. We also show the necessity of incorporating physical insights in model inputs, i.e., using inputs that are inspired by the vertically integrated thin-boundary-layer equations and the eddy population density scalings. We show that the inclusion of the above physics-based considerations would enhance extrapolation capabilities of a neural network to flow conditions that are not within the train data. Being cheap to evaluate and using only channel flow data at Reτ=1000, the trained networks are found to capture the law of the wall at arbitrary Reynolds numbers and outperform the conventional equilibrium model in a nonequilibrium flow.
@article{
title = {Predictive large-eddy-simulation wall modeling via physics-informed neural networks},
type = {article},
year = {2019},
pages = {34602},
volume = {4},
websites = {https://link.aps.org/doi/10.1103/PhysRevFluids.4.034602},
month = {4},
publisher = {American Physical Society},
id = {3981fbac-dbaf-3646-acef-d3d887d228e4},
created = {2021-04-09T15:24:52.317Z},
file_attached = {false},
profile_id = {75799766-8e2d-3c98-81f9-e3efa41233d0},
group_id = {c9329632-2a50-3043-b803-cadc8dbdfc3f},
last_modified = {2021-04-09T15:24:52.317Z},
read = {false},
starred = {false},
authored = {false},
confirmed = {false},
hidden = {false},
source_type = {article},
private_publication = {false},
abstract = {While data-based approaches were found to be useful for subgrid scale (SGS) modeling in Reynolds-averaged Navier-Stokes (RANS) simulations, there have not been many attempts at using machine learning techniques for wall modeling in large-eddy simulations (LESs). Large-eddy simulation differs from RANS simulation in many aspects. For one thing, LES is scale resolving. For another, LES is in and of itself a high-fidelity tool. Because data sets of higher fidelity are in general not frequently accessible or available, this poses additional challenges to data-based modeling in LES. Further, SGS modeling usually needs flow information at only large scales, in contrast with wall modeling, which needs to account for both near-wall small scales and large scales above the wall. In this work we discuss how the above-noted challenges may be addressed when taking a data-based approach for wall modeling. We also show the necessity of incorporating physical insights in model inputs, i.e., using inputs that are inspired by the vertically integrated thin-boundary-layer equations and the eddy population density scalings. We show that the inclusion of the above physics-based considerations would enhance extrapolation capabilities of a neural network to flow conditions that are not within the train data. Being cheap to evaluate and using only channel flow data at Reτ=1000, the trained networks are found to capture the law of the wall at arbitrary Reynolds numbers and outperform the conventional equilibrium model in a nonequilibrium flow.},
bibtype = {article},
author = {Yang, X I A and Zafar, S and Wang, J.-X. and Xiao, H},
doi = {10.1103/PhysRevFluids.4.034602},
journal = {Physical Review Fluids},
number = {3}
}
Downloads: 5
{"_id":"FN7KoXTZxR2KMAyBd","bibbaseid":"yang-zafar-wang-xiao-predictivelargeeddysimulationwallmodelingviaphysicsinformedneuralnetworks-2019","authorIDs":[],"author_short":["Yang, X., I., A.","Zafar, S.","Wang, J.","Xiao, H."],"bibdata":{"title":"Predictive large-eddy-simulation wall modeling via physics-informed neural networks","type":"article","year":"2019","pages":"34602","volume":"4","websites":"https://link.aps.org/doi/10.1103/PhysRevFluids.4.034602","month":"4","publisher":"American Physical Society","id":"3981fbac-dbaf-3646-acef-d3d887d228e4","created":"2021-04-09T15:24:52.317Z","file_attached":false,"profile_id":"75799766-8e2d-3c98-81f9-e3efa41233d0","group_id":"c9329632-2a50-3043-b803-cadc8dbdfc3f","last_modified":"2021-04-09T15:24:52.317Z","read":false,"starred":false,"authored":false,"confirmed":false,"hidden":false,"source_type":"article","private_publication":false,"abstract":"While data-based approaches were found to be useful for subgrid scale (SGS) modeling in Reynolds-averaged Navier-Stokes (RANS) simulations, there have not been many attempts at using machine learning techniques for wall modeling in large-eddy simulations (LESs). Large-eddy simulation differs from RANS simulation in many aspects. For one thing, LES is scale resolving. For another, LES is in and of itself a high-fidelity tool. Because data sets of higher fidelity are in general not frequently accessible or available, this poses additional challenges to data-based modeling in LES. Further, SGS modeling usually needs flow information at only large scales, in contrast with wall modeling, which needs to account for both near-wall small scales and large scales above the wall. In this work we discuss how the above-noted challenges may be addressed when taking a data-based approach for wall modeling. We also show the necessity of incorporating physical insights in model inputs, i.e., using inputs that are inspired by the vertically integrated thin-boundary-layer equations and the eddy population density scalings. We show that the inclusion of the above physics-based considerations would enhance extrapolation capabilities of a neural network to flow conditions that are not within the train data. Being cheap to evaluate and using only channel flow data at Reτ=1000, the trained networks are found to capture the law of the wall at arbitrary Reynolds numbers and outperform the conventional equilibrium model in a nonequilibrium flow.","bibtype":"article","author":"Yang, X I A and Zafar, S and Wang, J.-X. and Xiao, H","doi":"10.1103/PhysRevFluids.4.034602","journal":"Physical Review Fluids","number":"3","bibtex":"@article{\n title = {Predictive large-eddy-simulation wall modeling via physics-informed neural networks},\n type = {article},\n year = {2019},\n pages = {34602},\n volume = {4},\n websites = {https://link.aps.org/doi/10.1103/PhysRevFluids.4.034602},\n month = {4},\n publisher = {American Physical Society},\n id = {3981fbac-dbaf-3646-acef-d3d887d228e4},\n created = {2021-04-09T15:24:52.317Z},\n file_attached = {false},\n profile_id = {75799766-8e2d-3c98-81f9-e3efa41233d0},\n group_id = {c9329632-2a50-3043-b803-cadc8dbdfc3f},\n last_modified = {2021-04-09T15:24:52.317Z},\n read = {false},\n starred = {false},\n authored = {false},\n confirmed = {false},\n hidden = {false},\n source_type = {article},\n private_publication = {false},\n abstract = {While data-based approaches were found to be useful for subgrid scale (SGS) modeling in Reynolds-averaged Navier-Stokes (RANS) simulations, there have not been many attempts at using machine learning techniques for wall modeling in large-eddy simulations (LESs). Large-eddy simulation differs from RANS simulation in many aspects. For one thing, LES is scale resolving. For another, LES is in and of itself a high-fidelity tool. Because data sets of higher fidelity are in general not frequently accessible or available, this poses additional challenges to data-based modeling in LES. Further, SGS modeling usually needs flow information at only large scales, in contrast with wall modeling, which needs to account for both near-wall small scales and large scales above the wall. In this work we discuss how the above-noted challenges may be addressed when taking a data-based approach for wall modeling. We also show the necessity of incorporating physical insights in model inputs, i.e., using inputs that are inspired by the vertically integrated thin-boundary-layer equations and the eddy population density scalings. We show that the inclusion of the above physics-based considerations would enhance extrapolation capabilities of a neural network to flow conditions that are not within the train data. Being cheap to evaluate and using only channel flow data at Reτ=1000, the trained networks are found to capture the law of the wall at arbitrary Reynolds numbers and outperform the conventional equilibrium model in a nonequilibrium flow.},\n bibtype = {article},\n author = {Yang, X I A and Zafar, S and Wang, J.-X. and Xiao, H},\n doi = {10.1103/PhysRevFluids.4.034602},\n journal = {Physical Review Fluids},\n number = {3}\n}","author_short":["Yang, X., I., A.","Zafar, S.","Wang, J.","Xiao, H."],"urls":{"Website":"https://link.aps.org/doi/10.1103/PhysRevFluids.4.034602"},"biburl":"https://bibbase.org/service/mendeley/75799766-8e2d-3c98-81f9-e3efa41233d0","bibbaseid":"yang-zafar-wang-xiao-predictivelargeeddysimulationwallmodelingviaphysicsinformedneuralnetworks-2019","role":"author","metadata":{"authorlinks":{}},"downloads":5},"bibtype":"article","creationDate":"2020-05-26T19:26:45.513Z","downloads":5,"keywords":[],"search_terms":["predictive","large","eddy","simulation","wall","modeling","via","physics","informed","neural","networks","yang","zafar","wang","xiao"],"title":"Predictive large-eddy-simulation wall modeling via physics-informed neural networks","year":2019,"biburl":"https://bibbase.org/service/mendeley/75799766-8e2d-3c98-81f9-e3efa41233d0","dataSources":["hY2XmpNnxf2BGtC9Z","ya2CyA73rpZseyrZ8","2252seNhipfTmjEBQ"]}