Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations. Raissi, M., Yazdani, A., & Karniadakis, G. E. Science, 367(6481):1026–1030, February, 2020.
![Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex
Paper doi abstract bibtex For centuries, flow visualization has been the art of making fluid motion visible in physical and biological systems. Although such flow patterns can be, in principle, described by the NavierStokes equations, extracting the velocity and pressure fields directly from the images is challenging. We addressed this problem by developing hidden fluid mechanics (HFM), a physicsinformed deep-learning framework capable of encoding the Navier-Stokes equations into the neural networks while being agnostic to the geometry or the initial and boundary conditions. We demonstrate HFM for several physical and biomedical problems by extracting quantitative information for which direct measurements may not be possible. HFM is robust to low resolution and substantial noise in the observation data, which is important for potential applications.
@article{raissi_hidden_2020,
title = {Hidden fluid mechanics: {Learning} velocity and pressure fields from flow visualizations},
volume = {367},
issn = {0036-8075, 1095-9203},
shorttitle = {Hidden fluid mechanics},
url = {https://www.science.org/doi/10.1126/science.aaw4741},
doi = {10.1126/science.aaw4741},
abstract = {For centuries, flow visualization has been the art of making fluid motion visible in physical and biological systems. Although such flow patterns can be, in principle, described by the NavierStokes equations, extracting the velocity and pressure fields directly from the images is challenging. We addressed this problem by developing hidden fluid mechanics (HFM), a physicsinformed deep-learning framework capable of encoding the Navier-Stokes equations into the neural networks while being agnostic to the geometry or the initial and boundary conditions. We demonstrate HFM for several physical and biomedical problems by extracting quantitative information for which direct measurements may not be possible. HFM is robust to low resolution and substantial noise in the observation data, which is important for potential applications.},
language = {en},
number = {6481},
urldate = {2025-04-12},
journal = {Science},
author = {Raissi, Maziar and Yazdani, Alireza and Karniadakis, George Em},
month = feb,
year = {2020},
pages = {1026--1030},
file = {PDF:/Users/tkaiser/Zotero/storage/SEQBP7ZM/Raissi et al. - 2020 - Hidden fluid mechanics Learning velocity and pressure fields from flow visualizations.pdf:application/pdf},
}Downloads: 0
{"_id":"LxLmFct7jtQJEg6LN","bibbaseid":"raissi-yazdani-karniadakis-hiddenfluidmechanicslearningvelocityandpressurefieldsfromflowvisualizations-2020","author_short":["Raissi, M.","Yazdani, A.","Karniadakis, G. E."],"bibdata":{"bibtype":"article","type":"article","title":"Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations","volume":"367","issn":"0036-8075, 1095-9203","shorttitle":"Hidden fluid mechanics","url":"https://www.science.org/doi/10.1126/science.aaw4741","doi":"10.1126/science.aaw4741","abstract":"For centuries, flow visualization has been the art of making fluid motion visible in physical and biological systems. Although such flow patterns can be, in principle, described by the NavierStokes equations, extracting the velocity and pressure fields directly from the images is challenging. We addressed this problem by developing hidden fluid mechanics (HFM), a physicsinformed deep-learning framework capable of encoding the Navier-Stokes equations into the neural networks while being agnostic to the geometry or the initial and boundary conditions. We demonstrate HFM for several physical and biomedical problems by extracting quantitative information for which direct measurements may not be possible. HFM is robust to low resolution and substantial noise in the observation data, which is important for potential applications.","language":"en","number":"6481","urldate":"2025-04-12","journal":"Science","author":[{"propositions":[],"lastnames":["Raissi"],"firstnames":["Maziar"],"suffixes":[]},{"propositions":[],"lastnames":["Yazdani"],"firstnames":["Alireza"],"suffixes":[]},{"propositions":[],"lastnames":["Karniadakis"],"firstnames":["George","Em"],"suffixes":[]}],"month":"February","year":"2020","pages":"1026–1030","file":"PDF:/Users/tkaiser/Zotero/storage/SEQBP7ZM/Raissi et al. - 2020 - Hidden fluid mechanics Learning velocity and pressure fields from flow visualizations.pdf:application/pdf","bibtex":"@article{raissi_hidden_2020,\n\ttitle = {Hidden fluid mechanics: {Learning} velocity and pressure fields from flow visualizations},\n\tvolume = {367},\n\tissn = {0036-8075, 1095-9203},\n\tshorttitle = {Hidden fluid mechanics},\n\turl = {https://www.science.org/doi/10.1126/science.aaw4741},\n\tdoi = {10.1126/science.aaw4741},\n\tabstract = {For centuries, flow visualization has been the art of making fluid motion visible in physical and biological systems. Although such flow patterns can be, in principle, described by the NavierStokes equations, extracting the velocity and pressure fields directly from the images is challenging. We addressed this problem by developing hidden fluid mechanics (HFM), a physicsinformed deep-learning framework capable of encoding the Navier-Stokes equations into the neural networks while being agnostic to the geometry or the initial and boundary conditions. We demonstrate HFM for several physical and biomedical problems by extracting quantitative information for which direct measurements may not be possible. HFM is robust to low resolution and substantial noise in the observation data, which is important for potential applications.},\n\tlanguage = {en},\n\tnumber = {6481},\n\turldate = {2025-04-12},\n\tjournal = {Science},\n\tauthor = {Raissi, Maziar and Yazdani, Alireza and Karniadakis, George Em},\n\tmonth = feb,\n\tyear = {2020},\n\tpages = {1026--1030},\n\tfile = {PDF:/Users/tkaiser/Zotero/storage/SEQBP7ZM/Raissi et al. - 2020 - Hidden fluid mechanics Learning velocity and pressure fields from flow visualizations.pdf:application/pdf},\n}\n\n","author_short":["Raissi, M.","Yazdani, A.","Karniadakis, G. E."],"key":"raissi_hidden_2020","id":"raissi_hidden_2020","bibbaseid":"raissi-yazdani-karniadakis-hiddenfluidmechanicslearningvelocityandpressurefieldsfromflowvisualizations-2020","role":"author","urls":{"Paper":"https://www.science.org/doi/10.1126/science.aaw4741"},"metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://bibbase.org/network/files/KCdmLiYSME4xCdEx2","dataSources":["b7opvToJ7XWt8fFiB"],"keywords":[],"search_terms":["hidden","fluid","mechanics","learning","velocity","pressure","fields","flow","visualizations","raissi","yazdani","karniadakis"],"title":"Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations","year":2020}