{"id":{"repo_id":"cambridge","oai_identifier":"oai:www.repository.cam.ac.uk:1810/371397"},"canonical_url":"https://search.dev.ndltd.org/etd/cambridge/oai:www.repository.cam.ac.uk:1810/371397","repository":{"repo_id":"cambridge","name":"Cambridge University","base_url":"https://api.repository.cam.ac.uk/server/oai/request"},"display":{"title":"From local and bulk structure-flow correlations to macro-scale transport properties with MRI: applications using deformable porous media and deep learning","abstract":"This thesis investigates the relationship between structure and flow in complex porous media using nuclear magnetic resonance (NMR)/magnetic resonance imaging (MRI) and machine learning (ML) techniques. The evolving interactions between the solid framework and the local flow fields of deformable model porous systems were probed by both spatially and non-spatially resolved NMR methods and linked to bulk transport properties such as permeability and dispersion. Three-dimensional (3D) spatially resolved flow MRI measurements combined with machine learning techniques, namely a feedforward artificial neural network (ANN), were employed to predict the absolute permeability of 21 rock samples of various lithology. To investigate structure-flow correlations in deformable model porous media, a custom-made bead pack with a moving piston was developed to allow loading and unloading of the constituent deformable particles. 3D MRI velocimetry was used to probe the evolving local hydrodynamics during loading conditions of a system composed of 8 mm diameter silicone spheres. After high compression stress a redistribution of flow was observed, with the fluid being distributed more evenly among the dominant flow channels. This phenomenon was linked to the slower rate of permeability decline at high stresses compared to lower stresses, which is often observed in rock samples. Permeability hysteresis and its relationships with irreversible structural and local transport phenomena were investigated. These studies were performed during unloading conditions of a previously compressed system using 3D spatially resolved MRI velocity measurements. It was argued that the degree to which fluid is distributed evenly between the dominant flow channels affects the evolution of permeability hysteresis. The evolution of bulk transport and structural properties was investigated during loading of a deformable model porous system (composed of 1.26 mm diameter expanded polystyrene spheres) by conducting non-spatially resolved NMR propagator experiments. It was found that dispersion increased with decreasing permeability and that it almost doubled with strain due to long-range heterogeneities. Macro-scale heterogeneities introduced by differential compaction were examined using one-dimensional (1D) spatially resolved MRI propagator measurements. It was also shown that strain intensifies the power law dependence of dispersion on Péclet number. Further, NMR flow diffraction phenomena revealed the evolution of structural properties with strain. A novel technique combining 3D spatially resolved MRI propagator measurements and a feedforward ANN was used as a means for predicting the absolute permeability of 21 rock samples of various lithology. The first three moments of each per-voxel propagator were utilised for training the ANN as to develop associations between the training data and the corresponding permeability values. A correlation coefficient (R) of 0.943 between the predicted and the actual permeability values, was achieved, demonstrating the link between the spatially resolved propagator measurement and absolute permeability.","abstract_html":"This thesis investigates the relationship between structure and flow in complex porous media using nuclear magnetic resonance (NMR)/magnetic resonance imaging (MRI) and machine learning (ML) techniques. The evolving interactions between the solid framework and the local flow fields of deformable model porous systems were probed by both spatially and non-spatially resolved NMR methods and linked to bulk transport properties such as permeability and dispersion. Three-dimensional (3D) spatially resolved flow MRI measurements combined with machine learning techniques, namely a feedforward artificial neural network (ANN), were employed to predict the absolute permeability of 21 rock samples of various lithology. To investigate structure-flow correlations in deformable model porous media, a custom-made bead pack with a moving piston was developed to allow loading and unloading of the constituent deformable particles. 3D MRI velocimetry was used to probe the evolving local hydrodynamics during loading conditions of a system composed of 8 mm diameter silicone spheres. After high compression stress a redistribution of flow was observed, with the fluid being distributed more evenly among the dominant flow channels. This phenomenon was linked to the slower rate of permeability decline at high stresses compared to lower stresses, which is often observed in rock samples. Permeability hysteresis and its relationships with irreversible structural and local transport phenomena were investigated. These studies were performed during unloading conditions of a previously compressed system using 3D spatially resolved MRI velocity measurements. It was argued that the degree to which fluid is distributed evenly between the dominant flow channels affects the evolution of permeability hysteresis. The evolution of bulk transport and structural properties was investigated during loading of a deformable model porous system (composed of 1.26 mm diameter expanded polystyrene spheres) by conducting non-spatially resolved NMR propagator experiments. It was found that dispersion increased with decreasing permeability and that it almost doubled with strain due to long-range heterogeneities. Macro-scale heterogeneities introduced by differential compaction were examined using one-dimensional (1D) spatially resolved MRI propagator measurements. It was also shown that strain intensifies the power law dependence of dispersion on Péclet number. Further, NMR flow diffraction phenomena revealed the evolution of structural properties with strain. A novel technique combining 3D spatially resolved MRI propagator measurements and a feedforward ANN was used as a means for predicting the absolute permeability of 21 rock samples of various lithology. The first three moments of each per-voxel propagator were utilised for training the ANN as to develop associations between the training data and the corresponding permeability values. A correlation coefficient (R) of 0.943 between the predicted and the actual permeability values, was achieved, demonstrating the link between the spatially resolved propagator measurement and absolute permeability.","abstract_has_math":false,"creators":["Avrantinis, Nikolaos"],"institution":"University of Cambridge","degree_name":"Doctor of Philosophy (PhD)","degree_level":"Doctoral","degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Gladden, Lynn","Sederman, Andrew"],"committee_chairs":[],"committee_members":[],"year":2023,"date_issued":"2023-11-28","date_published":"2023-11-28","updated_at":"2026-07-22T22:24:21Z","subjects":["deformable porous media","deformation","dispersion","flow fields","MRI","neural networks","permeability","porous media","strain","stress"],"languages":["eng"],"rights":[],"rights_urls":["https://apollo8-f-pro.lib.cam.ac.uk/bitstreams/46ccc7a1-449b-4537-bcfe-61d79ffca284/download","https://www.rioxx.net/licenses/all-rights-reserved/"],"identifier_entries":[]},"links":{"outbound_url":"https://doi.org/10.17863/CAM.110590","outbound_label":"DOI","outbound_source":"dc:identifier.doi"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Gladden, Lynn","Sederman, Andrew"]},{"key":"dc:contributor.sponsor","label":"Sponsor","values":["EPSRC Shell"]},{"key":"dc:creator","label":"Author","values":["Avrantinis, Nikolaos"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.issued","label":"Date","values":["2023-11-28"]},{"key":"dc:publisher.institution","label":"Dc Publisher Institution","values":["University of Cambridge"]},{"key":"dc:relation.isreferencedby.uri","label":"Dc Relation Isreferencedby URI","values":["https://www.repository.cam.ac.uk/handle/1810/371397"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"dc:type.qualificationlevel","label":"Dc Type Qualificationlevel","values":["Doctoral"]},{"key":"dc:type.qualificationname","label":"Dc Type Qualificationname","values":["Doctor of Philosophy (PhD)"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["deformable porous media","deformation","dispersion","flow fields","MRI","neural networks","permeability","porous media","strain","stress"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["https://apollo8-f-pro.lib.cam.ac.uk/bitstreams/46ccc7a1-449b-4537-bcfe-61d79ffca284/download","https://www.rioxx.net/licenses/all-rights-reserved/"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.doi","label":"DOI","values":["https://doi.org/10.17863/CAM.110590"]},{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://apollo8-f-pro.lib.cam.ac.uk/bitstreams/f3566016-2569-481e-9222-8776e61a50dd/download"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["This thesis investigates the relationship between structure and flow in complex porous media using nuclear magnetic resonance (NMR)/magnetic resonance imaging (MRI) and machine learning (ML) techniques. The evolving interactions between the solid framework and the local flow fields of deformable model porous systems were probed by both spatially and non-spatially resolved NMR methods and linked to bulk transport properties such as permeability and dispersion. Three-dimensional (3D) spatially resolved flow MRI measurements combined with machine learning techniques, namely a feedforward artificial neural network (ANN), were employed to predict the absolute permeability of 21 rock samples of various lithology. To investigate structure-flow correlations in deformable model porous media, a custom-made bead pack with a moving piston was developed to allow loading and unloading of the constituent deformable particles. 3D MRI velocimetry was used to probe the evolving local hydrodynamics during loading conditions of a system composed of 8 mm diameter silicone spheres. After high compression stress a redistribution of flow was observed, with the fluid being distributed more evenly among the dominant flow channels. This phenomenon was linked to the slower rate of permeability decline at high stresses compared to lower stresses, which is often observed in rock samples. Permeability hysteresis and its relationships with irreversible structural and local transport phenomena were investigated. These studies were performed during unloading conditions of a previously compressed system using 3D spatially resolved MRI velocity measurements. It was argued that the degree to which fluid is distributed evenly between the dominant flow channels affects the evolution of permeability hysteresis. The evolution of bulk transport and structural properties was investigated during loading of a deformable model porous system (composed of 1.26 mm diameter expanded polystyrene spheres) by conducting non-spatially resolved NMR propagator experiments. It was found that dispersion increased with decreasing permeability and that it almost doubled with strain due to long-range heterogeneities. Macro-scale heterogeneities introduced by differential compaction were examined using one-dimensional (1D) spatially resolved MRI propagator measurements. It was also shown that strain intensifies the power law dependence of dispersion on Péclet number. Further, NMR flow diffraction phenomena revealed the evolution of structural properties with strain. A novel technique combining 3D spatially resolved MRI propagator measurements and a feedforward ANN was used as a means for predicting the absolute permeability of 21 rock samples of various lithology. The first three moments of each per-voxel propagator were utilised for training the ANN as to develop associations between the training data and the corresponding permeability values. A correlation coefficient (R) of 0.943 between the predicted and the actual permeability values, was achieved, demonstrating the link between the spatially resolved propagator measurement and absolute permeability."]},{"key":"dc:format.checksum.md5","label":"Dc Format Checksum Md5","values":["e6295fee1dbfef8acc40c403a40bfd67","87eda9de84448d1f82354d60eee3eb5f"]},{"key":"dc:title","label":"Title","values":["From local and bulk structure-flow correlations to macro-scale transport properties with MRI: applications using deformable porous media and deep learning"]}]}],"canonical_facts":{"dc:contributor.advisor":["Gladden, Lynn","Sederman, Andrew"],"dc:contributor.sponsor":["EPSRC Shell"],"dc:creator":["Avrantinis, Nikolaos"],"dc:date.issued":["2023-11-28"],"dc:description.abstract":["This thesis investigates the relationship between structure and flow in complex porous media using nuclear magnetic resonance (NMR)/magnetic resonance imaging (MRI) and machine learning (ML) techniques. The evolving interactions between the solid framework and the local flow fields of deformable model porous systems were probed by both spatially and non-spatially resolved NMR methods and linked to bulk transport properties such as permeability and dispersion. Three-dimensional (3D) spatially resolved flow MRI measurements combined with machine learning techniques, namely a feedforward artificial neural network (ANN), were employed to predict the absolute permeability of 21 rock samples of various lithology. To investigate structure-flow correlations in deformable model porous media, a custom-made bead pack with a moving piston was developed to allow loading and unloading of the constituent deformable particles. 3D MRI velocimetry was used to probe the evolving local hydrodynamics during loading conditions of a system composed of 8 mm diameter silicone spheres. After high compression stress a redistribution of flow was observed, with the fluid being distributed more evenly among the dominant flow channels. This phenomenon was linked to the slower rate of permeability decline at high stresses compared to lower stresses, which is often observed in rock samples. Permeability hysteresis and its relationships with irreversible structural and local transport phenomena were investigated. These studies were performed during unloading conditions of a previously compressed system using 3D spatially resolved MRI velocity measurements. It was argued that the degree to which fluid is distributed evenly between the dominant flow channels affects the evolution of permeability hysteresis. The evolution of bulk transport and structural properties was investigated during loading of a deformable model porous system (composed of 1.26 mm diameter expanded polystyrene spheres) by conducting non-spatially resolved NMR propagator experiments. It was found that dispersion increased with decreasing permeability and that it almost doubled with strain due to long-range heterogeneities. Macro-scale heterogeneities introduced by differential compaction were examined using one-dimensional (1D) spatially resolved MRI propagator measurements. It was also shown that strain intensifies the power law dependence of dispersion on Péclet number. Further, NMR flow diffraction phenomena revealed the evolution of structural properties with strain. A novel technique combining 3D spatially resolved MRI propagator measurements and a feedforward ANN was used as a means for predicting the absolute permeability of 21 rock samples of various lithology. The first three moments of each per-voxel propagator were utilised for training the ANN as to develop associations between the training data and the corresponding permeability values. A correlation coefficient (R) of 0.943 between the predicted and the actual permeability values, was achieved, demonstrating the link between the spatially resolved propagator measurement and absolute permeability."],"dc:format.checksum.md5":["e6295fee1dbfef8acc40c403a40bfd67","87eda9de84448d1f82354d60eee3eb5f"],"dc:identifier.doi":["https://doi.org/10.17863/CAM.110590"],"dc:identifier.uri":["https://apollo8-f-pro.lib.cam.ac.uk/bitstreams/f3566016-2569-481e-9222-8776e61a50dd/download"],"dc:language":["eng"],"dc:publisher.institution":["University of Cambridge"],"dc:relation.isreferencedby.uri":["https://www.repository.cam.ac.uk/handle/1810/371397"],"dc:rights":["https://apollo8-f-pro.lib.cam.ac.uk/bitstreams/46ccc7a1-449b-4537-bcfe-61d79ffca284/download","https://www.rioxx.net/licenses/all-rights-reserved/"],"dc:subject":["deformable porous media","deformation","dispersion","flow fields","MRI","neural networks","permeability","porous media","strain","stress"],"dc:title":["From local and bulk structure-flow correlations to macro-scale transport properties with MRI: applications using deformable porous media and deep learning"],"dc:type":["Thesis"],"dc:type.qualificationlevel":["Doctoral"],"dc:type.qualificationname":["Doctor of Philosophy (PhD)"]},"updated_at":"2026-07-22T22:24:21Z"}