{"id":{"repo_id":"sask","oai_identifier":"oai:harvest.usask.ca:10388/18299"},"canonical_url":"https://search.dev.ndltd.org/etd/sask/oai:harvest.usask.ca:10388/18299","repository":{"repo_id":"sask","name":"University of Saskatchewan","base_url":"https://harvest.usask.ca/server/oai/request"},"display":{"title":"Geometry-Aware Modeling of Rough-Wall Turbulence Using Reynolds-Averaged Navier-Stokes (RANS) Frameworks","abstract":"Turbulent flow over rough walls is widespread in engineering systems and plays a critical role in drag prediction, energy efficiency, and transport processes in internal and external flows. Despite extensive research, the representation of roughness effects in Reynolds-Averaged Navier–Stokes (RANS) turbulence models remains fundamentally challenging. Classical approaches rely on hydrodynamic equivalence, typically through the equivalent sand-grain roughness height and empirical shifts of the logarithmic velocity profile, which can reproduce bulk quantities for specific surface classes. The mechanisms by which surface geometry alters near-wall momentum transfer are often absorbed implicitly into modified boundary conditions (BCs) or an enhanced eddy viscosity, obscuring the role of pressure drag and geometry-induced flow structures. Surface geometry is not resolved in RANS explicitly; instead, its influence must be introduced indirectly through modeling assumptions at the wall. This thesis investigates how rough-wall effects can be represented within RANS frameworks beginning with fully developed internal flows. The work is structured around three interconnected studies that progressively address model assessment, near-wall representation, and geometry-aware parameterization. The first study evaluates the performance of RANS turbulence models including the SST k-ω and v²-f-k-ω models for turbulent channel flow over rough walls, with particular attention to how roughness is introduced through BCs. This analysis reveals a structural limitation shared across formulations: roughness effects are primarily absorbed through modified wall treatments, with limited transparency regarding how surface geometry influences turbulence momentum transfer. A new BC was introduced for the v²-f-k-ω model in which wall-normal stress is prescribed using relations based on both equivalent sand-grain roughness and experimentally derived surface metrics. This formulation demonstrated agreement with Direct Numerical Simulation (DNS)-generated roughness functions and friction factors, particularly in the transitionally rough regime, highlighting the sensitivity of model predictions to how roughness information is specified. Despite the agreement in global quantities, the near-wall velocity distribution remained ambiguous. This indicates that BC adjustments alone cannot fully clarify how roughness modifies the structure of the flow in the roughness sublayer. The second study introduces a reference-plane framework that explicitly partitions the flow into a geometry-dominated near-wall inner region and an outer region governed by the k-ε model. The reference plane is placed at the equivalent sand-grain roughness height, with velocity and turbulence quantities at this interface derived using high-fidelity DNS data. The inner region is represented kinematically through mean velocity-profile formulations calibrated against surface metrics obtained from an extensive DNS database. In addition, BCs imposed at the reference plane are formulated directly from surface metrics, providing a geometry-aware transmission of near-wall effects into the outer flow. The results show that properly resolving the portion of the flow below the reference plane can improve predictions for bulk quantities such as bulk velocity and friction factor. The reference-plane framework still includes the roughness function and equivalent sand-grain roughness as inputs. The third study addresses this limitation by developing geometry-aware correlations for the roughness function using a comprehensive database. A dataset comprising 172 DNS and Large Eddy Simulation (LES) cases spanning a wide range of roughness morphologies was assembled and analyzed. The analysis shows that height-based metrics alone are insufficient to capture roughness effects and that predictive capability emerges from the combined influence of roughness amplitude, effective slope, and surface asymmetry. Compact non-linear regression models are proposed that operate directly in roughness function space and reduce scatter relative to existing correlations. Because these relations predict the roughness function from metrics, they provide a pathway for introducing geometry effects into RANS models without relying exclusively on equivalent sand-grain scaling. The proposed developments establish that: 1) BC formulations informed by surface metrics can maintain the predictive capability, 2) a two-layer reference-plane representation allows roughness effects to be transmitted to the outer flow, and 3) morphology-based roughness function correlations provide an alternative and improved approach compare to single-parameter sand-grain scaling. These findings define a practical pathway toward geometry-aware roughness modeling in RANS.","abstract_html":"Turbulent flow over rough walls is widespread in engineering systems and plays a critical role in drag prediction, energy efficiency, and transport processes in internal and external flows. Despite extensive research, the representation of roughness effects in Reynolds-Averaged Navier–Stokes (RANS) turbulence models remains fundamentally challenging. Classical approaches rely on hydrodynamic equivalence, typically through the equivalent sand-grain roughness height and empirical shifts of the logarithmic velocity profile, which can reproduce bulk quantities for specific surface classes. The mechanisms by which surface geometry alters near-wall momentum transfer are often absorbed implicitly into modified boundary conditions (BCs) or an enhanced eddy viscosity, obscuring the role of pressure drag and geometry-induced flow structures. Surface geometry is not resolved in RANS explicitly; instead, its influence must be introduced indirectly through modeling assumptions at the wall. This thesis investigates how rough-wall effects can be represented within RANS frameworks beginning with fully developed internal flows. The work is structured around three interconnected studies that progressively address model assessment, near-wall representation, and geometry-aware parameterization. The first study evaluates the performance of RANS turbulence models including the SST k-ω and v²-f-k-ω models for turbulent channel flow over rough walls, with particular attention to how roughness is introduced through BCs. This analysis reveals a structural limitation shared across formulations: roughness effects are primarily absorbed through modified wall treatments, with limited transparency regarding how surface geometry influences turbulence momentum transfer. A new BC was introduced for the v²-f-k-ω model in which wall-normal stress is prescribed using relations based on both equivalent sand-grain roughness and experimentally derived surface metrics. This formulation demonstrated agreement with Direct Numerical Simulation (DNS)-generated roughness functions and friction factors, particularly in the transitionally rough regime, highlighting the sensitivity of model predictions to how roughness information is specified. Despite the agreement in global quantities, the near-wall velocity distribution remained ambiguous. This indicates that BC adjustments alone cannot fully clarify how roughness modifies the structure of the flow in the roughness sublayer. The second study introduces a reference-plane framework that explicitly partitions the flow into a geometry-dominated near-wall inner region and an outer region governed by the k-ε model. The reference plane is placed at the equivalent sand-grain roughness height, with velocity and turbulence quantities at this interface derived using high-fidelity DNS data. The inner region is represented kinematically through mean velocity-profile formulations calibrated against surface metrics obtained from an extensive DNS database. In addition, BCs imposed at the reference plane are formulated directly from surface metrics, providing a geometry-aware transmission of near-wall effects into the outer flow. The results show that properly resolving the portion of the flow below the reference plane can improve predictions for bulk quantities such as bulk velocity and friction factor. The reference-plane framework still includes the roughness function and equivalent sand-grain roughness as inputs. The third study addresses this limitation by developing geometry-aware correlations for the roughness function using a comprehensive database. A dataset comprising 172 DNS and Large Eddy Simulation (LES) cases spanning a wide range of roughness morphologies was assembled and analyzed. The analysis shows that height-based metrics alone are insufficient to capture roughness effects and that predictive capability emerges from the combined influence of roughness amplitude, effective slope, and surface asymmetry. Compact non-linear regression models are proposed that operate directly in roughness function space and reduce scatter relative to existing correlations. Because these relations predict the roughness function from metrics, they provide a pathway for introducing geometry effects into RANS models without relying exclusively on equivalent sand-grain scaling. The proposed developments establish that: 1) BC formulations informed by surface metrics can maintain the predictive capability, 2) a two-layer reference-plane representation allows roughness effects to be transmitted to the outer flow, and 3) morphology-based roughness function correlations provide an alternative and improved approach compare to single-parameter sand-grain scaling. These findings define a practical pathway toward geometry-aware roughness modeling in RANS.","abstract_has_math":false,"creators":["Sojoudi, Ata"],"institution":"University of Saskatchewan","degree_name":"Doctor of Philosophy (Ph.D.)","degree_level":"Doctoral","degree_discipline":"Mechanical Engineering","degree_department":null,"school":null,"contributors":[],"advisors":["Bergstrom, Donald"],"committee_chairs":[],"committee_members":["Yang, Qiaoqin","Sumner, David","Evitts, Richard","Ghaemi, Sina"],"year":2026,"date_issued":"2026-04-28","date_published":"2026-04-28","updated_at":"2026-07-24T04:26:45Z","subjects":["Turbulent channel flow","wall roughness","turbulence modelling"],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/10388/18299","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Bergstrom, Donald"]},{"key":"dc:contributor.committeemember","label":"Committee Member","values":["Yang, Qiaoqin","Sumner, David","Evitts, Richard","Ghaemi, Sina"]},{"key":"dc:creator","label":"Author","values":["Sojoudi, Ata"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2026-04-28T15:41:53Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2026-04-28T15:41:53Z"]},{"key":"dc:date.issued","label":"Date","values":["2026-04-28"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Mechanical Engineering"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Doctoral"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Doctor of Philosophy (Ph.D.)"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Saskatchewan"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Turbulent channel flow","wall roughness","turbulence modelling"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/10388/18299"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Turbulent flow over rough walls is widespread in engineering systems and plays a critical role in drag prediction, energy efficiency, and transport processes in internal and external flows. Despite extensive research, the representation of roughness effects in Reynolds-Averaged Navier–Stokes (RANS) turbulence models remains fundamentally challenging. Classical approaches rely on hydrodynamic equivalence, typically through the equivalent sand-grain roughness height and empirical shifts of the logarithmic velocity profile, which can reproduce bulk quantities for specific surface classes. The mechanisms by which surface geometry alters near-wall momentum transfer are often absorbed implicitly into modified boundary conditions (BCs) or an enhanced eddy viscosity, obscuring the role of pressure drag and geometry-induced flow structures. Surface geometry is not resolved in RANS explicitly; instead, its influence must be introduced indirectly through modeling assumptions at the wall. This thesis investigates how rough-wall effects can be represented within RANS frameworks beginning with fully developed internal flows. The work is structured around three interconnected studies that progressively address model assessment, near-wall representation, and geometry-aware parameterization. The first study evaluates the performance of RANS turbulence models including the SST k-ω and v²-f-k-ω models for turbulent channel flow over rough walls, with particular attention to how roughness is introduced through BCs. This analysis reveals a structural limitation shared across formulations: roughness effects are primarily absorbed through modified wall treatments, with limited transparency regarding how surface geometry influences turbulence momentum transfer. A new BC was introduced for the v²-f-k-ω model in which wall-normal stress is prescribed using relations based on both equivalent sand-grain roughness and experimentally derived surface metrics. This formulation demonstrated agreement with Direct Numerical Simulation (DNS)-generated roughness functions and friction factors, particularly in the transitionally rough regime, highlighting the sensitivity of model predictions to how roughness information is specified. Despite the agreement in global quantities, the near-wall velocity distribution remained ambiguous. This indicates that BC adjustments alone cannot fully clarify how roughness modifies the structure of the flow in the roughness sublayer. The second study introduces a reference-plane framework that explicitly partitions the flow into a geometry-dominated near-wall inner region and an outer region governed by the k-ε model. The reference plane is placed at the equivalent sand-grain roughness height, with velocity and turbulence quantities at this interface derived using high-fidelity DNS data. The inner region is represented kinematically through mean velocity-profile formulations calibrated against surface metrics obtained from an extensive DNS database. In addition, BCs imposed at the reference plane are formulated directly from surface metrics, providing a geometry-aware transmission of near-wall effects into the outer flow. The results show that properly resolving the portion of the flow below the reference plane can improve predictions for bulk quantities such as bulk velocity and friction factor. The reference-plane framework still includes the roughness function and equivalent sand-grain roughness as inputs. The third study addresses this limitation by developing geometry-aware correlations for the roughness function using a comprehensive database. A dataset comprising 172 DNS and Large Eddy Simulation (LES) cases spanning a wide range of roughness morphologies was assembled and analyzed. The analysis shows that height-based metrics alone are insufficient to capture roughness effects and that predictive capability emerges from the combined influence of roughness amplitude, effective slope, and surface asymmetry. Compact non-linear regression models are proposed that operate directly in roughness function space and reduce scatter relative to existing correlations. Because these relations predict the roughness function from metrics, they provide a pathway for introducing geometry effects into RANS models without relying exclusively on equivalent sand-grain scaling. The proposed developments establish that: 1) BC formulations informed by surface metrics can maintain the predictive capability, 2) a two-layer reference-plane representation allows roughness effects to be transmitted to the outer flow, and 3) morphology-based roughness function correlations provide an alternative and improved approach compare to single-parameter sand-grain scaling. These findings define a practical pathway toward geometry-aware roughness modeling in RANS."]},{"key":"dc:format.mimetype","label":"Dc Format Mimetype","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Geometry-Aware Modeling of Rough-Wall Turbulence Using Reynolds-Averaged Navier-Stokes (RANS) Frameworks"]}]}],"canonical_facts":{"dc:contributor.advisor":["Bergstrom, Donald"],"dc:contributor.committeemember":["Yang, Qiaoqin","Sumner, David","Evitts, Richard","Ghaemi, Sina"],"dc:creator":["Sojoudi, Ata"],"dc:date.accessioned":["2026-04-28T15:41:53Z"],"dc:date.available":["2026-04-28T15:41:53Z"],"dc:date.issued":["2026-04-28"],"dc:description.abstract":["Turbulent flow over rough walls is widespread in engineering systems and plays a critical role in drag prediction, energy efficiency, and transport processes in internal and external flows. Despite extensive research, the representation of roughness effects in Reynolds-Averaged Navier–Stokes (RANS) turbulence models remains fundamentally challenging. Classical approaches rely on hydrodynamic equivalence, typically through the equivalent sand-grain roughness height and empirical shifts of the logarithmic velocity profile, which can reproduce bulk quantities for specific surface classes. The mechanisms by which surface geometry alters near-wall momentum transfer are often absorbed implicitly into modified boundary conditions (BCs) or an enhanced eddy viscosity, obscuring the role of pressure drag and geometry-induced flow structures. Surface geometry is not resolved in RANS explicitly; instead, its influence must be introduced indirectly through modeling assumptions at the wall. This thesis investigates how rough-wall effects can be represented within RANS frameworks beginning with fully developed internal flows. The work is structured around three interconnected studies that progressively address model assessment, near-wall representation, and geometry-aware parameterization. The first study evaluates the performance of RANS turbulence models including the SST k-ω and v²-f-k-ω models for turbulent channel flow over rough walls, with particular attention to how roughness is introduced through BCs. This analysis reveals a structural limitation shared across formulations: roughness effects are primarily absorbed through modified wall treatments, with limited transparency regarding how surface geometry influences turbulence momentum transfer. A new BC was introduced for the v²-f-k-ω model in which wall-normal stress is prescribed using relations based on both equivalent sand-grain roughness and experimentally derived surface metrics. This formulation demonstrated agreement with Direct Numerical Simulation (DNS)-generated roughness functions and friction factors, particularly in the transitionally rough regime, highlighting the sensitivity of model predictions to how roughness information is specified. Despite the agreement in global quantities, the near-wall velocity distribution remained ambiguous. This indicates that BC adjustments alone cannot fully clarify how roughness modifies the structure of the flow in the roughness sublayer. The second study introduces a reference-plane framework that explicitly partitions the flow into a geometry-dominated near-wall inner region and an outer region governed by the k-ε model. The reference plane is placed at the equivalent sand-grain roughness height, with velocity and turbulence quantities at this interface derived using high-fidelity DNS data. The inner region is represented kinematically through mean velocity-profile formulations calibrated against surface metrics obtained from an extensive DNS database. In addition, BCs imposed at the reference plane are formulated directly from surface metrics, providing a geometry-aware transmission of near-wall effects into the outer flow. The results show that properly resolving the portion of the flow below the reference plane can improve predictions for bulk quantities such as bulk velocity and friction factor. The reference-plane framework still includes the roughness function and equivalent sand-grain roughness as inputs. The third study addresses this limitation by developing geometry-aware correlations for the roughness function using a comprehensive database. A dataset comprising 172 DNS and Large Eddy Simulation (LES) cases spanning a wide range of roughness morphologies was assembled and analyzed. The analysis shows that height-based metrics alone are insufficient to capture roughness effects and that predictive capability emerges from the combined influence of roughness amplitude, effective slope, and surface asymmetry. Compact non-linear regression models are proposed that operate directly in roughness function space and reduce scatter relative to existing correlations. Because these relations predict the roughness function from metrics, they provide a pathway for introducing geometry effects into RANS models without relying exclusively on equivalent sand-grain scaling. The proposed developments establish that: 1) BC formulations informed by surface metrics can maintain the predictive capability, 2) a two-layer reference-plane representation allows roughness effects to be transmitted to the outer flow, and 3) morphology-based roughness function correlations provide an alternative and improved approach compare to single-parameter sand-grain scaling. These findings define a practical pathway toward geometry-aware roughness modeling in RANS."],"dc:format.mimetype":["application/pdf"],"dc:identifier.uri":["https://hdl.handle.net/10388/18299"],"dc:language.iso":["en"],"dc:subject":["Turbulent channel flow","wall roughness","turbulence modelling"],"dc:title":["Geometry-Aware Modeling of Rough-Wall Turbulence Using Reynolds-Averaged Navier-Stokes (RANS) Frameworks"],"dc:type":["Thesis"],"thesis:degree_discipline":["Mechanical Engineering"],"thesis:degree_level":["Doctoral"],"thesis:degree_name":["Doctor of Philosophy (Ph.D.)"],"thesis:institution_name":["University of Saskatchewan"]},"updated_at":"2026-07-24T04:26:45Z"}