ORCID Identifier(s)

0000-0003-0969-0848

Graduation Semester and Year

Summer 2026

Language

English

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Civil Engineering

Department

Civil Engineering

First Advisor

Habib Ahmari

Second Advisor

Nick Fang

Third Advisor

Xinbao Yu

Fourth Advisor

Behzad Ghanbarian

Abstract

Bridge scour threatens transportation infrastructure during floods, particularly when water submerges a bridge deck and forces the flow through the opening beneath the superstructure. The resulting pressure-flow condition produced by this vertical contraction can intensify hydraulic forces and cause scour. Although pressure-flow scour has been investigated in non-cohesive sediments, limited information is available for cohesive streambeds, where erosion resistance depends on soil composition, mineralogy, moisture, compaction, plasticity, and strength. This dissertation investigated the hydraulic, geometric, and soil controls on vertical contraction scour beneath submerged bridge decks and integrated the findings into a predictive framework.

Three coordinated experimental programs were conducted in a laboratory flume under clear-water conditions. Forty unique pressure-flow experiments evaluated the effects of flow intensity, inundation ratio, proximity ratio, girder presence, blockage ratio, aspect ratio, soil type, plasticity, clay content, initial moisture content, degree of compaction, and soil strength. Kaolinite clay, Ball clay, and Kaolinite-sand mixtures were used to represent cohesive-bed compositions and preparation conditions. Scour response was characterized in terms of scour depth, planform area, eroded volume, and temporal development. A consolidated database of 37 experiments was analyzed through dimensional analysis, physical screening, and statistical modeling to develop an equation for maximum scour depth.

The hydraulic experiments showed that cohesive-bed scour was localized and time-dependent, initiating at weak zones and expanding progressively. Increasing Froude number increased scour severity, while greater inundation generally increased maximum depth. Under comparable hydraulic conditions, lower proximity ratios intensified vertical contraction and increased scour. The geometry experiments demonstrated that girder presence had limited influence on maximum depth when total superstructure height remained constant. Blockage ratio was the primary geometric control on scour depth, whereas aspect ratio more strongly influenced scour hole extent and eroded volume. The soil-property experiments showed that increasing compaction reduced scour, although its effectiveness depended on mineralogy and plasticity. Moisture content produced a nonmonotonic response, with minimum scour near the optimum moisture condition. Kaolinite–sand mixtures also exhibited nonlinear behavior, with the mixture containing 75% Kaolinite clay and 25% sand providing the greatest resistance. Clay content described mixture behavior better than plasticity index, while unconfined compressive strength was useful within soil groups but was not a universal indicator of scour resistance within the tested range.

Five candidate models, including power-law, additive power law and three linear additive formulations, were evaluated for predicting vertical contraction scour in cohesive soils. Among these, the adopted linear additive model provided the best overall performance, with cross-validation confirming good predictive capability. The model incorporated the inundation ratio, approach Froude number, aspect ratio, dimensionless bulk density of the compacted bed, and plasticity index. It achieved an R² of 0.72, an adjusted R² of 0.67, and an RMSE of 0.098 for dimensionless scour depth prediction. The proposed model outperformed the evaluated pressure-flow scour equations developed for non-cohesive sediments, as well as the cohesive soil equations developed for free-surface contraction scour, highlighting the need for a predictive relationship specifically developed for vertical contraction scour in cohesive soils.

Temporal analysis further showed that maximum scour depth approached its ultimate value more rapidly than the scour planform area and eroded volume. When the maximum scour depth reached 90% of its estimated ultimate value, only approximately 77% of the ultimate eroded volume had been removed. This difference indicates that scour-hole expansion continued after the maximum depth had largely stabilized and demonstrates that maximum scour depth alone does not fully characterize the temporal development of the scour hole.

This research provides an integrated experimental and predictive framework for assessing vertical contraction scour in cohesive streams. The findings can support improved scour-evaluation guidance and the design or retrofit of more resilient river-crossing bridges.

Bridge scour threatens transportation infrastructure during floods, particularly when water submerges a bridge deck and forces the flow through the opening beneath the superstructure. The resulting pressure-flow condition produced by this vertical contraction can intensify hydraulic forces and cause scour. Although pressure-flow scour has been investigated in non-cohesive sediments, limited information is available for cohesive streambeds, where erosion resistance depends on soil composition, mineralogy, moisture, compaction, plasticity, and strength. This dissertation investigated the hydraulic, geometric, and soil controls on vertical contraction scour beneath submerged bridge decks and integrated the findings into a predictive framework.

Three coordinated experimental programs were conducted in a laboratory flume under clear-water conditions. Forty unique pressure-flow experiments evaluated the effects of flow intensity, inundation ratio, proximity ratio, girder presence, blockage ratio, aspect ratio, soil type, plasticity, clay content, initial moisture content, degree of compaction, and soil strength. Kaolinite clay, Ball clay, and Kaolinite-sand mixtures were used to represent cohesive-bed compositions and preparation conditions. Scour response was characterized in terms of scour depth, planform area, eroded volume, and temporal development. A consolidated database of 37 experiments was analyzed through dimensional analysis, physical screening, and statistical modeling to develop an equation for maximum scour depth.

The hydraulic experiments showed that cohesive-bed scour was localized and time-dependent, initiating at weak zones and expanding progressively. Increasing Froude number increased scour severity, while greater inundation generally increased maximum depth. Under comparable hydraulic conditions, lower proximity ratios intensified vertical contraction and increased scour. The geometry experiments demonstrated that girder presence had limited influence on maximum depth when total superstructure height remained constant. Blockage ratio was the primary geometric control on scour depth, whereas aspect ratio more strongly influenced scour hole extent and eroded volume. The soil-property experiments showed that increasing compaction reduced scour, although its effectiveness depended on mineralogy and plasticity. Moisture content produced a nonmonotonic response, with minimum scour near the optimum moisture condition. Kaolinite–sand mixtures also exhibited nonlinear behavior, with the mixture containing 75% Kaolinite clay and 25% sand providing the greatest resistance. Clay content described mixture behavior better than plasticity index, while unconfined compressive strength was useful within soil groups but was not a universal indicator of scour resistance within the tested range.

Five candidate models, including power-law, additive power law and three linear additive formulations, were evaluated for predicting vertical contraction scour in cohesive soils. Among these, the adopted linear additive model provided the best overall performance, with cross-validation confirming good predictive capability. The model incorporated the inundation ratio, approach Froude number, aspect ratio, dimensionless bulk density of the compacted bed, and plasticity index. It achieved an R² of 0.72, an adjusted R² of 0.67, and an RMSE of 0.098 for dimensionless scour depth prediction. The proposed model outperformed the evaluated pressure-flow scour equations developed for non-cohesive sediments, as well as the cohesive soil equations developed for free-surface contraction scour, highlighting the need for a predictive relationship specifically developed for vertical contraction scour in cohesive soils.

Temporal analysis further showed that maximum scour depth approached its ultimate value more rapidly than the scour planform area and eroded volume. When the maximum scour depth reached 90% of its estimated ultimate value, only approximately 77% of the ultimate eroded volume had been removed. This difference indicates that scour-hole expansion continued after the maximum depth had largely stabilized and demonstrates that maximum scour depth alone does not fully characterize the temporal development of the scour hole.

This research provides an integrated experimental and predictive framework for assessing vertical contraction scour in cohesive streams. The findings can support improved scour-evaluation guidance and the design or retrofit of more resilient river-crossing bridges.

Keywords

Bridge scour; vertical contraction scour; pressure flow; submerged bridge decks; cohesive soils; hydraulic parameters; superstructure geometry; soil properties; temporal scour evolution; predictive modeling.

Disciplines

Civil Engineering | Geotechnical Engineering | Hydraulic Engineering | Water Resources Engineering

Download

Available for download on Thursday, August 10, 2028

Share

COinS