Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
8 CHAPTER 1. Introduction 1.1 Motivation Bridge scour is the removal of sediment from river and stream beds by water flowing around obstructions such as bridge abutments and piers, and through bridge contracted openings in response to large flood events. Examples of abutment scour, pier scour, and contraction scour are shown in Figure 1-1. Bridge damage or failure due to bridge scour has been a long-standing infrastructure problem because it is accompanied by costs associated with repairing or replacing bridges, economic losses incurred by the necessity to re-route commercial transportation, and even loss of life. According to Landers (1992), about 84% of the bridges in the United States are over waterways. One thousand bridges have collapsed over the last 30 years in the United States and 60% of those failures were associated with hydraulic failure including bridge foundation scour (Shirole and Holt, 1991). Recently, in Georgia, because of the 1994 flooding from Tropical Storm Alberto, over 500 state and locally owned bridges in South Georgia were damaged by scour and were scheduled for repair and/or replacement; total damage was estimated to be $130 million (Arneson et al., 2012). Subsequent severe flooding in the Atlanta, GA metro area occurred in 2009 when eighteen (a) Outflanking of wingwall abutment (b) Undermining of abutment (USGS) (c) Pier scour (USGS) (d) Contraction scour (USGS) Figure 1-1. Photos of different types of bridge scour.
9 stream gages recorded flood magnitudes in excess of the 500-year flood with widespread damage to bridges (Gotvald and McCallum 2010). During the 1993 upper Mississippi River basin flooding, more than 258 million dollars in federal assistance was requested to repair bridges due to flooding-related scour. Damage to bridges accounted for 18 percent of total relief costs, and the primary cause of damage was abutment and approach embankment scour (Parola et al. 1998). According to Macky (1990), in New Zealand, scour caused by rivers results in road expenditures of NZ$36 million per year. The expenditure on scour-related bridge damage amounted to about NZ$18 million (50%) per year, of which 70% was spent on abutments and approaches. Arneson et al. (2012) quoted a 1973 national study for the Federal Highway Administration (FHWA), showing that 25% of failures involved pier damage and 75% involved abutment damage (see Figure 1). Bridge failures can also lead to loss of life such as the I-90 bridge failure over Schoharie Creek near Albany, New York in 1987, the US-51 bridge over the Hatchie River in Tennessee in 1989, and the I-5 bridges over Arroyo Pasajero in California in 1995 (Morris and Pagan-Ortiz, 1999). 1.2 Problem Statement Given the widespread and devastating human and societal costs of bridge scour caused by large floods, intensive research on predicting bridge scour has been completed in the past 25 years or so. Summaries of research progress as well as suggestions for future research on pier scour, abutment and lateral contraction scour, and fluvial geomorphic processes can be found in Sturm et al. (2011), Ettema et al. (2011), and Zevenbergen et al. (2011), respectively. While significant advances have been made since about 1990 in describing and predicting pier scour, much remains to be done concerning abutment and contraction scour and their interaction, scour at solid-wall vs. erodible embankments and abutments, and embankment failure mechanisms. In particular, the interaction of abutment and contraction scour processes occurs because they are driven by similar predictive flow and sediment variables but in a complex interplay of different but simultaneous physical processes arising from turbulence and flow contraction. In addition, the occurrence of more frequent extreme flood events due to climate change introduces the complication of submergence of bridges in orifice flow and even overtopping of bridge decks in weir flow when in fact they were designed for free surface flow based on the historical flood record at the time they were constructed. Orifice flow and weir flow result in vertical contraction (or pressure) scour to further complicate scour-type interactions. Classes of flow that can occur at bridges and affect scour processes are shown in Figure 1-2. Concurrent pier, abutment, lateral contraction, and vertical contraction scour and their interactions greatly complicate the prediction of maximum scour depth for the purpose of designing the bridge foundation and bridge opening width. In addition, most current scour prediction formulas are based on simplified experiments in rectangular flumes that investigate each type of scour in isolation. As a conservative estimate, the engineer is recommended to estimate each scour component separately and then to add them as though they were independent of one another (Arneson et al. 2012). The degree of conservatism and inherent overdesign associated with this approach is unknown and forms the research problem of interest addressed in this report.
10 (a) Free surface or free flow (F) (b) Submerged orifice flow (SO) (c) Deep overtopping flow (OT) (d) Shallow overtopping flow (OT) Figure 1-2. Classes of bridge flow. 1.3 Research Objective In the face of uncertainties and unanswered questions associated with current understanding of how various components of total bridge scour interact during the same design flood event, the principal objective of this research (NCHRP Project 24-37) can be stated succinctly as Determine the relationships among individual scour components observed in the same flow event at a bridge, and determine how to combine them to produce realistic estimates of total scour.
11 In satisfying the objective, bridge foundation designs are expected to become more realistic and economical with the positive result of wider acceptance of and confidence in scour prediction methods by practicing engineers. Inherent in meeting the research objective is a research plan that overcomes some of the weaknesses of past research and implements several tools available in hydraulic engineering research including realistic physical modeling, application to field cases where possible, and introduction of state-of-the art numerical modeling to better understand scour interactions so as to devise a predictive methodology rooted in the physics of the processes involved. 1.4 Research Approach Interactions among pier, abutment, lateral contraction and vertical contraction (pressure) scour that are studied as part of the objective of this research to predict total scour are shown in Figure 1-3. Lateral contraction scour and abutment scour interaction are shown on the left side of the diagram as a continuum between the relative importance of flow acceleration caused by lateral constriction of the flow through the bridge opening and turbulent processes associated with local abutment scour. The processes responsible for total scour on this continuum are interacting such that their relative influences change as the abutment/embankment length increases or decreases, but the maximum depth of total scour is predicted by the same independent variables (Sturm 2006, Ettema et al. 2010). The formulation of abutment and lateral contraction scour acting in concert is extended in this research to include submerged orifice and overtopping weir flow as well as free surface flow (Hong 2013, Hong et al. 2015). Figure 1-3. Bridge scour components and their interactions. The interaction of pier scour and abutment/contraction scour depends on the pier location and on the observation that the maximum pier scour depth at its upstream face does not coincide with the location of the maximum abutment/contraction scour depth near the downstream side of the abutment face under the bridge. As a consequence, the influence of pier scour on maximum abutment/contraction scour is negligible while the influence of abutment/contraction scour on
12 pier scour is determined by the distance of the pier from the abutment. The two-way interaction between vertical contraction scour and abutment/contraction scour is treated as part of the total abutment/contraction scour calculation within the scour hole location but then becomes vertical contraction scour alone at some distance from the abutment in the floodplain if the abutment is set back far enough from the main channel. Pier scour and vertical contraction scour interact with each other outside the zone of influence of abutment/contraction scour. Based on the interactions shown in Figure 1-3, experiments were conducted in the laboratory by varying: ï· Discharge, Q, with flow type changing from free (F) to submerged orifice (SO) to overtopping (OT) as Q increased; ï· Tailwater elevation which increased with Q as in the field; ï· Abutment/embankment length which resulted in long setback abutments (LSA-scour hole in floodplain), short setback abutments (SSA scour hole straddling floodplain and main channel near the bank), and bankline abutments (BLA abutment at main channel bank with scour hole in main channel); o (Note: abutment/embankment length refers to the length of the roadway approach embankment from the edge of water, measured perpendicular to the stream, up to the toe of the abutment to indicate the length of flow obstruction. It will be referred to as the âabutment lengthâ herein for simplicity and in agreement with common practice). ï· Presence and location of piers (twin rectangular column and wall piers); ï· Abutment shape (spill-through and wingwall); ï· Presence or absence of embankment in SO and OT flows; ï· Geometric scale of bridge components and embankment; ï· Clear-water vs. live-bed scour. The experiments were conducted such that the different interactions shown in Figure 1-3 could be isolated and studied to determine relative contributions of various scour types to the total scour. The experimental bridge geometry and bathymetry were based on a common prototype bridge and a typical river geometry that included a compound channel and representative bed roughness. The experiments were conducted as physical model studies with Froude number similarity. The experiments included measurements of water surface profiles and flow fields using acoustic Doppler velocimeters according to established protocols. These measurements were made for a fixed bed to provide characterization of the velocities and turbulence quantities at the beginning of the scour process. In addition, detailed measurements were made of the flow field within an equilibrium scour hole. Numerous field studies have been conducted in the past to improve scour prediction formulas, but the imperative of âflood chasingâ to capture simultaneous measurements of the flow field and the scour footprint in the bridge opening at the peak flood discharge has proved to be a formidable logistics and safety challenge. Alternatively, continuous monitoring of the streambed and flow field in the vicinity of a few selected bridges presents its own challenges of debris
13 damage to the instruments and significant costs of maintaining a long-term intensive monitoring effort. Furthermore, many field sites have unique bathymetry and flow-field characteristics that thwart investigations of interactions of different types of scour and different classes of flow during extreme flood events. Accordingly, a hybrid technique was employed in this research whereby previous physical model studies of existing bridges were used and validated with field data so that the physical model itself then became the laboratory representation of an actual bridge, but with the ability to control the independent variables and study specific scour interactions in isolation under realistic conditions. Numerical modeling, or computational fluid dynamics (CFD), can be applied to compute scour- inducing flow fields to obtain values of velocity, shear stress, and turbulence quantities that are instrumental in causing scour. Turbulence is a key contributing factor to the scour process because it generates not only velocity and pressure fluctuations near the sediment bed but also because it leads to large-scale unsteadiness caused by aperiodic shifting of the horseshoe vortex and shedding of wake vortices which are three-dimensional processes that loosen and entrain sediment grains in pier and abutment scour. Although CFD models cannot as yet simultaneously simulate both the flow fields and the sediment transport processes associated with bridge scour over the long time period (of the order of several days) required to reach equilibrium, they have much information to offer on the influence of turbulence and flow acceleration on the interaction of scour components. The objectives of the numerical modeling component of the research approach were to: (1) expand the experimental study to a wider range of flow conditions in which multiple scour components occur; (2) examine the extent to which scour predictor variables can be obtained from simpler 2D numerical models for some bridge crossings; and (3) elucidate the physics of the flow processes involved in various types of scour to be studied so as to develop realistic predictions for combined scour depths and to establish connections between turbulence structures and scour depths. The numerical modeling added an important dimension to the study that enhanced the results while remaining integrated into the experimental and field aspects of the research plan. 1.5 Embankment Failure Conditions It is possible for scour to continue at an abutment until the embankment itself fails and the abutment stub is exposed. This type of geotechnical failure was considered by Ettema et al. (2010). However, to provide a consistent end point of scour for comparisons of scour depths for various combinations of scour types, this level of failure was not allowed in the present study. Riprap blankets were placed around the embankment and abutment toe, and on the embankment itself, as prescribed in HEC-23 (Lagasse et al. 2009) such that undermining of the embankment or abutment toe, did not occur. It is common practice to incorporate scour protection for bridge abutments in order to prevent failure due to scour, and this practice of designing for minimal scour is a developing international trend. The riprap protection described is not a study of channel stability, which is outside the scope of the present research project.
14 1.6 Organization of the Report In Chapter 2, a brief literature review is presented for the purpose of highlighting some formulas developed for predicting individual types of scour and facilitating their incorporation into methods for predicting maximum combined scour depths where possible. Some field studies that provide specific insights into combined scour processes are also reviewed. Preliminary experiments on combined scour in the laboratory are discussed although they are relatively few in number. This review includes physical model studies of prototype bridges to introduce representative real-world observations that inform the proposed research approach. Finally, the capabilities of CFD models to be used in the research are given for the challenging flow conditions encountered in combined bridge scour including turbulence properties, flow in compound channels, bridge flow contractions and bridge overtopping with a special emphasis on an advanced free surface model component. Chapter 3 encompasses a description of the experimental techniques used to study combined bridge scour including clear-water and live-bed scour at Georgia Tech and the University of Auckland. The experimental method was designed to isolate different types of scour and then methodically add other types of scour to study specific interactions in a manner not possible in the field. Also in this chapter, the application and objectives of CFD models to achieve the research objectives are described. The methods of Large-Eddy Simulation (LES), three- dimensional Reynolds-Averaged Navier-Stokes (3D RANS), and two-dimensional, depth- averaged RANS were incorporated into the overall research plan to complement and inform the experimental results. Results from all phases of the research plan are given in Chapter 4. They are organized to show the findings from investigations of each type of scour interaction as illustrated previously in Figure 1-3. Discussion of the results includes the rationale for proposed methods of combining scour depth predictions for different types of interactions, and comparisons of scour predictions with the proposed methods vs. existing methods. Applicability of results, including limitations and parameter evaluation, are also discussed. Support for the final recommendations, summarized in Chapter 6, is provided by both experimental and numerical model insights.
