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.
38 5.1 Introduction The literature on scour at bridge abutments and similar structures, such as spur dikes, is extensive. Useful overviews of the literature are given by Melville and Coleman (2000) and FHWA (1995) and (1997). This chapter is a concise review of the literature pertaining to scour countermeasures for abutment protection. The review encompasses the fol- lowing scour countermeasure concepts: ⢠Structures for controlling channel alignment, ⢠Guidebanks for approach-flow entry into a bridge water- way, ⢠Structures for controlling channel grade, ⢠Armoring of flow boundaries, and ⢠Local modification of flow around abutments. Channel-bank control serves to ensure that flow remains in a predetermined channel, thereby minimizing flow capacity to erode the boundaries (bed and bank) of the predetermined channel. Channel-grade control structures are used to limit bed erosion so as to impede the upstream progress of chan- nel degradation. For bridges in compound channels (a main channel with flanking floodplains), flow must be guided into a bridge waterway. Guidebanks direct flow into a bridge waterway in a manner that enhances flow alignment, mini- mizes flow turbulence, and thereby reduces scour of the waterway. By far the most common form of scour countermeasure is the armoring of flow boundaries prone to erosion. Armoring substantially increases the capacity of a boundary to resist erosion. Riprap is the customary form of armoring. Other forms of armoring that have been attempted for abutment scour mitigation are tied mats, ensembles of Toskanes, tetrapods, and soil reinforcement. The basic intent of locally modifying the flow field at an abutment is to reduce the scour capacity of the local flow field around an abutment by modifying the flow. The modification hypothetically could be achieved by attaching a form of vane, plate, collar, delta wing, or something similar to the abut- ment. No published studies appear to exist on the use of appurtenances for reducing scour at bridge abutments. The present review also includes the considerations associ- ated with the performance of countermeasure methods subject to the effects of woody debris, ice, scour of cohesive soil, and modeling issues. 5.2 Approach-Channel Alignment 5.2.1 Introduction The most common method for controlling approach- channel alignment to a bridge waterway entails the use of spur dikes or structures similar in function to spur dikes. Spur dikes have been used extensively in all parts of the world as river training structures to enhance navigation, improve flood control, and protect erodible banks (Copeland, 1983). Spur dikes are structures that project from the bank into the channel (Figure 5-1). There are a variety of terms that refer to these transverse structures, including spur dikes, transverse dikes, cross dikes, spurs, wing dams, jetties, groins, and deflec- tors. While there are some differences in the use of these terms, they may be taken to be generally synonymous. Fol- lowing usage of the U.S. Army Corps of Engineers (Franco, 1982; Copeland, 1983), the term spur dikes will be used here. Spur dikes may be permeable, allowing limited passage of water at a reduced velocity, or they may function to com- pletely block the current (impermeable). They may be con- structed out of a variety of materials, including masonry, concrete, earth and stone, steel, timber sheet-piling, gabions, timber fencing, or weighted brushwood fascines. They may be designed to be submerged regularly by the flow or to be sub- merged only by the largest flow events. The main function of spur dikes is to reduce the current adjacent to the streambank, often at the outside (concave C H A P T E R 5 Literature Review
39 bank) eroding bank of meander bends, the placement of which reduces the erosive ability of the flow and may cause deposition near the bank. Because of the deposition induced by spur dikes, spur dikes may protect a streambank more effectively and at less cost than revetments (Lagasse et al., 2001). Spur dikes are usually built in a group of two or more and may be at right angles to the bank, angled upstream, or angled downstream. The crest of the individual dikes might be level or sloping from the bank toward the channel. The crest of each succeeding dike in a system might be at the same elevation as, higher than, or lower than the one upstream, based on the low-water plane (Franco, 1982). There have been many studies on spur dikes for river train- ing, notably Kuhnle et al. (1997, 1998, 1999), Farsirotou et al. (1998), Molinas et al. (1998b), Zhang and Du (1997), Soliman et al. (1997), Tominaga et al. (1997), Wu and Lim (1993), Khan and Chaudhry (1992), Shields et al. (1995a, 1995c), Molls et al. (1995), Mayerle et al. (1995), and Muneta and Shimizu (1994). Richardson and Simons (1984) give design recommendations based on the literature. Lagasse et al. (1995) and Richardson et al. (1991) give design guidelines for impermeable and permeable spur dikes, guidebanks, and riprap stability factor design. Even with the widespread use of spur dikes, there has been no definitive hydraulic design criteria developed. Design guidance is based largely on experience and practice within specific geographical areas, usually on an ad hoc basis. Several hydraulic model studies have investigated the use of spur dikes upstream of abutments at specific sites (e.g., Herbich [1967]). The wide range of variables affecting the perform- ance of the spur dikes complicates the development of gen- eral guidelines for spur dike use. The main site-specific parameters affecting the performance of the spur dikes include channel width, depth, flow velocity, shape of flow hydrographs, sinuosity of the channel, bed material size, distribution and transport rate, and material characteristics of the bank (Copeland, 1983). Parameters that affect the per- formance of spur dikes include: length, width, height, shape, orientation angle, permeability, construction materials, and longitudinal extent of the spur dike field (Melville and Cole- man, 2000). Spur dikes may be classified based on their permeability: high permeability (retarder spur dikes), impermeable (deflec- tor spur dykes), and intermediate permeability (retarder/ deflector). Permeability of a spur dike may be defined as the percentage of the spur dike surface facing the flow that is open. A qualitative guide as to the type of spur dike to use for a specific situation is given in Table 5-1. This table provides preliminary advice on the type of spur dike that may be most suitable for a given circumstance. 5.2.2 Local Scour at Spur Dikes The flow adjacent to a spur dike is characterized by a sys- tem of vortices that is formed as the flow is diverted around the structure (Figure 5-2). Flow velocity is greatest at the edge of the structure, where the protrusion into the channel is greatest. This flow velocity peak and high turbulence causes bed material to be suspended intermittently and transported by the flow. For simple-shaped spur dikes (i.e., flat plates without overtopping flow), the maximum depth of scour occurs at the tip of the structure (Figure 5-3). For more com- plicated spur dike shapes and overtopping flows, the shape of the scour hole may become more complex (Figure 5-4). Pre- dicting the vertical and lateral extent of the local scour is crit- ical for determining the length of bank that the spur dike will protect and determining the depth of spur dike required to protect its base. Stable scour holes associated with spur dikes have been shown to benefit aquatic ecology in degraded streams (Shields et al., 1995a). Currently, there is no established procedure for predicting the maximum scour depths associated with spur dikes. The many complicating parameters of the stream and the spur dike design (see above) are undoubtedly a factor in the lack of established procedure for prediction of scour in the vicinity of spur dikes. Equations to predict the maximum depth of scour have been developed by several researchers. These equa- tions were derived from experiments in laboratory flumes, in which the maximum depth of scour associated with spur dikes was measured. Even with modern equipment, the max- imum scour depth associated with spur dikes in the field is very difficult to determine. Unsteady flows, nonuniform and sometimes varying sediment sizes, and the difficulty of deter- mining the actual location of the stream bottom make collec- tion of field data challenging and rare. Table 5-2 shows the variety of equations used to predict the maximum dept of scour at a spur dike. As is shown in the table, there has been a Figure 5-1. A hardpoint on Goodwin Creek Experi- mental Watershed, Mississippi. Flow direction is from right to left.
40 general lack of agreement on the important variables needed to predict maximum scour depth. This disagreement has been possibly settled by Melville (1992, 1997), in whose equa- tions the ratio of the length of the structure to the flow depth determines the form of the equation. Melvilleâs (1992) equa- tions were technically derived for bridge abutments; however, in many cases, particularly in experimental studies, model bridge abutments are similar to spur dikes. 5.2.3 Design Considerations for Spur Dikes There are three main design considerations for spur dikes: ⢠Length and spacing of spur dikes. The spacing and length dimensions of spur dikes have been related to the length of bank that is protected by each structure. The relationships vary with local variables, such as bank curvature, flow velocity, and whether the structures are designed for navi- gation of bank protection. Recommendations from several sources are given in Table 5-3. ⢠Orientation of spur dikes. There is a lack of agreement as to the most advantageous orientation to construct spur dikes (Figure 5-5). Permeable spur dikes are usually designed to decrease the flow near the bank. They are gen- erally not strongly affected by the angle and are usually built at 90 degrees from the bank to have the maximum effect on near-bank velocity and to use the least material (Lagasse et al., 2001). From studies of single spur dikes (orientation angels of 45, 90, and 135 degrees) in a straight channel, Kuhnle et al. (1998) concluded that the volume of the scour hole was greatest for upstream-facing spur dikes (135 degrees), while the potential for bank erosion was greatest for the downstream-facing spur dikes (45 de- grees). There are proponents of both upstream-facing and downstream-facing spur dikes (Copeland, 1983). As in other design factors of spur dikes, the best orientation is most likely a function of the local conditions and the pur- pose of the structures. ⢠Permeability. Spur dikes with permeability up to about 35 percent do not affect the length of the channel bank pro- tected. For permeability values above 35 percent, the length of bank protected decreases with increased permeability. High-permeability spur dikes generally are most suitable for mild bends where small reductions in flow are sought. Permeable spur dikes may be susceptible to damage from debris and ice (Lagasse et al., 2001). Table 5-1. Spur dike performance chart (from Lagasse et al., 2001).
Spur dikes situated at the bridge crossing oriented 90 de- grees to the main channel are studied for the first time in this study, and the results are presented in Chapter 6. 5.2.4 Application to Abutment Scour Countermeasure Spur dikes are commonly used to maintain predetermined alignment of the upstream-channel approach to a bridge abutment. A bridge abutment may be in danger of being severely eroded when it is subjected to high-velocity flow from a channel that has changed course due to meandering of the channel (Figure 5-6). Spur dikes may also be used to establish and maintain the alignment of a channel. They have been used to decrease the length of the bridge required and reduce the cost and maintenance of the bridge in actively migrating braided channels (Lagasse et al., 2001) 5.3 Vanes Vanes comprise a set of panels placed within a channel and oriented slightly obliquely to the flow through the channel. Figure 5-2. Flow patterns at a spur dike: (top) plan view and (bottom) cross-sectional view (from Copeland, 1983). 41
42 and Kennedy, 1983; Odgaard and Lee, 1984; Odgaard and Mosconi, 1987; Odgaard and Spoljaric, 1986; Odgaard et al., 1988; Odgaard and Wang, 1987, 1990a, 1990b, 1991) and for reducing sediment ingestion in intakes (Barkdoll et al., 1999). The installation of vanes and their effects on bank stabiliza- tion after 2 years are shown in Figure 5-7. The main utility of vanes is to ensure channel alignment toward a bridge opening. Prominent concerns incurred with channel shifting are ⢠Outflanking of an abutment or bridge approach and ⢠Adverse flow orientation to the abutment. The set of vanes acts to direct bed sediment toward a river bank, or the set (if vane orientation is suitably altered) may scour sediment from a region within a channel. The modest amount of work to date suggests that vane use likely is effec- tive primarily for guiding flow to a bridge opening. Vane use, though, is limited to reasonably well-defined, and compara- tively wide, channel conditions whereby the flow indeed will hold its alignment in its approach to a series of vanes, and to channels in which the vanes will not be damaged by material conveyed by channel flow (ice, logs, large bed material, etc.). Vanes have been used for erosion reduction on river bends (Ettema, 1990, 1992; Odgaard and Kennedy, 1982; Odgaard Figure 5-3. Scour hole contours at the end of spur dike experi- ments (from Rajaratnam and Nwachukwu, 1983a). Arrow indicates flow direction.
A factor associated with these concerns is the common propensity of scour at an abutment to attract a thalweg, espe- cially during conditions of diminished or constricted flow (ice, debris, etc.). Scour at an abutment may be minimized if the channel thalweg is kept at a distance from an abutment or pier. In concept, a series of vanes could be used to direct flow optimally through a bridge opening. This use might be explored further. A possible scenario entails vanes placed immediately upstream of a bridge opening in such a manner as to keep flow from impinging directly against an abutment, or to direct flow away from an abutment, and to promote sed- iment accumulation at an abutment or pier. There are only a few cases of vane use to mitigate channel shifting at a bridge site. Odgaard (unpublished report) describes the use of vanes to align the West Fork of the Cedar River at a bridge site in northeast Iowa (Figure 5-8). In both cases, the vanes, set in an array (two wide and about ten long), act to stop a river bend from migrating and severing an approach to a bridge. One study purports to use vanes for mitigating local scour at abutments (Johnson et al., 2001). However, that study actu- ally looked at submerged, angled dikes. The dikes, built from piled rock, are intended to direct flow away from an abutment and possibly to induce some sedimentation at an abutment. In effect, the submerged, angled dikes shift the channel thal- weg away from an abutment. That study did not examine the effects of submerged, angled dikes on flow blockage (with or without debris or ice). The application of such dikes may be fraught with inadvertent effects, such as directing flow adversely toward a pier. 5.4 Guidebanks When embankments span wide floodplains, the flows from high waters must be aligned to go smoothly through the bridge opening. Overbank flows on the floodplain can severely erode the approach embankment and can increase the depth of the scour at the bridge abutment. Guidebanks can be used to redirect the flow from the embankment and to transfer the scour away from the abutment. Guidebanks serve to reduce the separation of flow at the upstream abutment face and maximize the total bridge waterway area, and reduce the abutment scour by lessening the turbulence at the abut- ment face (Lagasse et al., 2001). Guidebanks are earth or rock embankments placed at abutments to improve the flow alignment and move the local scour away from the embankment and bridge abutment. The guidebank provides a smooth transition for flow on the floodplain to the main channel. Design guidelines for guide- banks are given by Bradley (1978), Neill (1973), Ministry of Works and Development (1979), Lagasse et al. (1995), and Central Board of Irrigation and Power (1989). Typically, the length of the guidebank will be longer than the width of the bridge opening. The plan shape is usually elliptical and is designed to provide acceptable flow alignment without flow separation. This requires long radius curves. The important factors for guidebank design are orientation rela- tive to the bridge opening, plan shape, length (upstream and downstream of the abutment), cross-sectional shape, crest elevation, and protection of the structure from scour (Figures 5-9 and 5-10). Protection from scour on the flow-facing side of guidebanks, usually by using riprap stone protection, is critical. 5.5 Grade-Control Structures Channel bed degradation poses a common erosion prob- lem for bridge abutments, as it does for piers. The problem is especially a concern for short bridges over smaller streams farther up within watersheds, because the degradation typi- cally is proportionately more severe with distance up a water- shed. A fairly usual form of bed degradation is head-cutting, which especially occurs in streams with beds having signifi- cant clay content. To stop head-cut migration upstream, various hydraulic structures have been developed, including sheet-pile weirs, con- crete spillways, and rock drop-structures. The methods used to 43 (A) Flow depth = 0.18 m (B) Flow depth = 0.30 m Reprinted with permission from ASCE. Figure 5-4. Topographic map of scour hole experi- ments conducted at National Sedimentation Labora- tory, USDA-ARS (Kuhnle et al., 1999). Contour interval 2 cm. Flow direction indicated by arrows.
44 halt the upstream advance of a knickpoint have been required to change in recent years because of concerns that fish and other aquatic species can be able to move along a stream or river. Though there has been extensive work done on the use of grade-control structures to impede the upstream progres- sion of bed degradation, there has not been much work done regarding the effect of bed degradation on the stabil- ity of bridge abutments. There exists, though, several ad hoc grade-control structures placed in channels so as to protect bridges. Early work to stop knickpoint migration on small streams requiring single-span bridges sometimes entailed the con- Table 5-2. Equations to predict the maximum depth of scour at spur dike. Table 5-3. Spur dike spacing recommendations. Equation Reference Equation # 33.0 âââ â ââ â = f QkyS k varies between 0.8 and 1.8 Inglis (1949) (5-1) 33.02 = bo s F qky k varies between 2.0 and 2.75 Blench (1969) (5-2) 67.0kqys = Ahmad (1953) (5-3) n ns FB B yKy = 2 1 Garde et al. (1961) (5-4) 33.0 4.0 1.1 nsds Fy L yyy += Liu et al. (1961) (5-5) 83.0 2 1 25.0 50375.8= B B y d yys Gill (1972) (5-6) ( ) â¥â¥â¦ ⤠â¥â¥â¦ ⤠â+ ââ = 11175.2 70.1 y yy ry yy y L s s ssd Laursen (1962a) (5-7) ââ â âââ â âââ â âââ â âââ â âââ â âââ â âââ â âââ â âââ â âââ â âââ â Here, B1 = original channel width, B2 = constricted channel width, CD = drag coefficient, d50 = median grain size, Fbo = Blenchâs zero bed factor (which is a function of grain size), Fn = Froude number, f = Lacey silt factor, k = function of approach conditions, K = function of CD (which varies between 2.5 and 5.0), Lsd = effective length of spur dike, n = Manning coefficient, Q = total stream discharge, q = discharge per unit width at constricted section, rs = assumed multiple of scour at dike taken as 11.5 by Laursen, y = average depth in unconstricted section, ys = equilibrium scour depth measured from water surface. Reference Spacing to length ratio Comments Acheson (1968) 3-4 depends on curvature and slope Ahmad (1951) 4.3 5 straight channels curved channels Copeland (1983) 2-3 concave banks Grant (1948) 3 concave banks Neil (1973) 1.5 2.0 2.5 concave banks straight banks convex banks Maza Alvarez (1989) 5.1-6.3 2.5-4 straight channels curved channels Neill (1973) 2 4 if 2 or more dikes if fewer than 2 dikes Richardson et al. (1990) 2-6 depends on flow and dike characteristics Strom (1962) 3-5 Suzuki et al. (1987) < 4 straight channels United Nations (1953) 1 2-2.5 concave banks convex banks
struction of a bridge waterway as a weir. The resulting bridge was called a Greenwood bridge, named for the county engi- neer who developed the concept for this form of bridge waterway. An example is shown in Figure 5-11. Sometimes a simple sheetpile wall was place across the channel. Such weirs now are not well received by environmental biologists because fish migrations upstream are prohibited. Vertical drop structures typically include weirs, check dams, grade-control dams, and stilling basins constructed of materials able to maintain sharp, well-defined crests over which river or stream flow spills. Drop structures constructed of logs and tightly constructed rock can also be used as ver- tical drop structures. Structures constructed of loose rock usually form a sloping sill. Figure 5-12 shows the typical configuration of flow and scour at such structures. A grade- control weir is depicted in Figure 5-13. Such weirs are (or at least were) commonly used to stop the upstream progression of general degradation of a channel bed. One concern with them these days is that they may inhibit fish passage along a stream. The literature on scour at weirs and drop structures is quite extensive. Novak (1955, 1961) conducted early experiments on weir scour. Useful and accessible summaries are given by Laursen and Flick (1983), Peterka (1984), Breusers and Raudkivi (1991), Hoffmans and Verheij (1997), and Raudkivi (1998). The two leading equations for estimating scour depths caused by flow pouring over a vertical drop structure were developed to estimate scour immediately downstream of verti- cal drop structures and sloping sills. Equation 5-8 was devel- oped by Peterka (1984) and is recommended for predicting scour depth immediately downstream of a vertical drop struc- ture and for determining a conservative estimate of scour depth for sloping sills. Equation 5-9 was proposed by Laursen and Flick (1983) and is specifically developed for scour downstream of sloping sills constructed of rock. When designing check dams, weirs, grade controls, and similar structures, it is recom- mended that the designer use these equations as needed (using professional judgment) to estimate expected scour depth immediately downstream of the structure. 45 Figure 5-5. Definition sketch for spur dike angle (from Lagasse et al., 2001). Figure 5-6. Example of spur dike design for realignment of a channel at a bridge crossing (from Lagasse et al., 2001).
46 Figure 5-7. Vanes installed along the concave bank just upstream from the bridge crossing and their effective- ness in stabilizing eroded bank, Wapsipinicon River in Iowa. (5-8) Where: dS local scour depth (below the unscoured bed level) immediately downstream of the vertical drop (m); q discharge per unit width (m3/s/m); Ht total drop in head, measured from the upstream to downstream energy grade line (m), dM tail water depth immediately downstream of scour hole (m), K 1.9, a dimensionless coefficient. The depth of scour calculated in Equation 5-8 is inde- pendent of bed particle diameter. If the bed contains large or resistant materials, it may take years or decades for scour to reach the depth calculated in Equation 5-9. (5-9) Where: dS local scour depth (below unscoured bed level) immediately downstream of vertical drop (m or ft); yc critical depth of flow (m or ft); d50median grain size of material being scoured (m or ft); R50 median grain size of stone that makes up the grade control, weir, or check dam (m or ft); and dM tail water depth immediately downstream of scour hole (m or ft). In recent years, considerable effort has been devoted to developing channel-control structures that do not block fish d y d R y y dS C C C M= ( ) â ( )â¡â£ â¤â¦{ } â4 350 0 2 50 0 1/ /. . d KH q dS t M= â 0 26 0 54. . and aquatic creatures from moving along streams, as illustrated in Figures 5-14 through 5-16. The structures typically have replicated the form and flow features of rock riffles, like small- scale rapids (Figures 5-14 and 5-15), or a weir fitted with a fish ladder (Figure 5-16). The rock rifle drop structure is favored by biologists because it resembles a natural rock riffle and enables fish and aquatic creature migration upstream or downstream. In some instances, grout is applied over the riprap rocks to pre- vent them from moving during extreme flow events. 5.6 Riprap This section discusses procedures already used for sizing and placing riprap at bridge abutments, and it includes an outline of the recommended practice. Riprap, one of the most commonly used materials for ero- sion protection, consists of loose, coarse elements of natural stone. The use of stone to help prevent erosion is certainly not new; for example, a 1914 researcher (Forchheimer, 1914) refers to an equation for choosing stone size, which is based on flow velocity. The increased weight of the riprap stones enables them to resist the increased flow velocities and turbulence associated with flow around an abutment and thereby provides an armor layer protection to the underlying sediments. Interlocking forces between adjacent stones also act to stabilize the riprap layer. Typically, the riprap is placed on the embankment slopes to protect the sediment from scour. In many applications, riprap bank protection has traditionally been placed from the toe of the slope to the top of the bank and has been kept free of vegetation. On large rivers like the Mississippi, however, riprap on the upper part of the bank is often combined with
articulated concrete mattress on the lower portion because of the difficulty and uncertainty of placing riprap underwater in large depths and high velocities. Conversely, riprap can be used on the lower portion of the bank, with vegetation on the upper portion, a technique used on some smaller streams. An alternative to extending the riprap down to the expected scour depth is to lay an equivalent blanket of riprap, known as a launching apron, onto the sides of a developing scour hole. This riprap acts to reduce the scour depth and protect the abutment foundation from undermining. The launchable apron method was apparently first used for large alluvial rivers in India and is described by the Central Board of Irrigation and Power (1989) in India. Advantages of using riprap include ⢠Relative ease of construction, ⢠Flexibility, ⢠Tendency to be self-healing, ⢠Extensive experience and design guidance to support its use, ⢠Ease of repair of local failures, and ⢠Natural appearance that can be enhanced by vegetation. Potential disadvantages of using riprap include ⢠Limited availability and relatively high cost in some areas, ⢠Environmental restrictions on use, ⢠Variations in quality, and ⢠Difficulties of transport and placement in some locations. Fairly numerous guides exist for sizing and placing riprap. Thorne et al. (1995) usefully summarize five general require- ments for riprap sizing and placement: ⢠Riprap must be capable of withstanding the combined impact of all the forces of water flow (and wave attack) responsible for erosion and destabilization. This determination is based on (a) Before installation, 1984 (b) After installation, 1989 Figure 5-8. Plan view of changes in channel alignment before and after installation of vanes, West Fork Cedar River, Butler County, Iowa. 47
48 such factors as stable stone size, lateral and vertical extent of protection, and alignment. ⢠Riprap layout must be safe with regard to geotechnical sta- bility, foundation settlement, and groundwater seepage. ⢠Riprap must be composed of sufficiently durable materials to retain the required erosion resistance and mass stability over the design life of the project. ⢠Ecological impacts and aesthetics of the riprap have to be acceptable. ⢠Riprap layout must be economical to build using available materials, equipment, and labor. 5.6.1 Riprap Failure Mechanisms Riprap is subject to certain failure mechanisms, depending on where it is placed with respect to a bridge abutment. Riprap placed in the apron is subject to failure mechanisms similar to those of riprap placed about a bridge pier,whereas riprap placed on the embankment slopes are subject to not only dislodgement by the flow, but also slump and slide failures where the riprap moves down the embankment slope. Figure 5-11. A concrete weir helps protect bridge abutments against general scour produced by head-cutting. Figure 5-9. Design details for guidebanks at bridge crossing (from Lagasse et al., 2001). Figure 5-10. Design details for guidebanks at bridge crossing (from Melville and Coleman, 2000).
Riprap placed in an apron at the base of wing-wall abut- ments may be subjected to shear failure, edge failure, win- nowing failure, and bed-form undermining (Parola, 1993; Chiew, 1995; Parker et al., 1998; Lauchlan, 1999). Shear fail- ure occurs where the individual riprap stones are not large enough to resist entrainment by the flow. Scour of the riprap stones at the edges of the riprap layer is termed edge failure, while winnowing describes the erosion of the finer bed mate- rial between voids in the riprap layer. Filter layers are often placed to prevent winnowing failure. Shear failure may be triggered at the edges of the riprap layer (i.e., shear and edge failure are often linked and are both a consequence of under- sized riprap). Bed-form undermining occurs under mobile- bed conditions due to the migration of the troughs of large bed forms through the bridge section. The riprap stones set- tle into the troughs of the passing bed forms, which may destabilize the riprap layer. In general, riprap size selection can be based on stability against shear and edge failure, as long as the other possible modes of failure are also addressed appropriately. The following four failure mechanisms of riprap layers at the embankments of spill-through abutments were observed during laboratory studies and in the field (e.g., Blodgett and McConaughy, 1985): ⢠Particle erosion failure. The hydrodynamic forces of the flowing water are able to dislodge individual riprap stones. Possible causes of particle erosion failure include the stone size being too small, the riprap gradation being too uni- form, the side slopes being too steep, and the removal of Figure 5-13. A grade-control weir helps protect a bridge against general degradation of a stream bed. Riprap is placed along the downstream side to pro- tect against scour. Figure 5-14. A riprap stone riffle placed to arrest head-cut pro- gression upstream toward a bridge waterway. The riffle halts erosion, but enables the passage of aquatic creatures. 49 Figure 5-12. Flow and scour over a simple drop structure.
50 individual stones by impact and abrasion. This mode of failure is similar to shear failure, with an additional factor being the effect of the slope. ⢠Translational slide failure. This type of failure occurs when a mass of riprap stones moves down the embank- ment slope, with a horizontal fault line. Failure is usually initiated by undermining of the riprap blanket by a scour hole in the channel. Possible causes include excess pore pressures in the embankment slope, undermining of the riprap toe, and the side slopes being too steep. ⢠Modified slump failure. Modified slump failure is a mass movement of riprap material occurring along an internal slip surface within the riprap layer. Possible causes include disturbance of critical material in the lower levels of the riprap layer and side slopes being too steep. ⢠Slump failure. A movement of material occurs along a ruptured surface that has a concave upward curve, when a shear failure of the underlying base material occurs. Possible causes of slump failure include the presence of nonhomogeneous base material with layers of imper- meable material that can act as fault lines when subjected to excess pore pressure and side slopes being too steep. Abutment side slope is a significant factor in riprap stabil- ity. Accordingly, it is desirable to decrease side slope steepness, thus increasing the stability of the riprap on the slopes. Rec- ommendations by various authors for the minimum value for side slopes vary from 1:2 to 1:1.5 (H:V). Parker et al. (1998) undertook an extensive survey of scour countermeasures installed at bridges throughout the United States. They noted two primary methods of fail- ure for riprap, beside direct entrainment by the flow. These are failures caused by instability of the riverbed and failures caused by an inadequate filter. Instability of the riverbed affects countermeasure stability by altering the flow condi- tions that the countermeasure experiences. 5.6.2 Riprap Stability Many early equations for riprap stability were based on flatbed conditions, several having origins in the research of Isbash (1935, 1936), who was concerned with the stability of rock dumped in flowing water. He proposed the following stability criterion: (5-10) Where Nsc is a dimensionless stability factor for the stone given as (5-11) Where: Vcr critical threshold velocity for the stone, Sr specific gravity of the riprap stones, and d50 effective diameter of the stone. The parameter E has a value of 0.86 for loosely placed stones in flowing water and 1.2 for those that have become embedded. Neill (1976) studied incipient motion of uniform gravel on a flatbed. From dimensional analysis and empirical results, the following stability equation was developed: (5-12)N d ysc = â ââ â â â â 2 5 50 0 20 . . N V g S dsc cr r = â 2 501( ) N Esc = 2 2 Figure 5-15. A rock riffle grade-control structure placed downstream of a bridge. Figure 5-16. A sheetpile grade-control structure placed downstream of a bridge. The structure includes a fish ladder.
51 Where y is flow depth. Neill (1973) provides a graph of correctly sized riprap material, as shown in Figure 5-17, based on the recommen- dations of four U.S. agencies for embankment protection. The local velocity is to be taken as approximately 1.5 times the mean velocity through the waterway opening. The results of Neill (1967) were generalized by Maynord (1987) for graded riprap to produce (5-13) Where d30 denotes the riprap size for which 30 percent by weight are finer. The following equation was proposed by Maynord (1993) for determining riprap stone sizes for use in channels with low turbulence: (5-14) Where: Sf safety factor (>1); Cs stability coefficient for incipient failure: 0.3 (angular rock), 0.375 (rounded rock); Cv vertical velocity distribution coefficient, 1.0 for straight channels on the inside of bends, 1.283 â 0.2 log(Rb/W) for outside of bends, 1.25 for downstream of concrete channels, and 1.25 at end of dykes; Rb centerline radius of curvature of bend, W water surface width at upstream end of bend, CT blanket thickness coefficient, given by Figure 1 in Maynord (1993), y local depth of water, w unit weight of water, s unit weight of stone, V local depth-averaged velocity, K1 side slope correction factor, and g gravitational constant. Maynord et al. (1989) provide a riprap design procedure for application when riprap is placed in channels, whether natural or human-made, that are not adjacent to structures that induce high turbulence levels. The procedure is based on local depth-averaged velocity in a low-turbulence environ- ment. It is applicable to a wide range of gradations for riprap blanket thickness equal to the maximum stone size. The following riprap sizing equation was proposed by Far- raday and Charlton (1983) for the general water environ- ment, with an additional coefficient to account for flow changes in certain situations: (5-15)d y C Fr50 3= * d S C C C y V K gyf s v T w s w 30 1 2 1 2 = â â ââ â â â ⡠⣠â¢â¢ ⤠⦠â¥â¥ / .5 N d ysc = â ââ â â â â 2 63 30 0 20 . . Where: C* coefficient determined from laboratory and field testing and Fr flow Froude number V/(gy)0.5. The mean channel flow velocity (V) should be multiplied by the following factors: 2.0 at noses of groins and guide- banks, 1.5 at bends, and 1.25 in straight reaches. For determining the recommended rock sizes to protect a streambed conveying uniform flow, Brown and Clyde (1989) give (5-16) Where V is the average stream velocity. Pilarczyk (1990) suggested Equation 5-17 in the form of stability criteria for revetments under either wave or current attack. Instead of using the traditional threshold values such as the Shields (1936) criterion, Pilarczyk combined many empirical formulas into the criteria: (5-17) Where: tp thickness of the protection unit, m relative density of protection system, c stability factor for current, KT turbulence and/or shear stress adjustment factor, Kh depth (or velocity profile factor), Kd slope factor, cr critical shear stress parameter, and V depth-averaged velocity. It is suggested that Equation 5-17 can also be used for other armoring devices, such as blocks, block mats, and gabions. Values for the various factors are given in Pilarczyk (1990). An alternative approach to the methods outlined above is to consider the stability of individual riprap stones. This approach is outlined by Stevens and Simons (1971), Stevens et al. (1976), and Simons and Senturk (1977). The method is based on the ratio of the moments of forces resisting overturning of the stone and the moments of forces promoting overturning. A disadvantage of using moment analysis to size riprap stones is that the interaction between the stones in the layer, which can enhance the stability of the stones, cannot be incorporated. 5.6.3 Current Guidelines for Riprap Use at Abutments HEC 23 is the FHWA document that specifies bridge scour and stream instability countermeasures, including rock riprap design, for piers and abutments. In addition, the guide- lines in Table 5-4 customarily are used for sizing and placing m p c T cr h dt K K K V g ⥠â 0 035 2 1 2 . d V y g Sr 50 3 1 2 3 2 3 20 387 1 = â . ( )/ / /
52 Figure 5-17. Riprap size selection (Neill, 1973).
riprap at bridge abutments. The recommendations cover some or all of the following riprap parameters: size, extent of protection, layer thickness, gradation, and filter design. The following sections discuss each of these parameters. Size A list of the riprap sizing equations for abutment protec- tion is provided in Table 5-5. The equations by Simons and Lewis (1971), Croad (1989), Pagan-Ortiz (1991) for spill- through abutments, Austroads (1994), and Atayee et al. (1993) for Fr 0.8 can be arranged into the form (5-18) Where C is a coefficient. For Fr 0.8 and flatbed condi- tions, the Simons and Lewis (1971) relation at the critical loca- tion of failure can be considered identical to that of Atayee et al. (1993) if the local velocity one-rock diameter over the bed is considered as 1.15 times the average contracted flow veloc- ity on the floodplain. For the same flow range and conditions, the Croad (1989) equation can be considered identical to that of Atayee et al. (1993) if the depth-averaged velocity at the crit- ical point of failure is considered to be 1.48 times the average contracted-flow velocity on the floodplain. A graphical comparison of the various equations for Sr 2.65 is shown in Figure 5-18. The equations of Pagan-Ortiz (1991) and Richardson and Davis (1995), which are expressed in terms of flow velocity and depth in the contracted (bridge) section, are plotted for different values of the contraction ratio,, where is the ratio of the channel width to the bridge opening width. It can be seen that the various equations give a wide range of recommended riprap sizes. The equations given by Croad (1989) and Richardson and Davis (1995) give larger riprap sizes in comparison with the other equations, whereas the equations given by Brown and Clyde (1989) and Pagan-Ortiz (1991) give relatively small riprap sizes. Three of these equations are applicable to wing-wall abut- ments, namely Brown and Clyde (1989), Pagan-Ortiz (1991), and Lagasse et al. (2001). The three equations give a wide range d y C S Fr r 50 2 1 = â( ) of suggested riprap size, with the equation by Brown and Clyde giving significantly smaller riprap than the other two equations. The basis of each of the three equations is limited,as discussed below. The equation given by Brown and Clyde (1989) was derived from a simple tractive force consideration for uniform flow.The analysis uses the Manning equation for flow resistance and the Shields entrainment function for stone stability. The equation was calibrated using field observations from a number of sites on U.S.rivers,but the data set did not include bridge sites. The recommended stability factors for riprap placed at bridge abutments were taken as equivalent to those for âhigh- turbulenceâ and âsharp-bendâ sites. The Pagan-Ortiz (1991) equation is based on a simplified laboratory investigation of wing-wall abutments sited on the floodplain. The experiments were undertaken using an idealized fixed-bed channel on which the test riprap stones were placed. The study consistently indi- cated failure of the riprap occurring at the apron toe upstream of the abutment centerline. The equation is of the same form as the classic Isbash (1936) equation. The equation recommended by Lagasse et al. (2001) is also based on the experiments of Pagan-Ortiz (1991) and can be expressed in the same form. Extent of Protection The recommended practice (e.g., Richardson and Davis, 1995; Austroads, 1994) is to extend the riprap right around the abutment and down to the expected scour depth. As mentioned previously, an alternative to extending the riprap down to the expected scour depth is to lay an equiva- lent blanket of riprap, known as a launching apron, on the existing bed. The launching apron protects the side of the scour hole as erosion occurs. Macky (1986) notes that this is seldom practiced in New Zealand and that the riprap rarely extends below the existing riverbed. Specific guidelines for riprap layout for a launching apron are as follows: ⢠The apron at the toe of the abutment slope should extend along the entire length of the abutment toe, around the curved portions of the abutment to the point of tangency Table 5-4. Guidelines for sizing and placing riprap. 53 Country Guidelines United States Richardson et al. (1998) Brown and Clyde (1989) Richardson and Davis (1995) Canada Harris (1988) India Central Board of Irrigation and Power (1989) Australia Austroads (1994) New Zealand Ministry of Works and Development (1979) Gregorius (1985) Melville and Coleman (2000)
54 Table 5-5. Equations for size of riprap. with the plane of the embankment slopes (Richardson and Richardson, 1993a). ⢠The apron should extend from the toe of the abutment into the bridge waterway a distance equal to twice the flow depth in the overbank area near the embankment, but need not exceed 7.6 m (Lagasse et al., 1997). The recommendations shown in Figure 5-19 are based on the studies carried out by Pagan-Ortiz (1991) and Atayee (1993). Atayee (1993) recommended that the width of the apron not exceed 7.5 m. Gregorius (1985) states that âthe apron should have a thickness of 1.25 times the largest stone size and a horizontal length such that, in the launched position, the apron extends to below the estimated scour depth.âNo allowance is made for the fact that the presence of the riprap will reduce the scour depth and also affect the position of the scour hole. Eve (1999) conducted riprap tests with approach-flow conditions at 90 percent of the threshold condition for the approach sandbed. Based on observations of progressive failure of the abutment embankments, Eve developed the following relation for determining the extent of protection: (5-19)B B â La ai â ââ â â â â +( ) â ââ â â â 180 180 W y W r y r d y a a t t 0 5 0 5 1 82 50. . .+ + â ââ â â â = â â ââ â â â Reference Applicability Equation Symbols Simons and Lewis (1971) Spill-through abutments ( ) 50 2 1 4.0 gdS V r r â =η (5-A) d50 = riprap stone size Vr = velocity at a level of one-rock diameter above the bed Sr = specific gravity of rock η = stability factor = 0.595, for flow over a horizontal bed Croad (1989) Spill-through abutments 12 50 025.0 â = slb KVd r sl slK θ Ï 2 2 sin sin1= â (5-B) Vb = velocity at abutment end = 1.5V Ksl = embankment slope factor Ïsl = slope angle θr = angle of repose Brown and Clyde (1989) 5.1 5.15.15.0 3 50 2.1)1( 0127.0 â = fa rsl S SKy Vd (5-C) Sf = stability factor varying from 1.6 to 2.0 for abutment protection y = flow depth Pagan- Ortiz (1991) Wing-wall abutment ( ) 81.023.0 2 2 2 50 1 064.1 â = gS yVd r (5-D) V2 = mean velocity in contracted (bridge) section y2 = flow depth in contracted section Pagan- Ortiz (1991) Spill-through abutment ( )gS Vd r 1 535.0 22 50 â = (5-E) â ââ â ââ â ââ â Austroads (1994) ( ) 250 1 026.1 Fr Sy d r â = (5-F) Fr = flow Froude number = V/(gy)0.5 Fr2 ⤠0.8 ( ) 2 2 2 50 1 Fr S K y d r s â = (5-G) Fr2 = Froude number in the contracted section Ks = shape factor = 0.89 for spill-through abutments = 1.02 for wing-wall abutments Atayee et al. (1993) and Lagasse et al. (2001) Fr2 > 0.8 ( ) 14.0 2 2 50 1 Fr S K y d r s â = (5-H) Ks = 0.61 for spill-through abutments = 0.69 for wing-wall abutments
55 Figure 5-19. Plan view of the recommended extent of rock riprap apron (Lagasse et al., 1997). 0 0.05 0.1 0.15 0.2 0.25 0.3 0.2 0.3 0.4 0.5 0.6 Fr d50/y Pagan-Ortiz (1991), = 1.2β? Pagan-Ortiz (1991), = 1.0β? S =2.65r Richardson & Davis (1995), = 1.2β? Richardson & Davis (1995), = 1.0β? Croad (1989), K = 0.7sl Croad (1989), K = 0.8sl Brown and Clyde (1989), S = 2.0, K = 0.8f sl Brown and Clyde (1989), S = 2.0, K = 0.7f s l Figure 5-18. Comparison of equations for riprap sizing at bridge abutments.
56 Where: y approach-flow depth, B upstream width of the flume, La abutment length, and rt radius of the spill-through abutment toe Wa, , and ai are defined in Figure 5-20. Layer Thickness To a certain extent, riprap layer thickness affects the stabil- ity and durability of riprap protection. Thickness is generally specified as a multiple of maximum size d100 or of median size d50. For relatively low-turbulence applications such as bank protection, the U.S. Army Corps of Engineers (1994) specifies a minimum thickness of d100 or 1.5d50, whichever is greater. For high-turbulence applications, the same reference specifies 1.5d100. Maynord (1988) showed that additional thickness above these minimums generally results in increased stability. It is common practice to use 50 percent more thickness under- water because of uncertainties in placement. For New Zealand conditions, the Ministry of Works and Development (1979) recommends a layer thickness of 2d50, as well as a suitably graded filter layer or filter cloth. Lagasse et al. (2001) recommend that the rock riprap thickness be at least the larger of 1.5 times d50 or d100. To allow for the uncer- tainties associated with placing riprap underwater, it is also recommended that the rock riprap thickness be increased by 50 percent when it is placed underwater. Austroads (1994) gives specific recommendations of riprap size and thickness for specific water velocity values. Table 5-6 shows the riprap diameter D as computed by Austroads from a recommended rock weight, assuming a roughly spherical shape and a specific gravity of 2.65. The table also shows the value of tr/D for the Austroad recommendations and shows an average value of 1.77. The small variances from this value are negligible when considering that riprap placement is inexact in practice. Gradation Riprap gradation is often specified in the form of upper and lower limit curves, with any intermediate gradations being regarded as acceptable. Generally, the narrower the specified limits, the higher the production costs. If the riprap is not correctly graded, the Ministry of Works and Development (1979) recommends the use of a filter for New Zealand conditions. Figure 5-21 shows the grading curve recommended by the Ministry of Works and Development. A criterion for correctly grading riprap for bridge abut- ment protection, given by Brown and Clyde (1989), is shown in Table 5-7. If gradation is sufficient, a filter fabric is not required. It is acknowledged that this gradation may be restrictive, and the 85-percent specification may be ignored if the riprap cannot be sourced to this specification. Austroads (1994) recommends that a filter be required when âthe face stones are nearly uniform in size and embankment material is vulnerable to scour.â No criterion is given to ascertain when riprap is nearly uniform in size, although it is assumed that this is true when riprap grada- tion is more uniform than that stated in the Austroads guidelines. Table 5-6. Design of rock slope protection (Austroads, 1994). Wa Ïai Figure 5-20. Definition diagram for placement of a riprap launching apron at a spill-through abutment (after Eve, 1999). Velocity (m/s) Riprap Diameter, D (m) tr/D Riprap Thickness, tr (m) <2.0 None --- --- 2.0-2.6 0.30 1.67 0.50 2.6-2.9 0.40 1.87 0.75 2.9-3.9 0.55 1.82 1.00 3.9-4.5 0.70 1.79 1.25 4.5-5.1 0.90 1.77 1.60 5.1-5.7 1.15 1.74 2.00 5.7-6.4 1.45 1.72 2.50 >6.4 Special --- ---
Filter Design Filters include granular filters, which make use of the fil- tering effect of graded sediments, and synthetic filters, com- monly called geotextiles. Filters are placed beneath riprap layers to meet one or both of the following objectives: ⢠To prevent groundwater behind the riprap from trans- porting bank material through the riprap (i.e., piping). The filter should be fine enough to prevent the base material from passing through, but more permeable than the sedi- ment being protected. ⢠To prevent large-scale turbulence in front of the riprap layer from sucking bank material through the riprap (i.e., winnowing). Granular filters. According to conventional theory, granular filter material is placed in layers of decreasing size, where the fil- ter material follows the Terzaghi and Peck (1958) filter criterion. Terzaghi and Peck recommended that the 15-percent size be at least four times as large as the coarsest particles of the material being protected, but not more than four times as large as the 85-percent size of the finest soil to be protected by the filter. Advantages of granular filters include ⢠Self-healing ability, ⢠Durability, ⢠Ability to deform without serious damage, and ⢠Relative ease of repair. Potential disadvantages include ⢠The careful control required to achieve specified gradation and thickness, ⢠Difficulty of compaction on steep slopes, and ⢠Difficulties in control of underwater placement. In practice, it is difficult to place a multilayered filter. A sin- gle layer is simple to construct and less likely to contain defects. Riprap produced to exact specifications is usually expensive to source locally. An alternative to granular filters is the use of synthetic filters. De Sousa Pinto (1959) tested the Terzaghi and Peck criterion for applicability to the riprap protection of piers. Favorable results were determined, with no winnowing of the finer material. Synthetic filters. Brown and Clyde (1989) identified the following advantages for the use of synthetic filters (com- pared to granular filters) in riprap revetments: ⢠Lower costs; ⢠Consistent, more reliable material quality; ⢠Faster, more labor-efficient installation; and ⢠Lack of limits on design based on local availability of suit- able granular filter material. Disadvantages of synthetic filters include the following: ⢠Problems with placement underwater; ⢠Unproven durability; ⢠Propensity for clogging; ⢠Bacterial activity, which can affect performance; ⢠Relative movement between fabric and bank material; ⢠Failure on steep slopes, due to sliding (the technique is typ- ically limited to slopes 2H:1V or flatter); ⢠Requirements for edge protection, especially in turbulent flow conditions (sudden failure can result if the scour exposes the filter fabric edge [Escarameia and May, 1992], a phenomenon that is not observed with granular filters). ⢠Susceptibility to damage; ⢠Difficulty of repair; and ⢠The careful design and installation needed to accommo- date settlement. Filter recommendations. The important parameters in selecting a filter fabric are pore size, permeability, and long-term soil/fabric permeability and shear strength, according to Hudson and East (1991). Gregorius (1985) 57 0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100 Percentage passing by weight D ia m et er a s a fra ct io n of D 10 0 Figure 5-21. Optimum riprap grading curve (Ministry of Works and Development, 1979). Table 5-7. Rock riprap gradation (Brown and Clyde, 1989). Stone Size Range Percent of gradation smaller than the stone range 1.5d50 to 1.7d50 100 1.2d50 to 1.4d50 85 1.0d50 to 1.15d50 50 0.4d50 to 0.6d50 15
by Pagan-Ortiz (1991), Macky (1986), Atayee (1993), Croad (1989), and Eve (1999). Pagan-Ortiz (1991) Pagan-Ortiz made laboratory-flume measurements of flow velocities near model abutments and observed the stability of riprap protection placed on the bed surrounding the abut- mentâthat is, in the region where a launching apron would be sited. A vertical wall abutment and a spill-through abut- ment were modeled. The riprap was placed directly on the floor of the flume, rendering the study essentially a fixed-bed investigationâthat is, scour did not occur at the abutment. Therefore, the study is useful in determining the likely posi- tion in which riprap, placed in an apron around an abutment, will first fail while the riverbed remains level. This informa- tion is useful only until a scour hole begins to develop, at which time the flow regime changes, and riprap in other posi- tions of the apron may become unstable. The tests were carried out under clear-water conditions, with V/Vcs values of approximately 0.9, where V is the mean approach-flow velocity and Vcs is the threshold value of V. The spill-through abutment was 116.9 cm (46 inches) wide and 25.4 cm (10 inches) high, as shown in Figure 5-24. The length of the model abutment varied from 63.52 cm to 101.63 cm in increments of 12.70 cm. The experiments were conducted in a 21.34-m long by 1.78-m wide rectangular flume with glass walls. The abut- ment model was placed against the side wall of the flume and surrounded by an observation area consisting of a gravel bed 58 Figure 5-22. End dumping of riprap (Smart, 1990). Figure 5-23. Riprap placement by grab (Smart, 1990). summarized various suggested guidelines for the use of synthetic filters beneath riprap for channel protection as follows: ⢠Piping. To prevent fines from passing through the filter fabric, the average filter fabric pore size, O50, must be less than the d85 size of the bed sediment. ⢠Permeability. The permeability of the filter should be greater than that of the unprotected sediment to prevent hydrostatic pressure buildup in the protected bed. 5.6.4 Ecological Impacts Riprap has been shown to support dense, diverse popula- tions of macroinvertebrates (Shields et al., 1995b). Also, Farabee (1986) found that uniform riprap supports higher fish populations than does graded riprap, presumably because the larger interstitial openings provide better habitat. 5.6.5 Riprap Placement Riprap performance as a scour countermeasure at bridge abutments depends on the accuracy of the placement of the riprap at the site. Riprap is often placed inaccurately because of the inherent difficulties of handling the large riprap stones, especially when they are being placed underwater. There are two main methods of placementâend dumping, where the riprap is tipped off the back of a truck, and individual place- ment by grab, where each riprap stone is positioned individ- ually. Individual placement is more costly, but results in a more effective riprap blanket. These placement methods are illustrated in Figures 5-22 and 5-23. 5.6.6 Prior Laboratory Experiments on Riprap Protection of Abutments Relatively few laboratory studies of abutment scour protec- tion have been conducted. Notable studies are those conducted
placed on the floor of the flume. The gravel-covered observa- tion area spanned the width of the flume and extended for 1.78 m of the length of the flume, with equal areas upstream and downstream from the abutment model. Gravel sizes of 7.6 mm and 10.2 mm were used. Experiments were per- formed to determine the location of the vulnerable zone for initial failure and critical condition for displacement of gravel. Flow conditions were those to simulate 100- and 500- year return period floods, the latter based on FHWA recom- mendations. The significant findings from this study are as follows: ⢠For a spill-through abutment, the initial failure zone begins at the armored floodplain downstream of the contraction near the toe. ⢠For a wing-wall abutment, the initial failure zone occurs at the upstream corner of the abutment. ⢠The rock riprap apron should be extended along the entire length of the abutment, both upstream and downstream, and to the parallel face of the abutment to the flow. ⢠It is reasonable to limit the rock riprap apron to a relatively small portion of the contraction at a bridge crossing because the velocity amplification decays rapidly with dis- tance from the toe of the abutment. ⢠Equations 5-D and 5-E in Table 5-5 were recommended for sizing riprap at vertical-wall and spill-through abutments, respectively. Macky (1986) Macky undertook laboratory experiments under clear- water conditions (V/Vcs 0.9) to compare the effectiveness of different methods of protecting bridge abutments, using an idealized scale model of an actual bridge (Waiharakeke River Bridge). Bridge piers with 20-m spans were included in the model. Alternatives to riprap protection were examined because of the high cost of riprap in some parts of New Zealand. Rock riprap, concrete akmon units (commonly used at coastal sites), flexible concrete mattresses, gabions laid on the abutment slope, stacked gabions staggered up the abut- ment slope, and boulder-filled wire mattresses laid on the bed sediment with stacked gabions above were tested in the study. 59 Figure 5-24. Diagram of spill-through abutment model (Pagan-Ortiz, 1991).
The tests were conducted to simulate typical construction practice rather than recommended construction practice. Thus, the protection measure was terminated at a level slightly below the existing bed levels. This differs from recommended practice, according to which the protection material should be placed to cover the entire face of the scour hole that is expected to develop. The riprap was placed upon an abutment constructed of the bed sediment material. The configuration used to test typ- ical riprap practice is illustrated in Figure 5-25. It is noted that the riprap terminates at the abutment toe without an apron. Macky observed that the flow conditions at the start of the testing period were characterized by very high velocities near the front face of the abutment, especially the upstream corner. Sediment was rapidly eroded from this region, and several rocks were lost from the riprap layer. Over time, the abutment face slumped to a gentler slope that became stable. Despite being undermined, the riprap continued to pro- tect the abutment structure. The protected slope slumped considerably, which enabled a stable, flatter abutment slope to be formed, which was armored by the riprap. A stable scour hole was attained with the abutment substantially undamaged. The principal findings from this study are as follows: ⢠While the downstream side of the abutment requires nom- inal protection only, the upstream corner of the abutment in particular is subject to strong attack and requires pro- tection not only above the existing bed but also down the slope of the developing scour hole. ⢠The high initial velocities in some of the experiments caused damage, which could possibly have been avoided by pre-excavating the scour hole. ⢠The bridge pier adjacent to any abutment needs special pro- tection because it can be sited in the abutment scour hole. Atayee (1993) Atayee studied the stability of a riprap apron using a model spill-through abutment situated on the floodplain of a com- pound channel (see Figure 5-26). The study was intended to build on that of Pagan-Ortiz (1991) by measuring the thresh- old of movement of the gravel material used to protect the floodplain and channel in the vicinity of the abutment. The hydraulic conditions that initiate gravel movement were measured. 60 Figure 5-25. Riprap placement used by Macky (1986). Figure 5-26. Atayee experimental setup (Atayee, 1993).
The abutment was 150 mm wide on an embankment 1.17 m wide, 0.25 m high, varying in length from 130 to 510 mm, with side slopes of 2:1 (H:V). Two sizes of uniformly graded gravels, with d50 7.94 mm and 11.11 mm, were used as model riprap. The riprap apron thickness was equal to two layers of gravel. Atayee (1993) defined failure as occurring at the instant when the unprotected surface (in this case, the bed of the flume) was clearly exposed. Degradation of the gravel layer to expose the flume bed could happen very rapidly (in sec- onds). In all experiments, failure occurred at the toe of the embankment just downstream of the abutment centerline, as shown in Figure 5-27. Croad (1989) As a follow-up study to Macky (1986), Croad investigated the performance of riprap protection in pre-excavated scour holes under clear-water conditions (V/Vcs 0.95). Results were compared with two cases not involving a pre-excavated scour hole: (1) riprap placed down to just below the initial bed level and (2) riprap placed down to the initial bed level with a launching apron. A spill-through abutment, formed from the bed sediment, was used (Figure 5-28). A model bridge foundation was placed in the model and protected using riprap. The flume used for the experiments was 10.6 m long and 2.0 m wide. The tests were run for 24 hours with a flow depth of 75 mm. The experiments used a uniform bed sediment with a mean sediment size of 2.2 mm and riprap material with a mean stone size of 18 mm. Croad does not define his criterion for âDegree of Dam- age.â Photographs included in the report suggest that the degree of damage is related to the amount of sediment removed from the abutment around the model foundation. 61 Figure 5-27. Typical riprap failure zone (Atayee, 1993). Figure 5-28. Bridge abutment and pier model, scale 1:40 (Croad, 1989). Examples of slight, moderate, and severe failure for Croadâs tests are shown in Figure 5-29. Croad concluded that pre-excavation of the scour hole that is expected to develop at an abutment is effective in reducing the amount of damage to the abutment and riprap protection at the abutment. The pre-excavated scour hole needs to be aligned around the upstream corner of the abutment, and the riprap protection needs to extend to the bottom of the pre-excavated hole.
Eve (1999) Eve studied criteria for selection of riprap protection at spill-through bridge abutments with launching apron pro- tection under clear-water and live-bed conditions. The size of the riprap and the extent of the launching apron were varied systematically in the tests. Experiments were run for 24 hours, at the end of which the abutment was assessed for failure. The abutment embankments were constructed using the bed sed- iment material. Three failure conditions were defined: ⢠Total failure, where large-scale movement of sediment and riprap occurred on the abutment slopes. The abutment fill material slumped and large areas of sediment were exposed. ⢠Partial failure, where the movement of riprap and sedi- ment was initiated in one part of the embankment, but did not result in a change of the embankment slope as a whole. Partial failure was typically observed at the water level, where a few riprap stones would be displaced and would move down the slope, and at the base of the slope if under- mining of the toe occurred. ⢠No failure, where no change was observed in the embank- ment slope and the riprap stones did not move. Examples of each failure type are shown in Figure 5-30. For the clear-water tests, two abutment lengths and three riprap sizes were used. Eve measured the position at which the maximum scour depth occurred in all experiments. Gener- ally, the point of maximum scour moved away from the toe of the abutment as the size of the launching apron and riprap stone size increased, as expected. The lateral extent of the riprap launching apron was ini- tially set at twice the flow depth, based on the HEC 18 rec- ommendations (Richardson and Davis, 1995). This criterion was found to be conservative in all cases. For subsequent experiments, the lateral extent (as defined by , , and W in Figure 5-20) was reduced in increments until failure occurred. Eve (1999) proposed Equation 5-19 on the basis of these tests. The live-bed experiments were conducted at 125 and 150 percent of the threshold velocity for the bed sediment. These tests were preliminary in nature, and Eve recom- mended further study under live-bed conditions. In all cases, the riprap failed rapidly, apparently because of win- nowing of bed sediment through voids between the riprap stones. Some of the tests were repeated with the addition of a filter fabric, which was found to improve the stability of the protection, especially at the lower flow velocity. At the higher flow velocity, the abutments failed, in spite of the presence of the geotextile, because of undermining of the 62 (a) Slight failure (b) Moderate failure (c) Severe failure Figure 5-29. Examples of slight, moderate, and severe failure (Croad, 1989).
abutment toe, which led to slumping of the sediment beneath the filter fabric. Three types of failure were observed in the live-bed tests: ⢠Catastrophic rapid failure, which occurred without a geo- textile where the embankment fill material was rapidly winnowed from between the riprap stones, leading to dis- integration of the structure. ⢠Slumping failure, where the abutment failed due to exposure of the underlying geotextile at the abutment toe, allowing the embankment fill material to slump beneath the geotextile. Exposure of the geotextile at the toe of the embankment slope exacerbated the failure process. ⢠Riprap failure, where the riprap layer failed, but the embankment remained intact at the end of the test. 5.7 Cable-Tied Blocks Cable-tied blocks consist of concrete blocks or slabs inter- connected with metal or nonmetal cables. The cables used can be fabricated from steel, copper, or synthetic materials, such as polypropylene (Pzedwojski et al., 1995). An example of cable-tied blocks is given in Figure 5-31. A key feature of cable-tied blocks is the interconnecting of small units, which may be unstable as individual blocks, into a framework capable of withstanding much higher flow velocities. The term is used typically to describe relatively small units. Articulated concrete mattresses, which rely on the same principles, are larger units commonly used for bank protection. Previous studies and experiments on the use of cable-tied blocks for scour protection of bridge foundations are limited and are focused on bridge piers (McCorquodale et al., 1993; Bertoldi and Jones, 1994; Jones et al., 1995; Parker et al., 1998). McCorquodale et al. (1993) conducted an experimen- tal investigation of the use of cable-tied blocks for protection of bridge piers (see Figure 5-32). They studied concrete blocks in the shape of truncated pyramids, interconnected by stainless steel cables. Their clear-water tests indicated that, in 63 (a) Total failure (b) Partial failure (c) No failure Figure 5-30. Examples of clear-water failure criteria defined by Eve (1999). Figure 5-31. Cable-tied blocks used as bank protection (Pzedwojski et al., 1995).
the absence of a filter layer, the underlying sediment could be entrained by the winnowing process. Subsequent tests by Jones et al. (1995) and Parker et al. (1998) confirmed the need for filter layer protection. Escarameia (1995) conducted a series of experiments with cable-tied concrete blocks in a highly turbulent environment. A block mat containing rectangular concrete blocks with ver- tical holes was used for the tests. The blocks were joined using horizontal cables running through two cable ducts in each block. An interlocking effect was achieved as the blocks were joined in a staggered fashion. Results of testing indicated that collapse was more easily reached under rapid flow conditions than under highly turbulent flows and that the amount of tension applied to the block mat did not appear to have a strong effect on the stability of the mattress. Higher stability was achieved when the cable direction was transverse, rather than parallel, to the main flow direction. Escarameia (1995) recommended the following equation for selecting block size for cable-tied blocks with similar geo- metric characteristics to those tested: (5-20) Where: Vb velocity defined at 10 percent of the water depth above the bed; Ss specific gravity of sediment; D C S V gn s b = â 1 1 2 2 ( ) Dn design diameter of the cable-tied blocks; and C 0.05 for TI 0.43 and 1.79 TI â 0.72 for 0.43 TI 0.90; and TI turbulent intensity defined at 10 percent of the water depth above the bed. Choi et al. (2000) investigated the potential to use âG- blocks,â a type of cable-tied block available in Korea, as scour protection at bridge piers. Two types of G-block were tested, as shown in Figure 5-33. The G2 blocks are designed to be tied by U-bolts in one direction and to be self-interlocked in the other direction, whereas the G3 blocks are tied using U-bolts in both directions. Generally, the G-blocks were found to be more stable than the equivalent weight of riprap stone. Plots of critical block weight, WCR, for stability in terms of mean flow velocity, U, are shown in Figure 5-34. 64 Figure 5-33. G-blocks tested by Choi et al. (2000). Reprinted with permission from ASCE. Figure 5-32. Scour protection using cable-tied blocks (McCorquodale et al., 1993).
Parker et al. (1998) identified three possible failure mech- anisms: ⢠Overturning and rolling up of the leading edge, which is exacerbated if the edge is not anchored; ⢠Uplift of the center of the mat, which typically occurs in cases where the edge is not adequately anchored; and ⢠Winnowing of sediment between the mat and the bridge pier, which typically occurs if the mat is not sealed tightly to the pier. Also, Parker et al. (1998) identified six instances of the use of cable-tied blocks as bridge pier protection in the United States. They also noted the significant use of an articulated concrete mattressâusing very large, flat con- crete slabsâas bank protection by the U.S. Army Corps of Engineers. More recently, various cable-tied block mattress configurations have been developed by a number of manu- facturers for erosion protection, including flood-control protection works, bank stabilization, and erosion protec- tion at outfalls. The extensive laboratory investigation by Parker et al. (1998), as part of NCHRP Project 24-7, demonstrated that cable-tied blocks can be designed to be a highly effective countermeasure against scour at piers. Parker et al. proposed the following relation for design of cable-tied block mat- tresses at bridge piers, which indicates that the concrete block units can be smaller than riprap: (5-21) Where: weight per unit area of the mattress, fluid density, cb density of the concrete blocks, and V approach-flow velocity. = â 0 20 2. cb cb V The height of concrete blocks, Hcb, and the volume fraction pore space in the mattress, p, are related to as follows: (5-22) Parker et al. (1998) assessed the use of cable-tied blocks for pier protection to be feasible in sand and gravel-bed streams, but not for rivers with large cobbles or rock, and made the fol- lowing recommendations for their use: ⢠Spacing between cable-tied block units: make spacing ade- quate to allow the mattress a sufficient degree of flexibility. ⢠Cable material: use stainless steel cable for harsh envi- ronments, given the critical nature of cable durability. Where galvanized cable is used, it should be single-strand galvanized. ⢠Mattress size: Make mattress length Lp 3B/cos, mat- tress width 4B/cos, in which Lp is pier length, B is pier width, and is pier skewness to the flow. ⢠Geotextile filter: use it in sand-bed streams, but not in gravel-bed streams. ⢠Geotextile size: make length Lp 2B/cos, width 3B/cos. Additionally, Parker et al. (1998) undertook an extensive field survey of countermeasures at bridge sites throughout the United States. They reported two instances of concern related to U.S. Army Corps of Engineersâ designed cable-tied block installations. Galvanized cables at the I-880 crossing of the Guadalupe River in San Jose, California, had rusted between the inner strands in spite of detailed design specifications. At another site, where cable-tied blocks were installed to prevent erosion downstream from a grade-control structure, geotex- tile matting beneath the cable-tied blocks had pulled away from the edge of the grade-control structure, allowing the blocks to be undermined. Hoe (2001) undertook preliminary tests of the use of cable- tied blocks to protect bridge abutments. The model spill- through abutment was constructed from the bed sediment (0.85-mm uniform sand) using the same mold as that used by Eve (1999).Ceramic tilesâmeasuring 25 mm square and about 5 mm thick and with specific gravity about 2.1âwere used to model the blocks. The tiles were joined together by gluing them onto a flexible, loose-weave net fabric. Tests were conducted under clear-water conditions with and without the use of a geo- textile filter. Figure 5-35 shows before and after photographs of a typical test, which incorporated a filter and was conducted at V/Vcs 0.66 (where Vcs is the critical velocity for bed sediment movement), with flow depth 150 mm. In spite of significant scour development, the spill-through slope remained stable at the end of the experiment (a 24-hour duration). = âcb cbgH p( )1 65 Figure 5-34. Stability of G-blocks tested by Choi et al. (2000).
5.8 Geobags Geobags can be used in lieu of riprap stone or other armor cover, such as cable-tied blocks, that in certain regions can be difficult and expensive to obtain. The potential advantages of the geobags are that they are readily transported (when empty), they can be filled with local sediments and soils (sometimes concrete), and they can be formed to a range of sizesâgeobags can be tailored to fit specific application situ- ations (e.g., Pilarczyk, 2000; and Heibaum, 2002, 2004). Geobags, however, rarely have been used as a scour counter- measure to prevent erosion or scour of the abutments of bridges spanning rivers and streams. Geobags, also called geosynthetic containers, are quite commonly used as an erosion countermeasure in various applications, but have seen limited application for bridge abutments. Small geobags are extensively used during land- development activities to protect exposed soil against erosion. Geobag use is quite common for coastal engineering applica- tions (e.g., Ray, 1977; Pilarczyk, 2000; and Heibaum, 2002, 2004) and as temporary protection against scour at exposed banks and embankments in river channels. Geobags are espe- cially useful for containing and protecting exposed soil dur- ing earthwork construction projects. They are also especially useful for use during repair or renovation work. Figure 5-36 shows geobags used to form and protect a bridge abutment that had experienced scour damage. Geobags are sized in accordance with a design method pro- posed by Pilarczyk (2000). The method estimates a geobagâs thickness, DB. The aerial extent of a geobag should exceed DB and otherwise can be sized for handling ease or to fit a site. The general form of Pilarczykâs relationship for geobag thick- ness is as follows: (5-23) Where: SSB specific gravity of the geobag, V depth-averaged mean velocity, g gravity acceleration, st stability parameter, C critical value of the Shields parameter for particle (geobag) entrainment, KT turbulence factor, Kh depth parameter, and Ksl slope parameter in which (5-24)ksl b C = â â¡ â£â¢ ⤠â¦â¥1 2 2 sin sin D S K K K V gB SB st C T h sl = â( ) 0 035 1 2 2 . 66 (a) Prior to testing, looking from upstream (b) After testing, looking from upstream Figure 5-35. Laboratory test of spill-through abut- ment protection using cable-tied blocks, Hoe (2001). Figure 5-36. Geobags form and protect a bridge abutment that had recently experienced damage owing to abutment scour.
Where: b angle of the boundary on which the geobag is placed and C angle of repose of the sediment forming the boundary. For the experiments, and were 26.7 and 30 degrees, respectively. Pilarczyk (2000), who gives the background to Equation 5-24, recommends for geobags that , C, and KT be 0.75, 0.05, and 2.0, respectively. The depth parameter Kh is defined as a function of water depth, y, and equivalent rough- ness ks. Pilarczyk (2000) suggests using ksDn. However, since Dn is unknown initially, the measured thickness of the geobag sample was used as a trial value. The required thickness of the geobags, Dn, was calculated as 22 mm, using a bulk-specific gravity of the model geobags measured to be 1.46. Several investigators studied the stability of geobags as a slope-protection unit in coastal applications (Ray, 1977; Jacobs and Kobayashi, 1983; Gadd, 1988; Pilarczyk, 1998). Pilarczyk (1990) provided an empirical equation for stability of revetment material under flow attack. His formula can be used for different materialsâsuch as riprap, geobags, geomattresses, gabions, and block or block matsâusing different coefficients provided for each material. Placing geobags, geosynthetic bags filled with sand, as an abutment countermeasure has an important advantage com- pared with riprap. A geobag is less prone to winnowing of the fine underlying bed sediment particles. To overcome win- nowing failure of riprap, geosynthetic or granular filters are used in combination with riprap, commonly when riprap is used as a scour countermeasure. However, laying filters is difficult to control in practice, especially in flowing water. Geobags do not require filters and, therefore, are relatively easy to place. Moreover, the size and weight of individual bags are solely the discretion of the designer, and one can always design large enough geobags so that they can resists shear ero- sion. Figure 5-37 shows several arrangements of geobags. As is discussed above, there are distinct advantages of using geobags as an abutment countermeasure over its riprap coun- terpart. However, before one can use geobags, a confirmation of their performance and suitable deployment is needed. 5.9 Other Forms of Armoring Besides riprap and cable-tied blocks, various other forms of armoring have been used to protect bridge abutments against scour. Frequently, the other forms of armoring have entailed the use of large elements, notably Toskanes, dolos, and large blocks of concrete or rock. Figure 5-38 depicts the use of large concrete blocks (locally termed hedgehogs), linked together by cables to armor the channel bank imme- diately upstream of an abutment. The channel bed in front of the abutment is armored with large, hinged concrete slabs. Burns et al. (1996) developed Toskanes as an alternative scour countermeasure where riprap is not feasible. Results of model studies and design guidelines are presented. Ruff et al. (1995) used Toskanes to protect bridge piers. Adams et al. (1999) used reinforced soil by bridge abut- ments, but they concluded that reinforced soil is not suitable for permanent bridges in scour zones. 67 Figure 5-38. Heavy hinged slabs and cable-tied blocks (termed hedgehogs) used to protect a bridge abutment on an ephemeral river. Figure 5-37. Geosynthetic containers: bag (top), and mesh with plant openings (bottom) (Heibaum, 2002).
