Countermeasures to Protect Bridge Abutments from Scour (2007)

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.

161 This chapter reports laboratory results of three flow modifi- cation countermeasures: parallel walls, spur dikes, and abut- ment collars. Section 9.1 describes the laboratory equipment and procedure, including a description of the flume, the abut- ment model, the velocity ration, the sediment characteristics, instrumentation, and the experimental procedure employed. Section 9.2 describes the baseline experiment results—that is, the scour depth at the bridge abutment without any counter- measures.Sections 9.3,9.4,and 9.5 discuss the results of the tests using parallel walls, spur dikes, and abutment collars, respec- tively. Section 9.6 summarizes the findings. 9.1 Experimental Apparatus and Procedure 9.1.1 Flume All of the experiments were conducted in a flume located in the hydraulic laboratory at the USDA-ARS National Sedi- mentation Laboratory in, Oxford, Mississippi. The flume channel was 30 m long, 1.2 m wide, and 0.6 m deep. It was supported in the center at two points and on the ends by four screw jacks that allowed the channel slope to be adjusted. The wing-wall abutment model was located over a 3-m long, 1.2-m wide, and 1.2-m deep recessed section of the flume 22 m downstream of the inlet tank. The test section was 22 m downstream from the inlet, and the channel was 1.2 m wide, thereby making the test section a distance downstream from the inlet of 18 times the channel width. This distance was enough to ensure fully developed flow at the test section. Uni- form flow was established for each experimental run by the adjustment of the flume slopes and pump speed until the water surface line, the bed surface, and the flume slope were parallel to one another along a 12-m transect in the approach channel. The channel plan and section views of the experi- ments are illustrated in Figure 9-1. Figure 9-2 shows all of the elements in the flume, such as the abutment model, flood- plain, main channel, instrument carriage, and flume inlet and outlet. It also shows scour by the abutment. All experiments used a compound channel, consisting of a 320-mm wide, asymmetric floodplain next to a main channel with a bank slope of 1:1. The elevation difference between the top of the floodplain and the main channel bed was 80 mm. The rigid floodplain was made of a galvanized steel plate and glued down onto the flume bottom. A layer of sand was glued onto the floodplain to add roughness. In addition, since most floodplains are heavily vegetated and therefore have high roughness, gravel with a mean diameter of 4.5 cm was placed in a staggered arrangement on the floodplain in later runs of baseline cases. Figure 9-3 shows this arrangement. By meas- uring the velocity profiles both in the main channel and on the floodplain, and by using Manning’s equation, the rough- ness of the floodplain and the main channel bed under clear- water conditions (a velocity ratio of 0.9) were found to be 0.030 and 0.014, respectively. 9.1.2 Abutment Model The wing-wall abutment model was made of sheet steel. The dimensions of the model are shown in Figure 9-4. The abutment terminated on the bank slope of the main channel, as illustrated in Figure 9-1, which corresponds to the Type III abutment of Melville (1992). The distance between the top of the floodplain and the top of the abutment was 60 mm. The abutment length was about one-third of the channel width and was observed not to alter the flow enough to interact with the far flume wall. 9.1.3 Sediment Characteristics The bed material sediment used in the main channel had a diameter of 0.8 mm. The standard deviation of the sediment diameter, [g  (D84/D16)1/2], was equal to 1.37. According to a modified version of the Shields diagram (Miller et al., 1977), C H A P T E R 9 Lab Results IV: Flow Modification

the critical shear velocity of the bed sediment is 1.995 cm/s. In experiments under live-bed conditions, the sediment was recirculated with the water. At the upstream inlet of the flume, a gradual-transition contraction was built to guide the sediment into the main channel. 9.1.4 V/Vc Ratio For all clear-water scour experimental conditions herein, a V/Vc ratio of 0.9 was used, with V being the overall average velocity in the whole cross section of the compound channel and Vc being the critical velocity of the sediment. Similitude between laboratory experiments and field scale was satisfied by the use of the aforementioned u*/u*c ratio, of which a value of just below 1.0 represents a condition called “clear-water scour.” This condition is extreme for scouring because the velocity is as high as possible without the movement of the channel bed, which causes infilling of the sediment hole. The mean velocity of the flow is given by the following equation: (9-1) Where: ks roughness height of the bed and Yo  distance above the bed. At the threshold condition, (9-2) So for clear-water conditions, where the bed is stable and ks is constant, (9-3) V V u uc c = * * V u Y k c c o s* . log .= ⎛ ⎝⎜ ⎞ ⎠⎟ +5 75 6 0 V u Y k o s* . log .= ⎛ ⎝⎜ ⎞ ⎠⎟ +5 75 6 0 162 L320Floodplain Bank 80 Main Channel Flow Bm=800 A1 A A-A1 Sediment yf ym Abutment Wa=720 La=440 Figure 9-1. Dimensions of experimental compound channel (mm).

For the experiments of this research project, u*c  0.01995 m/s. Thus, given that clear-water scour is If the flow depth is set, then (9-4) Thus, the slope of the flow should be able to be determined. V and Vc can also be determined by selecting ks  2d50 1.6 mm. u gRS * = u * .= 0 017955 m/s V V u uc c = = * * . ,0 9 then Three flows were used with velocity ratios, V/Vc, of 0.9, 1.5, and 2.3. The critical velocity of the bed material was calcu- lated using the velocity distribution relation for a rectangular cross section, rough wall, and free surface, as shown in Equa- tion 9-2 above. For clear-water conditions (V/Vc  0.9), the experiments were run for 80 hours so that the local scour had reached a near equilibrium value. For live-bed conditions (V/Vc  1.5 and 2.3), all experiments were run for 50 hours to ensure that at least 125 bed forms migrated past the abutment. 9.1.5 Instrumentation Velocity profiles were collected 15 m downstream of the inlet tank, with a 2-mm outside diameter total head tube mounted on a point gage at the channel centerline. Flow rate in the flume was measured using a pressure transducer con- nected to a Venturi meter in the return pipe. Flow depth was controlled by the volume of water in the flume and measured by taking the difference in elevation between the bed and water surface over a 12-m long transect in the approach sec- tion. Water surface and bed surface profiles were collected using two acoustic distance measurement devices, the remote measurement unit (RMU) that operates in air and the bed form and sediment information system (BASIS) that operates underwater. These instruments were mounted on an instrument carriage that traveled on rails over the chan- nel. The instrument carriage was equipped with a computer- controlled, three-axis precision positioning system that allowed transects of the scour hole to be automatically col- lected using the BASIS. The bed elevation of the area in the vicinity of the abutment was measured at the completion of the clear-water experiments using the BASIS. For the live-bed experiments, the bed elevation of a 2.5-m long flow- parallel transect from 13 to 50 mm from the abutment (depending on the size of the bed forms) was measured con- tinuously for 125 minutes after the scour reached equilib- rium state. The probe of the BASIS detects the bed elevation once every minute at a certain point along this transect. The distance between two successive points detected along the transect by the probe is about 1.5 cm. The time-averaged and instantaneous scour depth values adjacent to the abutment were determined from this record. Flow depth was also measured and checked using the point gage. 9.1.6 Experimental Procedure For this experiment, the researchers took the following steps: 1. Placed the abutment model (with or without counter- measure models) in the flume; 2. Leveled the sediment bed surface; 163 Figure 9-2. View of the flume looking upstream. Figure 9-3. Staggered placement of gravel on floodplain to provide roughness (cm).

3. Wet and drained the flume completely; 4. Collected the profile of the bed surface using the RMU; 5. Filled the flume with water and obtained the desired depth; 6. Collected an initial set of transects of the scour region around the abutment using the BASIS program; 7. Set the predetermined flume slope and started the pump; 8. Adjusted the pump speed to obtain uniform flow at the selected flow depth by measuring the water surface ele- vation at both ends of the 12-m transect in the center of the channel; 9. Checked the water surface slope along the 12-m transect using the RMU device to ensure the uniformity of the flow; 10. Maintained the same rate of flow and approach depth for the entire experimental run; 11. Collected transects of the scour region at 30-minute intervals; 12. Increased the time intervals to 60 to 90 minutes or more as the experiment progressed and as changes in the scour region became slower; 13. Continued the experiment until the changes in the scour hole became very slow (approximately 80 hours); 14. Stopped the pump, dewatered the flume carefully, and contoured the scour hole; and 15. Took a photo of the scour hole. 9.2 Baseline Experiment In order to study the efficiency of a certain type of counter- measure in preventing scour at the bridge abutment, the base- line scour—that is, the scour depth at the bridge abutment without any countermeasures—was determined as a reference. 9.2.1 Clear-Water Scour Baseline Experiments Experimental Results Several runs of experiments were carried out under clear- water scour conditions with a velocity ratio of 0.9 to deter- mine the worst scour scenario at the bridge abutment. These experiments were done with varying flow depths both in the floodplain and in the main channel. The floodplain was first roughened only with sand of the same size as the bed mate- rial and later was further roughened with staggered gravel, as mentioned previously. The gravel was used to simulate a rough floodplain and had a mean diameter of 4.5 cm. The placement of the gravel is shown in Figure 9-3. Table 9-1 gives the experimental results of these baseline tests. Scour Pattern Two scour patterns were discovered, depending upon the difference of the roughness on the floodplain. Without gravel on the floodplain, the scour pattern of Figure 9-5 occurred, in which there were five scour locations. Tests B1 through B4 (not roughened with gravel) had similar scour patterns. The first scour hole, Zone A, was at the upstream corner of the abutment and posed the greatest threat to the stability of the abutment. Scour in Zone B was located some distance away from the abutment face in the bridge crossing and was where the maximum scour hole was located. Since it was away from the abutment, it was considered not to be a threat to the abutment. Scour in Zone C was a short distance downstream 164 Figure 9-4. Dimensions of abutment model (mm).

of the abutment. This scour zone may pose a threat to the main channel bank immediately downstream of the abut- ment. Scour in Zone D was far out into the main channel and, therefore, posed no threat to the abutment. Scour in Zone E was located at a short distance upstream of the abutment cor- ner and seemed to be the upstream extent of Scour Zone B. For Test B5, which did have gravel on the floodplain to pro- vide roughness, there was a slightly different scour pattern, as shown in Figure 9-6, where scour Zones A, B, and E merged to be the maximum scour location, which was located at the upstream corner of the abutment. Scour Mechanism In the baseline experiments, it was found that each of the scour zones identified in Tests B1 through B5 was formed by different mechanisms. Figure 9-7 shows the flow patterns responsible for the various scour patterns observed. In Zone A, scour was caused by a combination of a downward roller from the water striking the upstream abutment corner, return flow from the floodplain flowing down toward the main channel bed, and vortex shedding from the upstream abut- ment corner. In Zone B, scour was caused by a secondary vor- tex oriented horizontally and parallel to the streamwise flow direction. Scour in Zone C was caused by the wake vortex induced by flow separation. Scour in Zone D was caused by an increase in main channel velocity above the critical value for sediment movement caused by abutment-induced con- traction scour. Scour in Zone E was simply the initial part of scour in Zone B and was caused by the flow coming down from the floodplain into the main channel. Effect of Floodplain Roughness on Scour Depth at Upstream Corner of the Abutment The mechanisms of the formation of the four scour holes in Tests B1 through B4 were explained above. In addition to these flow patterns, it was seen that the maximum scour depth was found in the bridge crossing a distance away from the abutment instead of being at the upstream corner 165 Experimental Result Test B1 Test B2 Test B3 Test B4 Test B5 Run time, te (min) 4,800 2,920 4,800 4,800 4,800 Total discharge, Q (m3/s) 0.0366 0.0335 0.0387 0.0353 0.0387 Flow depth on floodplain, yf (cm) 4.5 1.2 5.2 3.0 5.2 Flow depth in main channel, ym (cm) 13.2 9.9 13.2 11.0 13.2 Main channel bank height, h1 (cm) 8.7 8.7 8.0 8.0 8.0 Scour depth at upstream corner of abutment, dmax,up,abut (cm) 3.60 5.00 5.30 4.70 7.77 Scour depth at a short distance downstream of the downstream corner of the abutment, dmax,dn,abut (cm) 3.83 -- 3.91 3.44 2.90 Scour depth in the channel away from the abutment, dmax,ch (cm) 4.00 -- 7.00 4.00 1.50 Floodplain roughness Sand (0.8 mm) Sand (0.8 mm) Sand (0.8 mm) Sand (0.8 mm) Sand (0.8 mm) plus staggered gravel (Figure 9-3) Table 9-1. Baseline clear-water experimental results with V/Vc  0.9. Figure 9-5. String contour of baseline Test B3 (flow is from left to right). Figure 9-6. String contour of B5 with gravel on the floodplain (flow is from left to right).

of the abutment. The reason why the maximum scour depth was not located right at the upstream corner of the abutment was that the velocity ratio between the floodplain flow and the main channel flow was so high that the flood- plain flow was able to jet into the main channel a distance away from the bank. To solve this problem with the hope of the maximum scour depth taking place right at the upstream corner of the abutment, the floodplain was further rough- ened with gravel of an average diameter of 45 mm, as shown in Figure 9-3. As was expected, and as is shown in Figure 9-6, the scour in Zones A, B, and E merged and the maximum scour hole was found right at the upstream corner of the abutment. Effect of Main Channel Height on Scour In addition to the velocity ratio of floodplain and main channel, another factor in the location of scour away from the bank is the bank height. When the bank height approaches zero, the scour pattern turns into Type I scour (Melville 1995), which is abutment scour in a rectangular channel. In this case, the maximum scour will happen around the upstream corner of the abutment because the approach flow obstructed by the protrusion of the abutment always makes contact with the bed around the abutment corner upon entering the bridge crossing. Therefore, the downflow and secondary vortex will exert significant shear stress on the bed and cause scour. When the bank height increases, the flow coming from the floodplain must travel a distance before it hits the bed in the main channel after it enters the bridge crossing. In this process, the flow may avoid contacting the corner of the abutment. Figure 9-5 showed that when the velocity ratio between the floodplain and the main channel flow was relatively high, the floodplain flow came off the floodplain edge at a distance upstream of the upstream cor- ner of the abutment (Point E) and made full contact with the main channel bed at Point B, where the maximum scour hole was found. However, Figure 9-6 (slower floodplain velocity) showed that the flow contacted the main channel bed at the upstream corner of the abutment. Formation of Scour Holes Clear-water data of all runs indicated that the upstream scour holes (A and B in Figure 9-5 or EA(B) in Figure 9-6) developed faster at the beginning of the experiment than the downstream scour holes (C in Figures 9-5 and 9-6) because the vortex systems at the upstream corner of the abutment were generally stronger than those at the downstream end of the abutment. Therefore, the upstream scour hole reaches equilibrium more quickly than the downstream scour hole. Figure 9-8 shows the time evolution of the scour depths of both the upstream and downstream scour holes with time for the case with a smooth floodplain (Test B1) and a rough floodplain (Test B5). 9.2.2 Live-Bed Scour Baseline Experiments When the velocity of the flow gets higher and the velocity ratio or the shear stress ratio of the flow exceeds one, then the bed materials of the river begin to move and bed forms are initiated. Under live-bed conditions, it is believed that the fluctuating bed forms and the higher shear stress may pose more threat to the stability and practicability of the counter- measures, even though bed-load sediment may fill in the scour holes. Therefore, live-bed experiments were conducted to test successful countermeasures for clear-water conditions to confirm their efficiency in all scour conditions. The live- bed baseline—that is, the scour depth at the bridge abutment without any countermeasures—was determined first as a ref- erence. The live-bed baseline scour results were obtained under the flow conditions shown in Table 9-2. In the live-bed condition, because of the wavy water sur- face and the fast-moving and fluctuating bed forms, velocity 166 Figure 9-7. Flow patterns around a wing-wall abutment (flow is from left to right).

profile measurements turned out to be difficult. Therefore, the flow was mainly controlled by the discharge and the aver- age water surface profile and the average bed profile along a 12-m transect in the middle of the approach channel, where each bed profile was monitored. First, the discharge was set to be 1.5 times the discharge at the critical condition, and then the slope was adjusted until the average water surface slope and average bed surface slope were equal to the flume slope (i.e., the uniform live-bed flow condition was set through trial and error). The flow depth was also adjusted to be 132 mm deep because it was in the clear-water case. Similitude between laboratory experiments and field scale was satisfied by the use of the aforementioned V/Vc ratio, of which a value of above 1.0 represents a condition called “live- bed scour.” This condition is extreme for scouring of objects in the flow, such as rock, that make up the flow-altering coun- termeasures described in this chapter because the high veloc- ity can cause dislodging of individual rocks and therefore constitute failure. Also, because of the fast change of the bed profiles at the bridge crossing, it was impractical to measure the bed profile in the same manner as was done in the clear-water scour con- dition, which took 14 transects and more than 13 minutes to cover the whole scour region. Therefore, in the live-bed case, only one 2.5-m long transect at the main channel side of the abutment was chosen to monitor the time evolution of the scour at the edge of the abutment. This transect started from a point 1.5 m upstream of the upstream abutment tip and traveled parallel to the flow just to the right of the abutment. Each pass of the transect took 54.90 seconds, and a total of 133 loops were taken to capture the local scour, as well as the bed forms along this transect, in a period of 2 hours. The data obtained from this transect enabled the determination of the maximum local scour location at the abutment and how it evolved with time. The bed form shape across the channel was not uniform. Close to the opposite wall, where there was no floodplain, the bed form amplitude was relatively small; close to the flood- plain, the bed form amplitude was relatively high. Figure 9-9 shows the time-averaged local scour depth along the 2.5-m transect at the abutment versus the distance starting from a point 1.5 m upstream of the upstream abutment corner for the 1.5 velocity ratio case. Figure 9-10 shows the time evo- lution of baseline scour at the upstream abutment corner ver- sus the time after the scour reached equilibrium. 167 upstream scour hole Time t (min) Time t (min) Sc ou r d ep th d s (cm ) Sc ou r d ep th d s (cm ) downstream scour hole upstream scour hole downstream scour hole Figure 9-8. Plot of the time evolution of the scour depth of both the upstream (A) and downstream (C) scour holes at the abutment. Table 9-2. Experimental results for baseline scour depth for three velocity ratios, V/Vc. Experimental Result V/Vc = 0.9 V/Vc = 1.5 V/Vc = 2.3 Instantaneous maximum scour at abutment, dmax,abut,inst (mm) 77.7 150.0 172.8 Time-averaged baseline scour depth, dabut,avg (m) 77.7 72.3 75.2 Run time, t (hr) 80 50 50 Total discharge, Q (m3/s) 0.0387 0.0622 0.0966 Flow depth in flood plain, yf , is 52 mm. Flow depth in main channel, ym, is 132 mm.

From Figure 9-9, it can be seen that the maximum time- averaged local scour still took place at the upstream corner of the abutment, with a scour depth of 7.23 cm. This finding agreed with the baseline scour pattern in the clear-water scour condition. The scour depth dropped to 4.01 cm as it approached the downstream end of the abutment. Deposi- tion began to occur at a point 45 cm downstream of the downstream abutment end. From Figure 9-10, it can be seen that the scour depth at the upstream abutment corner varied from near 0 to 14 cm with time. However, the mean value was 7.23 cm, and the ampli- tude of variation was about 7 cm. The fluctuation was mainly due to the bed forms. When the crest of the bed forms passed, the scour depth reached its minimum value, which is a few millimeters above zero; when the trough of the bed forms came, the scour depth reached its maximum value. 9.2.3 Conclusions The following findings were made from the baseline exper- iments in a compound channel with various flow depths on the floodplain and main channel, various floodplain rough- ness values, and various bank heights. Under clear-water conditions: • Five zones of scour were found for equal floodplain and main channel roughness values and relatively high velocity ratio between the floodplain and the main channel veloci- ties. The floodplain flow tended to shoot into the main chan- nel at a distance away from the upstream corner of the abutment instead of being fully located at the abutment cor- ner. Under this condition, the maximum scour in the whole region was normally found in Zone B (see Figure 9-5). • The closer the floodplain flow was to the abutment, the deeper the scour hole was. By increasing the roughness on the floodplain, and thereby decreasing the velocity ratio between the floodplain and the main channel, the flood- plain flow was located closer to the upstream abutment corner. As a consequence, the scour Zones A, B, and E were combined and the maximum scour depth was found at the upstream corner of the abutment. • The principal and secondary vortex systems at the upstream corner of the abutment were stronger than the wake vortex systems at the downstream corner of the abut- ment. Consequently, the upstream scour hole reached equilibrium more quickly than the downstream scour hole. • Test B5 was used as the clear-water baseline condition for subsequent countermeasure experiments with a scour depth of 77.7 cm. Under live-bed conditions: • Maximum scour took place at the upstream corner of the abutment. Time-averaged scour depth at the upstream cor- ner was less than the scour depth under critical clear-water conditions, whereas instantaneous scour depths were 168 0 50 100 150 200 250 Distance D (cm) -2 0 2 4 6 8 Ti m e a v e ra ge d lo c al s c o u r de pt h d s (cm ) abutment Flow Figure 9-9. Time-averaged local scour depth along the 2.5-m transect at the abutment versus distance starting from a point 1.5 m upstream of the upstream abutment corner. Figure 9-10. Time evolution of baseline scour at the upstream abutment corner after equilibrium was reached.

between values near zero and values nearly twice the max- imum scour under clear-water flows because of the super- position of the trough of bed forms (see Figure 9-10). • Live-bed scour reaches equilibrium more quickly than clear-water scour. Results in Table 9-2 will be used as the references to evalu- ate the efficiency of countermeasures that will be tested later. 9.3 Parallel-Wall Countermeasure The first countermeasure reported herein is the parallel wall. It consists of a wall parallel to the flow attached to the upstream corner of the abutment. After a brief introduction, this section describes the flow patterns and then details solid flat wall countermeasures and rock wall countermeasures. 9.3.1 Introduction Scour at an abutment can cause damage or failure of bridges and result in excessive repairs, loss of accessibility, or even death. Scour mitigation at bridges has received much attention in the past few decades. Hydraulic countermeasures against bridge abutment scour can be classified as either river training structures or armoring countermeasures. Other than design constraints, considerations in choosing the appropriate method of mitigation include maintenance and inspection requirements, enhancement of the physical environment, and constructability. Design specifications for many of these scour mitigation techniques can be found in Hydraulic Engineering Circular 23 (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. The major use of guidebanks in the United States has been to prevent erosion by eddy action at bridge abutments or piers where concentrated flood flow traveling along the upstream side of an approach embankment enters the main flow at the bridge (Lagasse et al., 2001). There also have been various studies on guidebanks. Among those studies are Spring (1903), Karaki (1959, 1961) Neill (1973), Bradley (1978), Chitale (1980), Smith (1984), Richardson and Simons (1984), and Lagasse et al. (2001). Guidebank orientation, length, crest height, shape, size, downstream extent, and other aspects were investigated. Design guidelines for guidebanks are given by Neill (1973), Bradley (1978), Ministry of Works and Development (1979), Central Board of Irrigation and Power (1989),and Lagasse et al. (1996, 1999, 2001). However, despite the design guidelines and studies in the literature, issues remain to be dealt with for certain types of bridges in certain environments. For instance, for small bridges where wing-wall abutments prevail and terminate on the riverbanks, specific design guidelines have not been devel- oped. Guidebanks have been specifically designed for spill- through abutments on rivers with wide floodplains; in such designs, the slope of the guidebank can be made tangent to the slope of the abutment so that there is no protrusion of the abutment into the flow beyond the slope of the guidebank. However, in a wing-wall abutment, this design may not be achieved readily because of the vertical front faces of the abut- ment. In this situation, either the slope of the guidebank pro- trudes out beyond the abutment face or the abutment face protrudes out beyond the guidebank slope. The impacts of these configurations on local scour at abutments need to be studied. Another issue is that a careful review of the guidelines for determining the length of guidebanks shows that the guide- lines designed for spill-through abutments in wide floodplain rivers may not apply to smaller bridges (Bradley, 1978). Many factors were not addressed that may be important for small bridges. For instance, first, it is recommended that if the length read from the design chart is less than 9.1 m (30 ft), a guidebank is not needed. This might not be true for a small two-lane bridge whose width is about 9 m; for such a bridge, a 9-m long guidebank may make a great difference in pro- tecting the bridge abutments. Second, it is recommended in the guidelines that for charts 9 to 30 m long, a guidebank no less than 30 m long be constructed. However, according to Herbich (1967), the length of the guidebank appears to be unimportant in reducing velocities provided that the length is greater than a certain minimum length. Therefore, an unnecessarily long guidebank may increase the cost of the structure and not improve its effectiveness. Yet another issue is that the parameters defined and used in determining the length of guidebanks may not be easily available—for instance, the total stream discharge, Q; the lateral or flood- plain flow discharge, Qf; and the discharge in the 100 feet of stream adjacent to the abutment, Q100 (Bradley, 1978). In addition, although an elliptical-shaped end seems to be favorable by all design recommendations because the curved head can direct the flow smoothly into the main channel and reduce scour at the guidebank end, the flood- plain flow velocity may be relatively low and a curved head may not be justified for small rivers and streams whose floodplains are relatively narrow and are mostly farmlands under cultivation. Most importantly, for abutments termi- nating on the riverbanks, a curved end stretching out from the bank into the farmland may be aesthetically and practi- cally unacceptable. This section deals with design issues for parallel-wall coun- termeasures on small rivers with wing-wall abutments. These parallel walls are essentially scaled-down, simplified versions 169

of guidebanks. The work fills a need for low-cost counter- measures for small bridges with wing-wall abutments. 9.3.2 Conceptual Model Figure 9-11 shows the conceptual model of a parallel-wall countermeasure against abutment scour in a compound channel. A wall is attached at the upstream end of the abut- ment and is parallel to the flow direction. There are multiple ways in which the parallel-wall coun- termeasure alters the flow field favorably. First, it can push the scour-inducing downflow and secondary vortex upstream away from the abutment so that scour will not occur at the upstream corner of the abutment provided that the length of the wall is long enough. Second, the wall can create a slow- moving or dead-water zone behind itself on the floodplain. In the case where there is no dead-water zone, the return flow from the floodplain would flow along the roadway and bridge abutment embankment toward the main channel, causing embankment scour. The presence of this wall, and thus the dead-water zone, helps slow down the scour and erosion of the embankment. Third, the wall helps straighten and improve the flow through the bridge crossing. 9.3.3 Solid Parallel-Wall Countermeasure Results A series of rectangular solid walls made from 13-mm thick plywood of different lengths, Ls, attached to the upstream end of the abutment and parallel to the flow direction were tested first as preliminary, proof-of-concept experiments. Solid walls are tested because, in certain geographical areas, rocks may not be readily available and cost efficient. All the solid parallel walls were seated at the bottom of the bank slope and aligned with the abutment face parallel to the flume wall. The top of each wall was the same height as the top of the abut- ment except in one clear-water case, in which the wall height was 52 mm lower than the water surface. The flow depth on the floodplain, yf, was equal to 52 mm, and the flow depth in the main channel, ym, was 132 mm. The velocity ratio, V/Vc, was about 0.9, 1.5, and 2.3 in the center of the entire channel for each of the three flow conditions tested. Table 9-3 gives 170 Figure 9-11. Conceptual model of the parallel-wall counter- measure against abutment scour in a compound channel. Solid-Wall Length, Ls Maximum Scour Depth at Abutment, ds (mm) Scour Reduction Rate (%) Maximum Scour at the Countermeasure, dc (mm) 0.3L, rectangular 62.5 19.6 86.5 0.5L, rectangular 40.1 48.4 81.0 0.6L, rectangular 29.5 62.0 77.1 0.7L, rectangular 21.5 72.3 78.3 0.8L, rectangular 14.3 81.6 76.2 1.0L, rectangular 3.3 95.8 77.1 1.2L, rectangular (Figure 9-14) -4.0* 105.1 83.0 1.0L,submerged, top even with floodplain 40.0 48.5 55.0 * Negative scour depth indicates deposition. Table 9-3. Solid-wall experimental results for clear-water scour (run time  80 hours, Q  0.0379  0.003 m3/s, V/Vc  0.9).

results of the solid-wall experiments under clear-water con- ditions, and Table 9-4 gives results under live-bed conditions. Figure 9-12 is a string contour of the 1.2L solid wall run in clear-water scour conditions, where ym is 132 mm, yf is 52 mm, Q is 0.0379 m3/s, and te is 80 hours. Flow was from left to right. Discussion of Solid-Wall Length Figure 9-13 shows the maximum scour depths at the abut- ment and the maximum scour depth in the vicinity of the upstream end of the wall versus the length of the wall in terms of the abutment length, La, for both clear-water (V/Vc  0.9) and live-bed (V/Vc 1.5) experiments. It is seen that, as the length of the wall increases from 0.3La to 1.2La, the scour at the abutment decreases rapidly. There was no scour at the abutment corner in 80 hours of running when the wall reached a length of 1.1La. It is also seen from Figure 9-13 that as the length of the wall increases from 0.6La to 1.5La, the scour at the abutment decreases rapidly for the live-bed case. The average scour depth at the abutment corner tends to zero if the wall length reaches a length of 1.6La. The amplitudes of the bed forms were significant, however. For instance, the maximum trough depths of the bed forms in the approaching channel ranged from 9.32 to 11.50 cm. The presence of the solid wall did not affect the dune amplitudes because the bed forms migrated past the wall and abutment, thereby causing the scour depth at the abutment to fluctuate about its average value, as men- tioned above. From the live-bed experimental data, it is seen that as the time-averaged scour depth at the bridge abutment changes from 5.05 cm (0.6La case) to 0.22 cm (1.5La case), the maximum contribution from the bed forms to the scour depth varies from 7.93 to 4.91 cm. It can be seen that although increases in solid-wall length decreased the amplitude of the bed forms, the decreases are not significant. Therefore, if the height of bed forms constitutes a large part of the local instan- taneous scour depth, scour can only be completely eliminated when the presence of the solid wall can change the flow con- dition in the bridge crossing into a transition regime under which the dunes completely disappear and a flat bed with bed material transport is formed. This transition regime may or may not be readily achieved depending on the approach-flow conditions and the constriction ratio of the channel. It was also found from the live-bed experimental data that, with a velocity ratio of 2.3, when the length of the wall was increased from 1.6La to 1.9La the scour reduction rate at the abutment decreased from 70 percent to 53 percent instead of increasing. This decrease may be due to imperfect construction of the floodplain or wall, or it may be that these two scour val- ues are within the range of scatter of the scour data for this high-sediment-transport flow. In summary, it was found that, in general, a solid parallel- wall countermeasure attached to the upstream end of the 171 Experimental Results Lspw = 0.6L Lspw = 0.9L Lspw = 1.2L Lspw = 1.5L Lspw = 1.6L Lspw = 1.9L Velocity ratio, V/Vc 1.5 1.5 1.5 1.5 2.3 2.3 Time-averaged scour depth at abutment, dabut,avg (cm) 5.05 4.04 2.57 0.22 2.23 3.57 Percent reduction in time-averaged scour depth at abutment, %max,abut,avg 30 44 63 92 70 53 Maximum instantaneous scour depth at abutment, dmax,abut,inst (cm) 12.98 9.56 10.12 5.49 8.84 8.94 Percent reduction in maximum instantaneous scour depth at abutment, %max,abut,inst 13 36 33 63 49 48 Time-averaged scour depth at the countermeasure, dcm,avg (cm) 7.57 7.90 7.74 7.86 9.21 9.82 Maximum instantaneous scour depth at the countermeasure, dmax,cm,inst (cm) 14.36 15.86 13.81 14.20 16.43 16.73 Table 9-4. Solid-wall experimental results for live-bed scour for six wall lengths, Lspw (run time  50 hours, Q  0.0619  0.0015 m3/s for V/Vc  1.5, and Q  0.0619  0.0015 m3/s for V/Vc  2.3; all walls were rectangular-shaped and emergent). Figure 9-12. A string contour of the 1.2L solid wall run in clear-water scour (contour interval  1 cm).

abutment was able to move the scour hole upstream from the abutment corner and, therefore, was effective as a scour coun- termeasure. It was also found that, for clear-water scour con- ditions, as the length of the wall increased, the scour at the abutment declined. In live-bed experiments, however, when the length of the wall becomes longer than 1.6La, then the scour at the abutment begins to increase. 9.3.4 Rock Parallel-Wall Countermeasure Results While the preliminary experiments mentioned above using a solid plate for a parallel-wall countermeasure were success- ful, there may be some bridge sites where rock is available. In these cases, it may be more economical and easier to construct if the parallel-wall countermeasure is a pile of rocks instead of a solid plate. Therefore, a series of rock walls of different lengths, Lw, and different protrusion lengths, Lp, were tested under both clear-water and live-bed conditions, as shown in Figure 9-14. In these experiments, the flow depth on the floodplain, yf , was 52 mm, and the flow depth in the main channel, ym, was 132 mm. The velocity ratio was 0.9 along the centerline of the entire channel for clear-water experiments and 1.5 and 2.3 for live-bed experiments. The top of each wall was the same height as the top of the abutment so that they were not submerged by the flow. The experimental results are tabulated in Table 9-5 for clear-water scour conditions and Table 9-6 for live-bed scour conditions. From Tests 1, 2, and 3 in Table 9-5, it was found that with a protrusion length of the wall base of 0.5W beyond the abut- ment face into the main channel, there tended to be a separa- tion zone behind the downstream end of the wall, causing a significant local scour hole. When the length of the wall was 1.5La, the scour hole was relatively small (8.81 cm) and did 172 Figure 9-13. Scour depth at both abutment and upstream end of solid walls versus wall length in terms of abutment length, La, for V/Vc  0.9 (clear-water scour conditions) and 1.5 (live-bed scour conditions). 0 0.4 0.8 1.2 1.6 Length of wall in terms of length of the abutment (L) -2 0 2 4 6 8 10 Sc o u r de pt h (cm ) dabut,avg, V/Vc=0.9 dcm,avg, V/Vc=0.9 dabut,avg, V/Vc=1.5 dcm,avg, V/Vc=1.5 Subscript “abut” denotes scour at the abutment that could threaten the abutment, and subscript “cm” denotes scour at the upstream end of the parallel-wall countermeasure tested (not threatening to the abutment).

173 Flow Floodplain A A1 Abutment ym Floodplain Section B-B1 Abutment Flow Riverbed Riverbed Wa Bm B1B Floodplain Parallel Wall V H LpW Parallel Wall Apron Apron Parallel Wall Lw Top View Main Channel La Section A-A1 Abutment Top and Bottom of Bridge Deck Figure 9-14. Parallel wall (aprons were present only in live-bed experi- ments) with Lw  1.2L. Table 9-5. Experimental data of parallel rock walls in clear-water scour (Q  0.0385  0.003 m3/s; te  80 hours; V/Vc  0.9; side slope, Sb  18/13.2; end slope, Sn  30/13.2). Experimental Result Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Test 7 Test 8 Test 9 Test 10 Test 11 Test 12 Gravel diameter, D (mm) 6.7~9.5 6.7~9.5 6.7~9.5 6.7~9.5 6.7~9.5 6.7~9.5 6.7~9.5 6.7~9.5 6.7~9.5 6.7~9.5 6.7~9.5 19.0~50.0 Wall length, Lw × La 1.5 0.5 1.0 1.5 0.5 1.0 1.5 1.0 0.5 0.25 2.0 1.5 Wall protrusion, Lp × W 0.5 0.5 0.5 0.3 0.3 0.3 0.0 0.0 0.0 0.0 0.0 0.0 Maximum scour at abutment, ds (mm) 2.0 52.1 50.6 4.5 53.0 23.5 19.0 19.5 20.0 28.0 20.0 3.0 Scour reduction (%) 97 29 35 94 32 70 76 75 74 64 74 96

not pose a direct threat to the abutment. However, when the length of the wall was 0.5La, the scour hole was 12.96 cm and the abutment was highly threatened. These scour holes could pose significant threat to a pier if a pier is located near the abutment. From Tests 4, 5, and 6 in Table 9-5, it was found that with a protrusion of the wall base of 0.25W beyond the abutment face into the main channel, there was still a separation zone right behind the downstream end of each wall. However, the scour holes caused by the separation in the 0.25W protrusion cases were smaller than they were in the corresponding 0.5W cases. For instance, the scour hole depths were 43.0 mm, 56.7 mm, and 80.4 mm for wall lengths of 1.5La, 1La, and 0.5La, respectively. However, these holes were closer to the abutment because of the reduced protrusion of the wall into the main channel. From Tests 7, 8, 9, 10, and 11 in Table 9-5, it was found that when the wall base did not protrude beyond the abutment, separation that appeared in the 0.25W and 0.5W protrusion cases disappeared. However, the abutment now was partly protruding out beyond the wall slopes, causing constriction of the flow coming from the wall slope. Fortunately, because of the high roughness of the wall, the near-wall velocity of the flow was retarded; as a result, the constriction of the flow did not cause significant scour at the abutment. It was found that the scour depth at the abutment ranged from 19 to 20 mm when the length of the wall varied from 0.5La to 2La. The scour depth was about 28 mm when the wall was 0.25La long, which indicated that for zero wall base protrusion, the scour depth at the upstream corner was not significantly affected by the length of the wall. Figures 9-15, 9-16, and 9-17 show the scour contours of Tests 3, 6, and 9, respectively, in Table 9-5. For live-bed scour conditions, it can be seen in Table 9-6 that a length of 0.5L requires the least amount of rock to build the wall and results in the same level of scour protection as longer walls. Even with a velocity ratio of 2.3, there is a 76-percent reduction in the maximum instantaneous scour depth at the abutment. The instantaneous scour depth is con- sidered to be more critical than the average scour depth because the abutment could collapse even in the short time in which the instantaneous scour depth was at its deepest level. 174 Experimental Result Lw = 0.5L Lw = 1.0L Lw = 1.5L Lw = 0.5L Velocity ratio, V/Vc 1.5 1.5 1.5 2.3 Time-averaged scour depth at abutment, dabut,avg (mm) 31.9 25.9 18.7 22.3 Maximum instantaneous scour depth at abutment, dmax,abut,inst (mm) 51.3 51.4 47.3 40.7 Percent reduction in time-averaged scour depth at abutment, %max,abut, avg 56 64 74 70 Percent reduction in maximum instantaneous scour depth at abutment, %max,abut,inst 66 66 68 76 te = 50 hours. All walls were rectangular shaped and emergent. Q = 0.0619 ± 0.0015 m3/s for V/Vc = 1.5. Q = 0.0966 ± 0.003 m3/s for V/Vc = 2.3. Table 9-6. Rock wall experimental results in live-bed scour for three different wall lengths, Lw. There is a small apron at the end. The wall base protruded out into the main channel from the abutment half-wall width. Flow is from left to right. Figure 9-15. Scour contours of Test 3 in Table 9-5. Figure 9-16. Scour contours of Test 6 in Table 9-5. There is a small apron at the end. The wall base protruded out into the main channel from the abutment a quarter-wall width. Flow is from left to right.

Discussion Figure 9-18 shows the scour depth at the bridge abutment versus rock wall length for different wall protrusion amounts for clear-water scour conditions. It can be seen that, for pro- trusion lengths, Lp, of 0.25W and 0.5W, increases in wall lengths can reduce scour at the abutment significantly. How- ever, for the case of no protrusion, increases in wall lengths do not show obvious effects in reducing scour at the abutment except when the wall is less than 0.5W. Figure 9-19 shows the maximum scour depth caused by the wall in the channel versus rock wall length for different wall protrusion lengths for the clear-water experiments. The fig- ure shows that for the 0.25W and 0.5W protrusion lengths, increases in wall lengths can significantly reduce the maxi- mum scour depth that is induced by the presence of the walls. For walls with protrusion length of zero, increases in wall length result in essentially no reduction in scour depth (i.e., scour at abutment) when wall lengths are greater than 0.5L. Figure 9-20 shows both time-averaged and maximum instantaneous scour depth at the bridge abutment versus rock wall length for zero protrusion length under live-bed condi- tions. The time-averaged scour depth was calculated by meas- uring the scour depth at regular time intervals and then averaging the depths over time, thereby giving a sense of the average depth of scour. The maximum instantaneous scour depth is the maximum scour measured at any time in the scour time series data collected. Even though this maximum scour value would not persist at the abutment, it could cause some brief undermining of the abutment structure and, therefore, is reported here as a parameter of interest. It is found that increases in wall lengths from 0.5L to 1.5L were able to reduce the time-averaged scour depth at the bridge abutment from 31.9 to 18.7 mm (a 41-percent reduction). 175 Figure 9-17. Scour contours of Test 9 in Table 9-5. 0 0.4 0.8 1.2 1.6 2 Stone wall length (L) 0 4 8 12 16 Sc o u r de pt h a t a bu tm e n t d s (cm ) Protrusion length, Lp, =0.5W Protrusion length, Lp, =0.25W Protrusion length, Lp, =0 a) Figure 9-18. Scour depth at bridge abutment versus rock wall length for different wall protrusion lengths under clear-water conditions (V/Vc  0.9). There is a small apron at the end. The wall base was even with the abutment. Flow is from left to right.

However, the increases in wall lengths only reduced the maximum instantaneous scour from 51.3 to 47.3 mm (an 8-percent reduction). The following general findings were made: • Walls that were set back onto the floodplain such that the base of the walls were even with the abutment (protru- sion length, Lp  0) were most effective in protecting the abutment. • For zero protrusion length (Lp  0), the scour protection of the walls was not sensitive to the length of the wall unless the length was less than 0.5L. • Walls whose base protruded into the main channel beyond the abutment (Lp  0.25W or 0.5W) tended to produce sig- nificant scour in the bridge crossing and potentially threaten the middle and downstream abutment end when the length of these walls became shorter than a certain length. Regarding the amount of mass transfer or water flow through the rock wall itself during the experiments, dye was injected in the floodplain side of the parallel walls. For all of the flow conditions investigated, no significant amount of dye 176 0 0.4 0.8 1.2 1.6 2 Stone wall length (L) 0 4 8 12 16 20 M a x im u m s c o u r de pt h ca u s e d by w a ll (cm ) 0.5W protruding 0.25W protruding No protruding a) Figure 9-19. Maximum scour depth caused by the wall in the entire channel versus rock wall length for different wall protrusion lengths under clear-water conditions (V/Vc  0.9). Figure 9-20. Time-averaged and maximum instantaneous scour depth at bridge abutment versus rock wall length for zero protrusion under live- bed conditions (V/Vc  1.5).

was observed to flow through the rocks. This indicated that there is no significant flow transfer through the wall. In addi- tion, a parallel solid wall with length 0.8La, with 85 circular 8-mm diameter holes uniformly distributed on the wall, was tested and compared with the 0.8La impermeable wall. The results were similar. This showed that when the permeability of the wall is smaller than a certain value, the solid wall acts like an impermeable one. Cases of clear-water and live-bed scour were investigated. It could be that during flow conditions in which there was less downstream velocity, there is more flow through the wall. This would most likely not be critical for scour, however, since clear-water and live-bed scour are the two worst scour cases. In all of the experiments described in this report, there was no physical gap between the countermeasure and the abut- ment. Care should be taken to ensure this, as a high-velocity jet may form if such a gap exists, which could exacerbate the scour depth in unknown ways. 9.3.5 Design of Parallel-Wall Countermeasure for Scour Prevention at Wing-Wall Abutments General preliminary design guidelines can be established for the use of parallel solid walls and parallel rock walls to reduce scour at typical bridges with wing-wall abutments ter- minating on or protruding beyond the main channel banks. Protrusion of Wall The best position for solid walls would be such that the solid wall’s face is aligned with the abutment’s face so that there is no protrusion by either structure. For parallel rock walls with lateral protrusions of 0.25W and 0.5W, a separation zone is formed behind the walls, caus- ing scour holes and threatening the abutment foundation. Also, the protrusion of the wall into the main channel further constricts the bridge crossing and reduces its conveyance capacity. Therefore, parallel rock walls of zero protrusion are recommended. Length of Wall For a solid parallel-wall countermeasure, the clear-water experimental results showed that a solid wall with a length of 1.1La will completely eliminate the local scour at the bridge abutment, while a solid wall of length 1.6La with a velocity ratio of 1.5 will be able to reduce the time-averaged scour to zero. For a velocity ratio of 2.3, a solid wall of length 1.6L will reduce the time-averaged scour up to 70 percent. Further increases in the length of the wall will result in a decrease in scour reduction rate. Therefore, it is recommended that the length of the parallel solid wall be 1.6La. For a parallel-wall countermeasure made of piled rock of zero protrusion, a rock wall length of 0.5La is recommended. Height and Width of Wall Crest Heights for both solid and rock parallel-wall countermea- sures should be high enough to prevent the flow from enter- ing the bridge crossing at the abutment, even in the worst case scenario. Slope of Wall and Apron The side slope of the rock wall must be less than the rock’s angle of repose to ensure stability. However, the side slope should be as high as possible so that the protrusion of the abutment beyond the wall slope is a minimum. It is recom- mended that the side slope of the wall be about 5 degrees less than the angle of repose of the rocks. Aprons are always needed at both the bottom of the side slope and the upstream end slope. When scour occurs in the riverbed adjacent to those locations, the rocks from the apron will launch (that is, fall) into the scour hole. This allows the rocks that make up the side slope and upstream end slope to remain intact. Apron thickness, area limit, and relative posi- tion with the parallel rock wall should be determined accord- ing to the scour depth and position of the scour hole along the wall. Even though Figure 9-11 shows the water hitting the abutment and coming back, the water’s velocity is so slow that erosion will be negligible. Comparison of Solid and Rock Parallel Walls If the best designs for both solid and rock parallel-wall countermeasures are compared, it can be seen that the rock walls have advantages over the solid walls. Table 9-7 shows scour depths for both solid and rock walls for both clear- water and live-bed scour conditions. The table shows that the rock wall allows less scour at the abutment than the solid wall does. The solid wall, therefore, seems to be feasible only when the cost of rocks is prohibitively high. 9.3.6 Conclusions Regarding Parallel-Wall Countermeasures The following conclusions were made about parallel-wall countermeasures: • A parallel solid wall attached at the upstream corner of the abutment parallel with the flow can be used as a counter- measure against abutment scour. The length of the solid 177

wall should be 1.6La to obtain acceptable scour reduction rate at the abutment for the conditions tried in this study. • A parallel solid wall attached at the upstream corner of the abutment parallel with the flow may or may not be able to reduce the amplitude of the bed forms that pass through the bridge opening, depending on the changes of the flow parameters from the approaching channel after entering the bridge crossing. • There may be significant scour at the upstream solid wall end, so no other structures should be located in this region. • Parallel rock walls attached at the upstream of the abut- ment can also be used as countermeasures against scour at the abutment. The foot of the wall should not protrude into the main channel beyond the abutment, and a top wall length of 0.5L will provide sufficient protection. The side slope of the rock wall should be on the order of 30 degrees, but in no case should it be steeper than about 70 percent of the rocks’ angle of repose. • Rock walls have more advantages than solid walls in terms of efficiency, stability, and cost. 9.4 Spur Dike Countermeasure The next flow-altering countermeasure described is a combi- nation of spur dikes located locally to the abutment. The prob- lem is described, and then the results of the lab tests are given. 9.4.1 Introduction Spur dikes—structures that project from the bank into the channel—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).They may be classified according to their permeability: high-permeability spur dikes are “retarder”spur dikes, impermeable spur dikes are “deflector” spur dikes, and intermediate-permeability spur dikes are “retarder/deflector” spur dikes (Brown, 1985b). They may be constructed out of a variety of materials, including masonry, concrete, earth and rock, 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 submerged only by the largest flow events. 178 Countermeasure Type Clear-Water Scour Live-Bed Scour Solid Parallel Wall -4.0 at abutment* 83.0 at solid wall 54.9 at abutment 78.6 at solid wall Rock Parallel Wall 3.0 max. at abutment 43.3 max. at rock wall 47.3 max. at abutment 85.8 max. at rock wall * Negative scour depth indicates deposition. Table 9-7. Comparison of rock and solid wall countermeasure performance (scour depth, mm). A spur dike serves one or more of the following functions: • It trains the stream flow. For instance, spur dikes are commonly used to realign streams as they approach 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 because of meandering. • It protects the stream bank (which may or may not contain bridge abutments) from erosion. • It increases the flow depth for navigation (Garde et al., 1961) or improves aquatic habitat. In recent years, porous and overflow-type spur dikes have been shown to provide improved pool habitats for fish and other aquatic life in severely degraded streams (Shields et al., 1995c). Volumes of the scour hole in the vicinity of model spur dikes were measured in a laboratory flume under clear-water overtopping flows with varying angles and contraction ratios to maintain bank protection and enhance aquatic habitats (Kuhnle et al., 1997, 1998, 1999). In addition, because of the deposition that spur dikes induce, spur dikes may protect a stream bank more effectively and at less cost than revetments (Lagasse et al., 2001). Spur dikes constructed on or adjacent to an abutment to counter local scour have not been previously tested. This technique would constitute a combination of bank hardening and flow altering to counter local scour. When spur dikes are properly placed around the abutment, the flow can be redi- rected and scour near the abutment can be reduced. Local scour can threaten the spur dike itself, however, as shown in Figure 9-21. 9.4.2 Conceptual Model Figure 9-22 shows the flow patterns around a spur dike as a countermeasure against abutment scour in a compound chan- nel. A spur dike is placed a certain distance upstream of the abutment and is perpendicular to the flow direction. Flow on the floodplain can only go around the main channel end of the spur dike. A spur dike thus installed is expected to be able to block the floodplain flow from hitting the abutment face and

direct the flow into the main channel. It may create wake vor- tices behind itself. The effects of these wake vortices at the spur dike structure on the abutment scour were evaluated experi- mentally to determine the best configuration of spur dikes as countermeasures. In the experimental studies described next, the number of spur dikes, the distance between spur dikes and between spur dikes and abutment, the protrusion length, and the construction material of spur dikes will be tested as param- eters. Figure 9-23 shows the variables involved. 9.4.3 Results Results are given next for both solid and rock spur dikes. Solid Spur Dikes A preliminary proof-of-concept series of solid spur dike experiments were performed first. The spur dike was made from 13-mm thick plywood and was tested under the same flow condition as in the clear-water baseline test (V/Vc 0.9). The flow depth in the main channel, ym, was 132 mm, and the flow depth in the floodplain, yf, was 52 mm. Only one spur dike located upstream of the abutment was tested in each experiment. The variables experimented with for the solid spur dikes were length, Lsp; distance upstream of the abut- ment, Ds; and orientation angle with respect to the flow, . In all cases, the top of the spur dikes was higher than the water surface. The experimental results are listed in Table 9-8. The flow perpendicular length of the spur dike was found to be an important variable in protecting the abutment. Flow per- pendicular lengths were restricted to the length of the abut- ment or less in this experimental series to prevent excessive contraction of the flow in the main channel or backwater effects. The six test cases showed that spur dikes of the same flow perpendicular length as the abutment do not protect the abutment from scour regardless of spacing or orientation angle, as shown in Figures 9-24, 9-25, and 9-26.When the spac- ing (Ds) of the spur dikes was less than the flow perpendicular length of the abutment (La), the scour hole induced by the spur dike will encompass the upstream corner of the abutment (Figure 9-26). When the spur dikes were far away (Ds  1.5La) from the abutment, a narrow channel formed between the deposited sand and the abutment corner (Figure 9-25). This caused scour at the upstream corner of the abutment. In addi- tion, the spur dikes caused the formation of huge scour holes that would undoubtedly threaten the stability of the stream bank and the stability of the spur dike itself. In summary, the effective reduction of local scour at bridge abutments using spur dikes requires that their flow perpendicular length be greater than the flow perpendicular length of the abutment. Rock Spur Dikes Four cases of rock spur dikes, each with the same length perpendicular to the flow, were tested under clear-water flows. These rock spur dikes were constructed from the flume wall 179 Figure 9-21. Excessive scour around a poorly positioned spur dike (flow is from left to right). Figure 9-22. Flow patterns around a spur dike as a counter- measure against abutment scour in a compound channel.

180 Figure 9-23. Definition sketch for spur dike scour countermeasure. Experimental result Test Sp-1 Test Sp-2 Test Sp-3 Test Sp-4 Test Sp-5 Test Sp-6 Spur dike description notes Rectangular, protrusion length equal to the width of floodplain Figure 9-24 Figure 9-25 Figure 9-26 Spur dike protrusion length, Lsdp (La) 0.7 1.0 0.7 0.7 0.7 1.0 Distance between the farthest spur dike tip at the main channel end and abutment tip, Ds (La) 2.0 2.0 1.0 1.5 1.5 0.6 Spur dike orientation angle with respect to the flow, θ (deg) 90 90 45 45 45 90 Run time, te (hours) 20.9 25.7 80.0 28.0 24.0 43.0 Time-averaged scour depth at abutment, d,abut,avg (mm) 45.2 46.0 46.9 45.0 43.3 58.0 Percent of scour reduction, %max,abut,avg (%) 32.2 34.8 39.6 35.2 36.5 20.6 Maximum scour depth at spur dike, dmax,sp,avg (mm) -- 105.2 144.0 139.0 135.3 -- Table 9-8. Preliminary solid spur dike experimental results (Q  0.0387  0.003 m3/s, V/Vc  0.9, ym  132 mm, yf  52 mm). Flow Floodplain A Abutment ym Floodplain Section B-B1 Abutment Flow Riverbed θ Riverbed Wa ssd Lsd Bm B1B Hsd Lsd Spur Dike Floodplain HdWsd ss Spur θ Top View Main Channel A1 La Section A-A1 Abutment Top and Bottom of Bridge Deck

was 1.5La (66 cm). Each spur dike had a top width of 100 mm and a bottom width of 400 mm. The rocks were 6.7 to 9.5 mm in diameter. In these cases, the flow depth on the floodplain, yf,was 52 mm and the flow depth in the main channel, ym,was 132 mm. The mean velocity ratio, V/Vc, was 0.9 in the middle of the main channel. The top of each spur dike was the same height as the top of the abutment. The experimental data for rock wall spur dikes in clear-water scour conditions are sum- marized in Table 9-9. The clear-water rock spur dike data, illustrated in Fig- ures 9-27 through 9-30, show that that spur dikes with a top length of 1.0La and bottom length of 1.5La provided protec- tion to the bridge abutment from scour. These lengths are the minimum lengths required to be sufficient to protect the abutment. It was found from Tests Sp-8 (Figure 9-28) and Sp-9 (Figure 9-29) that as the distance between two successive spur dikes increased from 1.0La to 2.0La, the scour depth between the first two spur dikes increased from 76.2 mm to 131.4 mm, more than a 70-percent increase. Although this increase in scour depth did not pose a direct threat to the abutment, it did threaten the two spur dikes and could have caused these spur dikes to partially fail. Therefore, the increase in scour depth ultimately may pose an indirect threat to the abutment. From the scour condition between the second and the third spur dikes in these two tests, it was found that in Test Sp-8, the scour depth was 66.1 mm and greater than the scour depth in Test Sp-9. However, the scour hole in Test Sp-8 was deflected out into the main channel by the spur dikes and had no effect on the abutment, while in Test Sp-9, the abutment was threat- ened. Therefore, the best spacing between spur dikes is con- cluded to be 1.0La. Generally, the larger the number of spur dikes, the better the abutment will be protected, but at a higher cost. To min- imize the cost, the number of spur dikes should be mini- mized. With this in mind, Sp-10 (Figure 9-30) was performed with only two end-slope spur dikes attached at the upstream and downstream corners of the abutment. It was found that the two spur dikes located in this configuration directed the flow away from the abutment, thereby protecting the abut- ment and upstream and downstream banks. The advantages of these two end-slope spur dikes are that (1) they can direct the flow away from the abutment and protect the abutment; (2) they can provide extra stabiliza- tion to the abutment, especially when some rocks fill in the existing scour holes at both corners of the abutment; (3) as the scour in the bridge crossing develops, the two spur dikes sink and collapse to the original bed level and the rock mate- rial is distributed by the flow such that they are still able to function as riprap to protect the abutment; and (4) the two spur dikes attached at the abutment face use less rock than spur dikes attached at the floodplain bank. The disadvantage of the two end-slope spur dikes is that they may cause 181 Figure 9-25. Photograph of Sp-5 (flow is from left to right). Figure 9-24. Photograph of Sp-3 (flow is from left to right). Figure 9-26. Photograph of Sp-6 (flow is from left to right). extending out into the main channel perpendicular to the flow direction. At the stream end of each spur dike was a transverse slope of 22/13.2. Therefore, the top protrusion length of the spur dike was La (44 cm) and the bottom protrusion length

contraction scour. In relatively wide bridge crossings, this may not cause a significant problem, but it may be a prob- lem for narrower ones. No significant backwater increase due to the addition of the rock spur dikes was observed in any of the experiments, so increased backwater is not seen as a disadvantage. Positioning of the spur dikes is important to the protection of the abutment. As shown in the baseline case, the upstream corner of the abutment was the point that was most likely to scour. Another scour-prone location would be the down- stream end of the abutment where the flow leaves the bridge crossing and separates. Therefore, spur dikes located at both ends of the abutment yielded the best results, as shown in Test Sp-10. The advantage of this configuration is that by being attached to the face of the abutment instead of the bank of the flood channel, the amount of rock material is minimized and the cost of the construction of the spur dikes is reduced. 182 Experimental Result Test Sp-7 (Figure 9-27) Test Sp-8 (Figure 9-28) Test Sp-9 (Figure 9-29) Test Sp-10 (Figure 9-30) Number of spur dikes, Nsd 2 3 3 2 Spacing between spur dikes or spur dike and abutment Ds (La) 1 1 2 1 Time-averaged scour depth at abutment, dabut,avg (mm) 0 0 20.0 0 Percent of scour reduction, %max, abut (%) 100 100 74.3 100 Maximum scour depth behind the first spur dike, dmax,sp1,avg (mm) 87.5 76.2 131.4 75.6 Maximum scour depth behind the second spur dike, dmax,sp2,avg (mm) 110.0 66.1 35.0 103.0 Maximum scour depth behind the third spur dike, dmax,sp3,avg (mm) -- 77.1 68.9 -- Table 9-9. Clear-water experimental data of rock spur dikes (Q  0.0368  0.0016 m3/s, te  80 hours). Figure 9-27. Scour contour of Test Sp-7 with two spur dikes upstream of the abutment (flow is from left to right). Figure 9-28. Scour contour of Test Sp-8 with three spur dikes, including the two formed by the abut- ment (flow is from left to right). Figure 9-29. Scour contour of Test Sp-9 with three spur dikes, including the one formed by the abut- ment (flow is from left to right). Figure 9-30. Scour contour of Test Sp-10 with two spur dikes, both located at the abutment (flow is from left to right).

From the clear-water experimental data, the configurations of Test Sp-8 and Sp-10 were considered to have the most potential for protecting the abutment and were tested further under live-bed flows. No further experiments were conducted with the configurations used in Sp-7 or Sp-9. The same spur dike configurations used in Tests Sp-8 and Sp-10 were tested under live-bed conditions as countermea- sures against scour at the abutment. Test Sp-11 had basically the same spur dike configuration as Case Sp-10. Test Sp-12 had a similar spur dike configuration as Test Sp-11 except that there was initially a semicircular ring-shaped apron around each of the spur dike ends. Each apron had a width of about 200 mm and was about three rock diameters thick. Tests Sp-13 and Sp-14 had similar spur dike configurations as Test Sp-10 except the aprons and rock size varied. There were similar aprons around the first two spur dikes in the later cases. In these cases, the flow depths were the same as for the clear- water experiments. The velocity ratios used were 1.5 and 2.3. The rock diameters used were 19 to 50 mm for the 1.5 veloc- ity ratio and 500 to 700 mm for the 2.3 velocity ratio. These rock sizes were chosen to avoid rock entrainment and trans- port by the flow. The top of each spur dike was approximately the same height as the top of the abutment. The live-bed experimental data are listed in Table 9-10. Data of Test Sp-11, illustrated in Figure 9-31, showed that with a velocity ratio of 1.5, two dikes attached at both ends of the abutment were able to reduce scour over 100 percent (i.e., cause deposition). Deposition of 10 to 40 mm occurred between the two slopes at the abutment during 50 hours of running time. The maximum scour at these spur dikes hap- pened at Point A, where the foot of the upstream bank and the upstream side of the first spur dike meet. This was different from the scour pattern in clear-water scour conditions. Sedi- ment moved along a scour zone starting from Point A and proceeded along the edge of the launched apron of the spur dikes toward Points B, C, and D, covering part of these aprons and extending past the bridge crossing while keeping a dis- tance from the abutment face. Therefore, after equilibrium was reached, the deposition between the two slopes at the abutment was not affected any more. Local scour holes were found at Points C and D, where flow separated between the two spur dikes. The top of the two spur dikes was initially as high as the flow surface in Test Sp-11. As the scour devel- oped, the rocks on the slope of the spur dikes kept sliding into the scour hole, causing the top of the two spur dikes to sub- side until the equilibrium state of the scour process was reached. At the end of the test, the top of the first spur dike sank 75.0 mm and the top of the second spur dike sank 20.0 mm. To reduce this sinking, Test Sp-12 was performed with two semicircular aprons of 200-mm width and three-rock- diameter thicknesses placed around the edge of each spur dike. After 50 hours of running time, the presence of the aprons helped improve the deposition between the two spur 183 Experimental Result Test Sp-11 (Figure 9-31) Test Sp-12 Sp-13 (Figure 9-32) Sp-14 (Figure 9-33) Number of spur dikes, Nsd 2 3 3 2 Velocity ratio, V/Vc 1.5 1.5 1.5 2.3 Time-averaged scour depth at abutment, dabut,avg (mm) -10.3* -14.2* -26.6* -3.0* Percent reduction in time-averaged scour depth at abutment, %max,abut,avg (%) 114 120 136 100 Time-averaged scour depth in front of the first spur dike, dmax,sp1,avg (mm) 51.1 44.5 53.9 69.7 Maximum instantaneous scour depth in front of the first spur dike, dmax,sp1,inst (mm) 103.5 95.1 89.0 109.4 Time-averaged scour depth at the second spur dike, dmax,sp2,avg (mm) 49.2 50.3 46.8 -- Instantaneous scour depth at the second spur dike, dmax,sp2,inst (mm) 99.3 74.9 77.4 -- * Negative scour depths indicate deposition. Table 9-10. Live-bed experimental data of rock spur dikes (Q  0.0627  0.003 m3/s for a velocity ratio of 1.5 and 0.0985 m3/s for a velocity ratio of 2.3, ym  132 mm, yf  52 mm, running time, te  50 hours). Figure 9-31. Test Sp-11 with two spur dikes, one at each end of the abutment (flow is from left to right).

dikes at the abutment and reduced the scour depth at the upstream side of the first spur dike, as shown in Table 9-10. However, the top of the first spur dike still sank 50.0 mm. Although the configuration in Tests Sp-11 and Sp-12 can protect the bed around the abutment, there is a concern about the portion of the floodplain at the upstream corner of the abut- ment. For erodible floodplains, the spur dikes thus placed are not able to protect the floodplain. Compared with Tests Sp-11 and Sp-12, data in both Table 9-10 and Figure 9-32 showed that the configuration of Test Sp-13 can protect not only the channel bed around the abutment but also the floodplain and the abutment fill. The minimum deposition of sediment around the abutment was found to be 26.6 mm. Also, because of the protection of the first spur dike, the spur dikes at both corners of the abutment experienced very little subsidence. The same spur dike configuration as in Test Sp-13 was tried further in Test Sp-14 with a velocity ratio of 2.3. The spur dikes were made of rocks with diameter sizes 500 to 700 mm in order to resist transport by the flow. The spur dikes still protected the abutment from scour successfully, as shown in Table 9-10 and Figure 9-33. Rock spur dikes have more advantages than solid spur dikes. First, unlike solid spur dikes, rock spur dikes do not need a traditional foundation. Instead, they can use aprons around the structure edges as the scour holes develop and prevent them from failing. Second, a sloped end at the main channel end of a spur dike whose top protrusion length is La will provide extra protrusion length and more deflection. Third, rock spur dikes may make deposition at the upstream end of the abutment possible. The reason that the sediment did not deposit at the upstream corner of the abutment is that the abutment structure had a very smooth surface, and the flow velocity near the abutment surface was relatively high. This unimpeded velocity prevents settling of sediment. To conquer this problem, a pile of gravel placed at the upstream abutment end increases the roughness of the abutment and decreases the flow velocity so that sediment can deposit at the upstream corner of the piled rocks. Fourth, upward-sloping rock spur dikes with relatively high friction roughness slow and guide the flow to climb up the slope instead of producing scour-inducing downflow. 9.4.4 Design of Spur Dikes for Scour Prevention at Wing-Wall Abutments From the experimental results, it was concluded that spur dikes with a top protrusion length of 1.0La and a bottom length of 1.5La were sufficiently long to protect the bridge abutment. The amount of material in the spur dikes can be greatly reduced if the spur dikes are attached at the face of the abutment. The top length 1.0La is believed to be the min- imum length required to protect the abutment, while the bot- tom length is designed as a function of the rock’s angle of repose. It was concluded that, for straight channels, the best spacing between successive spur dikes was 1.0La. A spacing of 1.0La or less was able to restrict the flow from full separation behind each spur dike. As a consequence, the scour depth behind each spur dike was reduced, and the scour hole was pushed farther away from the spur dike end into the main channel. It was concluded that three spur dikes, with the first one located 1.0La distance upstream of the upstream abutment corner, and the remaining two attached at the upstream and downstream corners of the abutment, would be the best con- figuration for preventing scour of the bed near the abutment for this experimental setup (streamwise width of abutment is around 1.0La). For a bridge abutment whose streamwise width is longer than 1La, a spur dike attached at the downstream end is still recommended, but with additional spur dikes located upstream at distances of 1.0La until the upstream corner of the abutment is met. One more spur dike is preferred upstream of the one at the upstream corner of the abutment. For a bridge abutment whose streamwise width is less then the flow- perpendicular length of the abutment (La), three spur dikes are recommended: one at the upstream corner of the abutment, one at a distance of 1.0La upstream of the abutment, and the other at the downstream corner of the abutment. 184 Figure 9-32. Test Sp-13 with three spur dikes (flow is from left to right). Figure 9-33. Test Sp-14 with three spur dikes (flow is from left to right).

In addition to the variables directly tested in the laboratory experiments described above, the following design guidelines are offered. The top height of each spur dike should be high enough so that it is not overtopped by the flow during flood- ing, since all of the experiments described here are for emer- gent spur dikes and the flow patterns will be greatly altered if the spur dikes were overtopped. The rock size of the spur dikes should be great enough to resist the flow stress in the worst flood situation. It was observed in the laboratory that end slopes of spur dikes can be constructed as steep as possi- ble since they are able to adjust themselves to a stable state as scour holes develop around them. Semicircular aprons launch rocks into the scour hole during its development. 9.4.5 Conclusions Regarding Spur Dike Countermeasure For the clear-water and live-bed scour at a Type III abut- ment configuration in a straight channel, it can be concluded that: • A single spur dike made of a solid plate, having a protru- sion length the same as or shorter than the abutment, and placed upstream of the abutment was not able to protect the abutment. The downflow and the principal vortex are very strong at the stream end of the structure. As a conse- quence, a significant scour hole was always found at the end of the structure, and this hole threatened both the structure and the channel bank. • Rock spur dikes show several advantages over rigid spur dikes and are preferred. • Three rock spur dikes, as configured in Tests Sp-9 and Sp-13, were considered the best configuration for protecting the abutment. This configuration can provide 100-percent pro- tection to the abutment under the velocity ratios of 0.9, 1.5, and 2.3. Two spur dikes at the upstream and downstream corners of the abutment were also successful at preventing scour in both clear-water and live-bed experiments. 9.5 Abutment Collar Countermeasure A flow-altering countermeasure that has not previously been tested for abutments in a compound channel is a hori- zontal collar. After describing the flow patterns around a col- lar, results are given for various collar configurations. 9.5.1 Introduction Collars attached to piers have been studied as either an armor layer of the bed or a downflow-halting device by Kapoor and Keana (1994), Kumar et al. (1999), and Borghei et al. (2004). Collars block the downflow found at the leading edge of piers and abutments and eliminate scour-inducing secondary vortices. This chapter describes laboratory experiments with collars at a vertical-face wing-wall abutment placed at the main channel edge, an abutment configuration typical of older bridges on smaller streams. 9.5.2 Flow Pattern at Collar Attached to Abutment Figure 9-34 shows the flow patterns of a collar as a counter- measure against abutment scour in a compound channel. A collar is attached around the bridge abutment and has a cer- tain aerial coverage. A collar thus installed is expected to be able to prevent the bed materials from being entrained by the return flow from the floodplain, the downflow, and the sec- ondary vortex systems. In the experimental studies described in this section, the aerial extent in all directions and the verti- cal elevation of the collar were studied under clear-water 185 Figure 9-34. Flow patterns of a collar as a countermeasure against abutment scour in a compound channel.

conditions to determine the best configuration for a collar to be a successful countermeasure. Figure 9-35 shows the collar countermeasure. 9.5.3 Collar Results A series of collars of different lengths and widths were attached to the bridge abutment under clear-water conditions as countermeasures against scour at the abutment. These col- lars were made from steel and were seated horizontally at the desired elevation. The flow depth on the floodplain, yf, was 52 mm, and the flow depth in the main channel, ym, was 132 mm. The velocity ratio, V/Vc, was 0.9 at the center of the entire channel, as in the baseline tests. Table 9-11 gives the dimensions of each collar configuration tested, as well as the experimental results. Figure 9-36 shows scour contours for the equilibrium con- dition for Test T3. It was found that the collars were able to protect the bridge abutment efficiently by isolating the return flow and the secondary vortices from the bed around the abutment that ordinarily would cause local scour. The mini- mum collar dimensions that eliminated local scour were those with a width of 0.23La (La is the abutment length per- pendicular to the flow direction) for elimination of local scour, a width of 0.8L for maximum reduction of scour at the edge of the collar, and a vertical location of 0.08ym below the mean bed sediment elevation (ym is the main channel flow depth). After removal of the collar, there was no scour observed under the collar around the abutment for any of the cases tested. 9.5.4 Discussion Protrusion Width Figure 9-37 shows the maximum scour depth at both the bridge abutment and the main channel edge of the collar ver- sus the transverse collar width for all collar cases when the collar elevation was 10 mm below the initial bed level. It can be seen from Figure 9-37 that the maximum local scour depth under the main channel edge of the collar decreased from −71.0 mm to −10.0 mm as the width of the collar beyond the abutment increased from 100 mm to 350 mm. Further examination of the experimental results shows that the maximum local scour depths at the main channel edge of 186 Flow Floodplain La ym Floodplain Section B-B1 Abutment Top and Bottom of Bridge Deck Flow Riverbed Wa Bm B1 Floodplain Collar Xu WcB yc Wc Collar Riverbed Collar ym X Top View Main Channel A A1 Abutment Section A- 1 Abutment Top and Bottom of Bridge Deck Figure 9-35. Abutment scour collar countermeasure.

each of these collars had a similar magnitude as the scour depth at the same location in the baseline case with no coun- termeasures. Figure 9-38 shows the transverse bed profile in the bridge crossing of the baseline case and the scour profile formed by the maximum local scour depth values under the edge of the various collars of different widths. Figure 9-38 suggests that the presence of the collar did not change the strength of the vortex, but protected the abutment from scour by not allowing the scour-inducing secondary vortex to inter- act with the bed sediment. Collar Elevation To determine the optimal collar elevation, three different ele- vations of the collars were used in the experiments. Figure 9-39 shows the scour depth at the abutment and at the edge of the collars versus collar elevation for collars with a width of 100 mm. It is evident that a collar elevation of 10 mm below the original bed level had the least scour.This corresponds to an ele- vation of 1/13.2, or 0.08ym, where ym is the flow depth in the main channel. The collar should be lower than the bed in order to keep the secondary vortex above it and not interacting with the bed sediment. Streamwise Collar Length At the upstream edge of the collar, a shallow scour hole per- pendicular to the flow was found in Tests T2 through T4. This scour hole started from the main channel bank and went trans- versely toward the opposite channel wall and was connected to the scour hole at the main channel edge of the collar. This scour hole remained at the collar leading edge and, therefore, did not 187 Test No. Dimensions (mm) Elevation dmax.abut (mm) Percent reduction in time-averaged scour depth at abutment, %max,abut,avg dmax.col (mm) T1 Floodplain level 44.8 42.3 -- T2 Bed level 35.1 54.8 81.4 T3 10 mm below bed (Figure 9-36) 19.2 75.3 71.0 T4 20 mm below bed 20.0 74.3 78.9 T5 10 mm below bed 10.0 87.1 45.4 T6 10 mm below bed 10.0 87.1 10.0 Note: dmax.abut = maximum scour depth at the abutment foundation; dmax.col = maximum scour depth at the collar edge. Table 9-11. Dimensions and positions of collars tested (run time  80 hours, ym 132 mm, yf 52 mm, Q 0.0387  0.001 m3/s, V/Vc 0.9). Figure 9-36. Elevation contours of Test T3 with collar at 1 cm below bed elevation (flow is from left to right).

188 -80 -70 -60 -50 -40 -30 -20 -10 0 0 50 100 150 200 250 300 350 400 Transverse Collar Width (mm) Be d El ev (m m ) At Abutment At Collar Edge Figure 9-37. Scour at both bridge abutment and the main channel edge of the collar versus the transverse collar width for a collar elevation of 10 mm below the bed. 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 Transverse Distance from Abutment (mm) B ed El ev at io n (m m ) Without collar With collar Figure 9-38. Transverse scour profile in the bridge crossing of the baseline case and the scour profile formed by the maximum local scour depths under the main channel edge of the various collars of different widths at the end of 80 hours (view is looking downstream).

threaten the abutment. In Tests T5 and T6, the upstream end of the collars was still buried in the sand at the end of the exper- iments and, therefore, no scour was found. The upstream end of the collar should be long enough that the scour hole won’t threaten the abutment. There was always scour downstream of the trailing edge of the collar. For Tests T3 and T4, the trailing edge of the collar ended in the middle of the bridge crossing, and scour holes of more than 19.2 mm at the abutment were found in both cases. These scour holes posed a threat to the middle of the abutment structure and may be eliminated simply by extend- ing the trailing edge of the collar to a location that is down- stream of the abutment structure, as shown in Figure 9-40. The extension of the downstream collar length may increase scour magnitude. This was observed in Tests T5 and T6, where the scour hole was more than 5 cm in Test T5 and more than 6 cm in Test T6. The scour location is not in the bridge crossing, however, and, therefore, should not correspond with a pier location. Temporal Scour Variation It was observed that, unlike the rapid scour at the upstream and downstream abutment corners in the baseline case, the scour in the first 10 hours under the main channel edge of the collar was very slow in all collar cases. Figure 9-41 shows the temporal evolution of scour under the edge of the plate in Test T3. Note the delayed scour in the first 10 hours by collars. This delayed scour constitutes another advantage of using the abutment collars. 9.5.5 Conclusions From these clear-water experimental data, the following conclusions can be made: • Collars were found to be effective at preventing local scour at wing-wall bridge abutments. The collars isolated the tur- bulent flow and vortex systems from the bed material and thereby prevented the bed underneath the collar from scouring. • The farther the collar extended downstream of the abut- ment, the farther downstream the scour hole was located. As the transverse width of the collars increased, the depth of the scour hole at the edge of the collar decreased. 189 Figure 9-39. Bed elevation at the abutment and at the edge of the collars versus collar elevation (all collars had a transverse width of 10 mm from the abutment face). -70 -60 -50 -40 -30 -20 -10 0 0 20 40 60 80 100 Time (hr) Be d El ev at io n (m m ) Edge of Collar Upstream, w/o collar Downstream, w/o collar Figure 9-40. Scour contour of Test T6 with a collar attached along the abutment (collar width is 350 mm, collar elevation is 10 mm below the original bed, flow is from left to right). Figure 9-41. Scour depth variation under the main channel edge of the collar and for both the upstream and downstream scour holes versus time in Test T3. 0 10 20 30 40 50 60 70 -20-15-10-50 Collar elevation (mm) Be d el ev at io n (m m ) Abutment Collar edge

The scour became insignificant as the main channel edge of the collar was extended beyond the local scour hole area measured in the baseline case without countermeasures. The trailing edge of the collar should extend to a location downstream of the abutment. • Based on these experiments, the collar elevation should be 0.08ym below the original bed level and the collar width should be at least 0.23La, where La is the abutment length perpendicular to the flow direction. 9.6 Summary Scour at bridge abutments can cause damage or failure of bridges and result in excessive repairs, loss of accessibility, or even death. To mitigate abutment scour, both clear-water and live-bed laboratory experiments in a compound channel were performed using parallel walls and spur dikes. In addition, collars were also tested under clear-water conditions only. A series of experiments were performed in an open- channel flume with a compound channel for clear-water and live-bed scour conditions. Two types of parallel walls were tested: the first type was made of a wood plate, and the second was made of piled rocks. For solid parallel walls, a series of rectangular straight plates of different lengths were used attached to the upstream end of a wing-wall abutment parallel to the flow direction. The velocity of the flow for the three cases was 0.9, 1.5, or 2.3 times the incipient motion velocity for bed sediment move- ment. The bed material was sand with a mean diameter of 0.8 mm and a standard deviation of 1.37. All the plates were seated at the bottom of the compound channel bank slope and were even with the abutment face. It was found that straight plates thus situated caused the scour hole to be shifted away from the upstream corner of the abutment and to be effective as a countermeasure to prevent scour there. As the length of the plate increased, the scour at the abutment declined. It was found that a length of 1.6La, with La being the length of the abutment perpendicu- lar to the flow, caused the scour to be eliminated at the abut- ment for a velocity ratio (V/Vc) of 0.9 (clear-water scour). Similarly, a 1.6La-long wall can eliminate the time-averaged scour depth at the abutment 100 percent for a velocity ratio of 1.5 and 70 percent for a velocity ratio of 2.3. If the upstream end of the wall is anchored below the scour depth, this countermeasure can be feasible for situations where rock is expensive. For parallel rock walls, various values of wall length and protrusion length into the main channel were tested. It was found that a wall that does not protrude into the main chan- nel with a length of 0.5La minimizes scour at the abutment for all three different flow velocity ratios (0.9, 1.5, and 2.3). A series of configurations of spur dikes with varying lengths, spacings, number, and positions with respect to the abutment were tested. The most effective configuration to prevent local scour at the abutments consisted of three spur dikes composed of rock located upstream of the abutment and at the two corners. In addition, collars at the abutment were tested. It was found that these collars were able to protect the bridge abutment efficiently by eliminating secondary vortices that ordinarily would cause local scour. The minimum collar dimensions that eliminated local scour were a flow- perpendicular width of 0.23La (La is the abutment length per- pendicular to the flow direction) and a flow-parallel length of 0.7 times the flow-parallel abutment width. It was determined that a vertical location of 0.08ym (where ym is the main chan- nel flow depth) below the mean bed sediment elevation gave the best results of scour reduction. In addition, the collar not only reduced scour magnitude near the abutment, but also retarded the development of the scour hole. 190