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.
3.1 3. LABORATORY TESTING - DOCUMENTATION OF UNDERWATER FILTER INSTALLATION PROCEDURES 3.1 Overview A variety of means and methods are available for placing filter materials underwater, and in flowing water. The selection of any particular technique depends on the materials to be used and the flow conditions at the time of placement (e.g., flow depth and velocity. A previous NCHRP project, 24-07 (2) on pier scour protection demonstrated the viability of placing sand- filled geocontainers around a bridge pier in flowing water to achieve a filter prior to placing the armor layer (riprap in that case). The resulting research report, NCHRP Report 593, "Countermeasures to Protect Bridge Piers from Scour," (Lagasse et al. 2007) documents the successful placement using a hydraulic excavator with thumb to drop individual geocontainers from above the water surface. The geocontainers were filled with filter sand prior to placement and weighed 200 pounds each. Figures 3.1 (a) and (b) illustrate the placement technique used in that project. Refer to Figures 2.20 through 2.26 and related text for a more thorough discussion of that research effort. (a) 200-lb sand-filled geocontainer placed from above. NCHRP Project 24-07 (2). (b) Geocontainers partially covered by riprap armor around bridge pier. NCHRP Project 24-07 (2). Figure 3.1. Sand-filled geocontainers placed as a filter around a bridge pier. For the current study, alternative filter materials were considered, especially for cases where access for construction equipment is limited (beneath a bridge deck with little clearance above ordinary low water, for example). Certified divers to assist the placement of both granular and geotextile filter materials were used to demonstrate placement of filters in flowing water as a proof-of-concept demonstration. Given the wide range of capabilities of the large-scale outdoor River Engineering flume at CSU, the following filter materials were selected for diver- assisted placement around a prototype-scale pier in flowing water:
3.2 Granular filters: ⢠Pea gravel mixed equally with 0.6 mm masonry sand and placed with a 2-inch diameter rigid tremie pipe at various elevations above the bed; d50 = 4.0 mm. ⢠Nominal 3/8-inch (9 mm) pea gravel placed loosely around the pier using sand pump and 2-inch diameter flexible tremie hose; d50 = 6.0 mm. Composite filters: ⢠Empty geobags placed on bed adjacent to the pier and filled by divers under water with pea gravel using a sand pump and 2-inch diameter flexible tremie hose; d50 = 6.0 mm. Geotextile filters: ⢠Non-woven, needle-punched geotextile sheets (buoyant), unrolled in the direction of flow adjacent to and downstream of the pier. ⢠Double-sided drainage composite sheets (negatively buoyant), unrolled in the direction of flow adjacent to and upstream of the pier. 3.2 Laboratory Testing Plan 3.2.1 Preparation of Testing Facility Preparation of the test setup in the outdoor River Engineering Flume at CSU commenced in November 2016. The flume is 180 feet long, 20 feet wide, and 8 feet deep. In the floor of the flume, approximately halfway along its length, is a 30 foot long by 8 feet deep pit, referred to as the sediment recess. A prototype-scale bridge pier measuring 4 feet long by 1.5 feet wide was constructed of timber and plywood and bolted into place in the center of the sediment recess. The recess was filled with 0.8 mm uniform concrete masonry sand, which acted as riverbed sediment around the pier. Due to frozen conditions during the winter of 2016 â 2017, the installation was not completed until May 2017. Figures 3.2 (a) and (b) show the installation. A 20-foot wide moveable data collection carriage spans the width of the flume and traverses its length on rails. It was used to collect point velocity data during the tests, and while the divers were in the water, it was stationed directly above the pier to simulate a bridge deck with limited overhead clearance. The lower part of the data collection carriage can be seen in Figure 3.2 (b). 3.2.2 Testing Approach and Documentation On May 25 and 26, 2017, a certified dive team was used to place the filter materials around the bridge pier as a "proof-of-concept" demonstration to determine the advantages and limitations of various filter materials and underwater placement techniques. The flume was filled to a 4-foot depth of water, and two steady flow rates were established.
3.3 (a) Construction of lower section of pier within the sediment recess (looking downstream). (b) Completed installation (looking upstream). Note data collection carriage above pier. Figure 3.2. Installation of prototype-scale bridge pier in the River Engineering Flume at Colorado State University. The first flow rate was held steady at 50 ft3/s, resulting in a modest approach velocity of about 0.6 ft/s along the flume centerline. The second flow rate was established at 200 ft3/s, which resulted in a centerline approach velocity in excess of 3.5 ft/s upstream of the pier. The dive team consisted of two scuba divers and a topside line handler who managed the tethers. Underwater communications were established via comm lines in the tether ropes. Figures 3.3 (a) and (b) show the test setup with flowing water and divers in the flume. Both underwater and above-water video and still photos were taken before, during, and after each individual dive to document filter placement. After the dive team finished the filter placement, point velocity data were taken with a Marsh-McBirney 1-dimensional electromagnetic flow meter at three predetermined cross sections at both the 50 ft3/s and 200 ft3/s flow rates. Velocity data were taken at the bed, and at 20%, 60%, and 80% flow depths. Figures 3.4 (a) and (b) are contour maps of the point velocity data at 60% flow depth for 50 ft3/s and 200 ft3/s, respectively. The black dots in these figures indicate the locations of the point velocity measurements at longitudinal flume stations 12.6, 54.6, and 81.6. Figures 3.5 (a) through (c) provide contour maps of point velocity data across the 20-foot flume width at cross section stations 12.6, 54.6, and 81.6 for the 50 ft3/s tests. Figures 3.6 (a) through (c) provide the point velocity contour maps at the same stations for the 200 ft3/s tests.
3.4 (a) Flume test at 4-foot depth and 200 ft3/s, approach velocity â 3.5 ft/s. Note bow wave around nose of pier. (b) Tethered divers preparing to place geotextile filters during 200 ft3/s test at 4-foot depth. Figure 3.3. Prototype-scale bridge pier in the River Engineering Flume at Colorado State University. (a) Point velocities at 50 ft3/s flow rate. (b) Point velocities at 200 ft3/s flow rate. Figure 3.4. Plan view of flume showing point velocity contours at 60% depth. Distance along flume, ft Di st an ce a cr os s f lu m e, ft Di st an ce a cr os s f lu m e, ft Distance along flume, ft 200 cfs 60% depth 0 10 20 30 40 50 60 70 80 90 100 0 5 10 15 20
3.5 (a) Cross section 12.6 (40 feet upstream of pier). (b) Cross section 54.6 (adjacent to pier). (c) Cross section 81.6 (25 feet downstream of pier). Figure 3.5. Point velocity contours (ft/s) at 50 ft3/s. Distance across flume, ft De pt h, ft Distance across flume, ft De pt h, ft
3.6 (a) Cross section 12.6 (40 feet upstream of pier). (b) Cross section 54.6 (adjacent to pier). (c) Cross section 81.6 (25 feet downstream of pier). Figure 3.6. Point velocity contours (ft/s) at 200 ft3/s. De pt h, ft Distance across flume, ft
3.7 3.3 Underwater Installation of Granular Filters 3.3.1 Critical Velocity vs. Particle Size Placing a granular filter in flowing water requires knowledge of the flow velocity that would be expected at the job site during the construction season. Granular filters should not be used alone if the local velocity (e.g., at a pier) exceeds the critical velocity for particle movement. As a rule of thumb, local velocities at a pier are estimated to be 1.5 times the approach velocity at circular or round-nose piers, and 1.7 times the approach velocity for rectangular or blunt-nose piers. Critical velocity for a given particle size of granular filter material can be estimated from Arneson et al. (2012): 3/16/1 uc )d(yKV = (3.1) where: Vc = Critical velocity above which particles of size d50 or smaller will be transported, ft/s (m/s) y = Flow depth, ft (m) d = Particle size ft (m) (typically taken as the median size d50) Ku = Coefficient for system of units = 11.17 English units (ft-lb-s) = 6.19 SI units (m-kg-s) Equation 3.1 can be rearranged to calculate the critical particle size dc at the threshold of motion for a given velocity and depth. Figure 3.7 is a graph of critical velocity vs. particle size for granular filters having a specific gravity of 2.65 and ranging in size from d50 = 1 mm (coarse sand) to 100 mm (cobbles). The pea gravel used in the flume tests at the prototype-scale pier had a d50 of 6.0 mm and a critical velocity of 3.9 ft/s at a flow depth of 4 feet. Figure 3.7. Critical velocity vs. particle size (particle specific gravity = 2.65). 0 2 4 6 8 10 12 14 1 10 100 Cr iti ca l v el oc ity , f t/ s Particle size, mm 8 ft 20 ft Particle S.G. = 2.65 depth = 4 ft
3.8 3.3.2 Dispersion and Segregation Issues When placed in flowing water, granular filter materials will disperse and segregate unless placed directly on the bed. Therefore, it is important to utilize uniformly-sized filter gradations and not allow loose dumping from the water surface. To quantify the effect of dispersion while dropping granular filter material through the water column, a well-graded mixture of sand and pea gravel was placed at various heights above the bed in 4 feet of flowing water using a rigid 2-inch diameter tremie pipe. This was done in the large-scale, outdoor River Engineering Flume at Colorado State University for two different flow rates: 1) 50 ft3/s, average velocity of 0.6 ft/s and critical particle size dc = 0.02 mm 2) 80 ft3/s, average velocity of 1.0 ft/s and critical particle size dc = 0.11 mm Figures 3.8 (a) and (b) show the test setup. (a) Procedure for granular filter dispersion tests using 5-gallon bucket and 2-inch diameter tremie. (b) Piles of filter material after dispersion tests. Upper row: Test at 80 ft3/s (V = 1.0 ft/s) Lower row: Test at 50 ft3/s (V = 0.6 ft/s) Figure 3.8. Granular filter dispersion tests. For each flow rate, a 5-gallon bucket of granular filter material was emptied into a funnel connected to the top of the tremie pipe. The lower end of the tremie was positioned at the following locations in the flow: 1. At the water surface (4.0 ft. above the bed) 2. At 80% flow depth (3.2 ft. above the bed) 3. At 60% flow depth (2.4 ft. above the bed) 4. At 20% flow depth (0.8 ft. above the bed) 5. Directly on the bed This resulted in the filter material falling through the water column and landing on the bed in five discrete piles for each of the two flow rates, as seen in Figure 3.8 (b). Samples from each pile were taken and analyzed for particle size gradation and compared to a sample from the stockpile. The grain size distribution curves for the tests at 50 ft3/s and 80 ft3/s are provided in Figures 3.9 and 3.10, respectively.
3.9 Figure 3.9. Grain size distribution curves for dispersion test at 50 ft3/s. (approach velocity 0.6 ft/s) Figure 3.10. Grain size distribution curves for dispersion test at 80 ft3/s. (approach velocity 1.0 ft/s) The gradation curves clearly show the effect of dispersion as a function of both velocity and height of the tremie pipe above the bed. Table 3.1 provides information on the resulting median particle size, uniformity coefficient, and geometric standard deviation of particle size of the filter material after placement. Placing the tremie pipe on or near the bed resulted in less dispersion compared to locating the end of the pipe higher up in the water column. In addition, Test 1 at a lower flow velocity resulted in less dispersion than Test 2. 3-in 2-in. 1-in. 0.5-in #4 #8 #10 #16 #20 #40 #70 #100 #140 #200 0 10 20 30 40 50 60 70 80 90 100 0.010.1110100 Pe rc en t F in er b y w ei gh t - % Grain Size in millimeters GRAIN SIZE DISTRIBUTION at 50 cfs (Approach velocity 0.6 ft/s) Stockpile Water Surface (4.0 ft) 80% Flow Depth (3.2 ft) 60% Flow Depth (2.4 ft) 20% Flow Depth (0.8 ft) Bed (0.0 ft) Sieve Size GRAVEL Coarse Fine SAND Coarse Medium Fine SILT 3-in. 2-in. 1-in. 0.5-in #4 #8 #10 #16 #20 #40 #70 #100 #140 #200 0 10 20 30 40 50 60 70 80 90 100 0.010.1110100 Pe rc en t F in er b y w ei gh t - % Grain Size in millimeters GRAIN SIZE DISTRIBUTION at 80 cfs (Approach velocity 1.0 ft/s) Stockpile Water Surface (4.0 ft) 80% Flow Depth (3.2 ft) 60% Flow Depth (2.4 ft) 20% Flow Depth (0.8 ft) Bed (0.0 ft) Sieve Size GRAVEL Coarse Fine SAND Coarse Medium Fine SILT
3.10 Table 3.1. Results of Granular Filter Dispersion Tests. Location d50, mm Uniformity coefficient Cu = d60/d10 Geometric standard deviation Ïg = (d84/d16)1/2 Stockpile 4.0 (5.3)/(0.31) = 17.1 [(7.6)/(0.44)]1/2 = 4.2 Dispersion Test 1: 50 ft3/s (approach velocity 0.6 ft/s, critical particle size 0.02 mm) Bed 3.6 (5.1)/(0.39) = 13.1 [(7.6)/(0.51)]1/2 = 3.9 20% flow depth (0.8 ft) 2.6 (5.0)/(0.38) = 13.2 [(7.6)/(0.55)]1/2 = 3.7 60% flow depth (2.4 ft) 4.7 (5.7)/(0.60) = 9.5 [(7.6)/(0.91)]1/2 = 2.9 80% flow depth (3.2 ft) 6.2 (6.9)/(0.92) = 7.5 [(8.1)/(3.4)]1/2 = 1.5 Water surface (4.0 ft) 5.4 (6.1)/(1.8) = 3.4 [(8.0)/(1.4)]1/2 = 2.4 Dispersion Test 2: 80 ft3/s (approach velocity 1.0 ft/s, critical particle size 0.11 mm) Bed 5.0 (5.9)/(0.46) = 12.8 [(7.7)/(0.73)]1/2 = 3.2 20% flow depth (0.8 ft) 5.7 (6.2)/(0.98) = 6.3 [(7.7)/(1.5)]1/2 = 2.3 60% flow depth (2.4 ft) 6.0 (6.7)/(1.3) = 5.2 [(7.8)/(2.0)]1/2 = 2.0 80% flow depth (3.2 ft) 6.3 (6.9)/(2.5) = 2.8 [(8.0)/(4.0)]1/2 = 1.4 Water surface (4.0 ft) 6.3 (6.9)/(2.0) = 3.5 [(8.0)/(3.9)]1/2 = 1.4 The values in Table 3.1 clearly show that the granular material released higher off the bed results in a filter placement that is increasingly coarser and more uniform compared to the original stockpile. 3.3.3 Flexible Tremie-Type Approaches Placing a granular filter under water is relatively easy to accomplish with a standard solids handling pump (also known as a trash pump) and a flexible tremie pipe or hose. In this way, even limited access or clearance beneath a bridge deck is no impediment to granular filter placement. The solids handling pump can be used to suction up a slurry of gravel and water with up to about 40% solids by volume, and deliver it to divers working under water at the site of filter placement. Figures 3.11 (a) and (b) illustrate the setup at CSU's River Engineering Flume in May 2017. In Figure (a) below, the solids handling pump and 40-hp engine is shown. The pump is capable of passing solids up to 3 inches in diameter. The pump has 4-inch diameter suction and discharge capacity; note we reduced both suction and discharge hoses to 2-inch diameter for ease of handling by the divers. The wooden bin containing the pea gravel and suction hose can be seen in the foreground in Figure 3.11 (b). We also tested a composite filter consisting of empty geobags placed on the bed adjacent to the pier and filled by divers under water with pea gravel using a sand pump and 2-inch diameter flexible tremie hose. Figures 3.12 (a) and (b) show the geobag with fill port and divers filling it under water, respectively.
3.11 (a) Solids handling pump. (b) Submerged bin with pea gravel (looking upstream). Note 2-inch diameter suction hose. Figure 3.11. Test setup for granular filter placement. (a) Nonwoven needle-punched geobag with fill port. (b) Filling geobag with pea gravel under water. Note 2-inch diameter flexible tremie hose. Figure 3.12. (a) Geobag with fill port; (b) Divers filling geobag under water at the pier. Figures 3.13 (a) and (b) show the pea gravel delivered to the pier during and after the 50 ft3/s run. For this demonstration, the tremie hose approach was successful in placing the material adjacent to the pier, as well as for underwater filling of 3-foot by 3-foot square geobags having an integral fill port. During the 200 ft3/s demonstrations, the loose pea gravel could not be uniformly placed due to the turbulence around the pier; however, the divers were able to fill the geobags successfully, even at the higher flow rate. Figures 3.14 (a) and (b) show the placement after the 200 ft3/s run.
3.12 (a) Underwater photo of diver placing loose pea gravel around pier at 50 ft3/s flow rate. Note gravel exiting the tremie hose. (b) After 50 ft3/s test, showing loose pea gravel and gravel-filled geobags, filled in flowing water. Figure 3.13. Tremie-placed granular filter during and after the 50 ft3/s demonstration. (a) Loose pea gravel and geobags after the 200 ft3/s test. (b) After 200 ft3/s test. Note dune bedforms downstream of pier. Figure 3.14. Tremie-placed granular filter after the 200 ft3/s demonstration. 3.4 Underwater Installation of Geotextile Filters Both buoyant and self-sinking panels of nonwoven needle-punched geotextiles were placed around the pier, first at a flow rate of 50 ft3/s and again at a flow rate of 200 ft3/s. A standard 8-ounce per square yard polypropylene geotextile (Mirafi 180N) was used as the buoyant fabric. A 6 mm thick double-sided drainage composite (CETCO Aquadrain G25) with High Density Polyethylene (HDPE) geonet core was used as the self-sinking fabric; it was made negatively buoyant by affixing metal strips at 4-foot intervals along its length to achieve a handling capability comparable to the sandmat commonly used in Europe (see Figures 2.9 and 2.10). Figure 3.15 is a close-up photo of the self-sinking fabric showing the drainage net core with geotextile on top and bottom. All geotextile panels measured 4 feet wide by 16 feet in length, and were placed around the pier in a four-foot depth of water. Satisfactory placement of both materials was accomplished during the 50 ft3/s flow rate as shown in Figure 3.16 (a) and (b). One gallon sand-filled Ziploc bags and 9" riprap stones were used to temporarily weigh down the fabrics while they were being unrolled at the pier in the direction of flow.
3.13 Figure 3.15. Self-sinking geotextile filter. (a) Rolls of geotextile filters. (b) After 50 ft 3/s run; satisfactory placement. (looking upstream) Figure 3.16. Geotextile filter placement after the 50 ft3/s demonstration. During the 200 ft3/s test, the turbulence around the pier created significant difficulty for placing the buoyant geotextile. The divers reported that the geotextile was "sailing" and "twisting" in the current, and that the rocks and sandbags could not effectively hold it down. The divers used heavier pieces of concrete rubble to weigh the fabric down. The installation of the buoyant fabric was unsatisfactory during the 200 ft3/s test. The divers reported that the self-sinking drainage composite was easier to handle because it was stiffer and was easier to unroll and control, although the resulting placement was still not ideal. Figures 3.17 (a) and (b) show the results of the geotextile placement after the 200 ft3/s test. 3.5 Appraisal of Underwater Installation Testing Results Placement of granular and geotextile filters at a prototype-scale bridge pier was successful during all tests conducted at 50 ft3/s, with an approach velocity of about 0.6 ft/s. The tests at 200 ft3/s were conducted with a centerline approach velocity somewhat in excess of 3.5 ft/s. The tests at the higher flow rate encountered difficulties placing sheets of geotextile fabric in flowing water.
3.14 (a) Unsatisfactory placement of the buoyant geotextile during the 200 ft3/s test run. (Flow from right to left) (b) Self-sinking drainage composite after the 200 ft3/s test run. (Flow from right to left) Figure 3.17. Geotextile filter placement after the 200 ft3/s demonstration. Placing granular filter material (in this case, pea gravel) using a standard solids-handling pump and flexible tremie hose proved successful at both flow rates. Underwater filling of standard 3- by 3-foot square geobags with integral fill port, also known as dewatering bags, with the pea gravel material was found to be rapid and efficient, even at the higher flow rate. The geobags were unrolled in the downstream direction with the fill port on the upstream side. One diver held the bag in place against the pier while the other filled it using the flexible tremie hose. The divers noted that with this approach, "the bags almost filled themselves," even during the high-flow test. Given the rate of gravel delivery from the solids-handling pump, each bag could be filled in about 3 minutes. The practical limit for buoyant geotextile placement by divers was exceeded at an approach velocity of approximately 3.5 ft/s. Placing a self-sinking fabric was barely manageable at this approach velocity. Based on the divers' first-hand reports, recommended limits on approach velocity for placing buoyant vs. self-sinking geotextiles are 2.5 ft/s and 3.5 ft/s, respectively. Dispersion of granular filter materials was investigated using a rigid 2-inch diameter tremie pipe with its bottom end placed at different heights above the bed in four feet of flowing water. At each height, a five-gallon bucket of well-graded granular filter material was injected into the flow via the tremie. After slowly draining the flume, the resulting piles of filter material were analyzed for their grain size distribution. In general, the filter material landing on the bed became increasingly coarser and more uniform with increasing height of placement, with the finer particles being winnowed out and swept downstream as the material fell through the water column. The dispersion tests were conducted with two velocities: 0.6 and 1.0 feet per second. The dispersion effect was significantly more pronounced at the higher flow velocity, even though the grain size distribution of the original stockpiled filter material indicated that less than 2 percent of the particles were smaller than the critical size for initiation of motion at this velocity. Note that safety precautions were taken during all phases of the installation demonstration tests. Divers were tethered at all times, and maintained constant communication with the topside line handler. Also, care was taken to ensure that the divers were never located directly downstream of the filter materials during placement, so that potential tangling or entrapment by loose geosynthetic fabrics caught in the turbulence was minimized.
