Click for next page ( 12


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 11
11 CHAPTER 2 FINDINGS FINDINGS FROM LITERATURE STUDY GRS bridge-supporting structures with flexible facings have been the subject of many studies and recently have seen Over the past two decades, GRS has been used in the con- some actual applications, in the United States and abroad. This struction of various earth structures, including retaining study synthesizes the measured behavior and experiences walls, embankments, slopes, and shallow foundations. In gained from case histories of flexible facing GRS bridge- actual construction, GRS structures have demonstrated many supporting structures from around the world. Observations distinct advantages over their conventional counterparts. were made in relation to performance, design, and construc- GRS structures typically are more ductile, more flexible tion of flexible facing GRS bridge-supporting structures. The (hence more tolerant to differential settlement), more adapt- case histories were organized into two groups: in-service struc- able to low quality backfill, easier to construct, and more tures and field experiments. Most of these studies were on economical. They also require less overexcavation. bridge abutments, with a few on bridge piers. The design and In recent years, applications of the GRS technology to construction of GRS bridge abutments are similar in principle bridge-supporting structures have gained increasing atten- to GRS walls, except the former typically are subject to a tion. The facing of GRS bridge-supporting structures can be rather high surface load close to the wall face. Also, some U.S. grouped into two types: rigid and flexible. A rigid facing is a states do not permit the use of segmental concrete facing in continuous reinforced concrete facing, either precast or cast- GRS bridge-supporting structures because of concerns with in-place. A flexible facing, on the other hand, typically takes the durability of masonry units when exposed to chemical the form of wrapped geosynthetic sheets, dry-stacked con- agents such as de-icing fluids. Based on the measured perfor- crete modular blocks, natural rocks, or gabions. In contrast mance of the case histories, observations were made in rela- to a flexible facing, a rigid facing offers a certain degree of tion to performance, design, and construction of GRS bridge- global bending resistance along the entire height of the fac- supporting structures. Some of the material properties and the ing, thus offering greater constraint to lateral earth pressure- methods for determining the properties are not reported induced "global" bending deformation. because they are not available in the source materials. Since 1994, the Japan Railway has constructed many full- height facing GRS bridge abutments and piers (e.g., Tateyama et al., 1994; Kanazawa et al., 1994; Tatsuoka et al., 1997) using In-Service Bridge-Supporting Structures a rigid facing GRS wall system developed by Tatsuoka and his The construction-related information and measured perfor- associates at the University of Tokyo. These GRS bridge- mance of six in-service GRS bridge abutments are described supporting structures have been constructed in two stages. The below. The six abutments are the Vienna railroad embank- first stage involves constructing a wrapped-faced GRS wall with ment in Austria (Mannsbart and Kropik, 1996), the New the aid of gabions, and the second stage involves casting in place South Wales GRS bridge abutments in Australia (Won et al., a full-height reinforced concrete facing over the wrapped face. 1996), the Black Hawk bridge abutments in Colorado (Wu et Field measurement has shown that these structures experienced al., 2001), the Founders/Meadows bridge abutments in Col- little deformation under service loads and have performed far orado (Abu-Hejleh et al., 2000), the Feather Falls Trail bridge better than conventional reinforced concrete retaining walls and abutments in California (Keller and Devin, 2003), and the abutments in the 1995 Japan Great Hansin earthquake that mea- Alaska bridge abutments in Alaska (Keller and Devin, 2003). sured 7.2 on the Richter scale (Tatsuoka et al., 1997). Most recently, Tatsuoka and his associates developed a preload- prestress method for improved performance of the GRS bridge- Case A1: Vienna Railroad Embankment, Austria supporting structures (Tatsuoka et al., 1997; Uchimura et al., (Mannsbart and Kropik, 1996) 1998). Despite their success, the rigid facing GRS bridge- supporting structures have found applications only in Japan, A temporary GRS embankment was constructed in mostly because of their higher cost and longer construction time Vienna, Austria, to support a railroad track. The railroad compared with GRS walls with flexible facings. embankment had a height of 2.1 m and a slope inclination of

OCR for page 11
12 63 deg from the horizontal. A needle-punched nonwoven fiberglass dowels were used to interlock block layers verti- geotextile was used as the reinforcement. The geotextile had cally. Foundation conditions at the site consisted of a 1- to a tensile strength of 23 kN/m with elongation at break of 45 3-m-thick layer of loose silty sand containing thin discontin- percent. The reinforcement spacing and length were 0.3 m uous silty clay layers overlying a medium dense silty sand and 1.7 m, respectively. The backfill was a compacted grav- layer varying in thickness from 7 m to 10 m. Sandstone elly sand. Its placement unit weight was 21 kN/m3, and the bedrock was present at 13 m depth. design internal friction angle was 35 deg. The two abutments are referred to as Abutment A and Abut- The individual layers of the structure were built using a ment B. Abutment A consisted of three terraced segmental removable formwork consisting of steel angles and wooden block walls with 12 layers of a Tensar HDPE geogrid, SR 110, bars. To get adequate friction between the adjacent geotextile beneath the sill beam. The tensile strength of the geogrid was layers, a thin layer of sandy gravel was placed on each lift 110 kN/m at 11.2 percent strain. Total tiered height was 6.5 m. before the installation of the next layer. Given that the struc- Abutment B consisted of four terraced segmental block walls ture had to fulfill only a temporary function, a wrapped-around with 17 layers of SR110 geogrid beneath the sill beam. Total wall face was used and the surface protection was omitted. tiered height was 9.5 m. To account for creep, temperature vari- Above the reinforced structure, a 0.9-m-high unreinforced ation, and construction damage, the allowable long-term design embankment with a slope of 45 deg was built as a buffer for strength for the SR110 geogrid was taken as 27 kN/m. The ver- the traffic. The design traffic load was 60 kPa, exerted at 1.45 tical spacing of geogrid layers was 40 cm or 60 cm. The maxi- m from the top edge of the unreinforced embankment. The mum reinforcement length was 15 m. The backfill material, a cross-section of the temporary embankment is shown in fine sand, was compacted to at least 95 percent Standard Rela- Figure 2-1. Weekly settlement measurement was carried out tive Density to have a design friction angle of 32 deg. Addi- on 6 points along the 100-m-long embankment. The results tional layers of geogrid, 5 m long with a wrap-around face, indicated that under traffic load, the measured settlement was were used to reduce active earth pressure behind the sill beam. nil at four of the six points, and at the other two points the The unreinforced concrete sill beam was 20 cm thick and 2.5 settlement was less than 1 mm. m wide. It was set back 2.5 m from the edge of the top wall to reduce the effects of horizontal pressure because of sill beam load distribution through the reinforced soil. In view of the Case A2: New South Wales GRS Bridge loose nature of the foundation soil, the top 1 m was excavated Abutments, Australia (Won et al., 1996) and compacted in the vicinity of the lowest-tiered wall. A comprehensive monitoring program was implemented to Geogrid reinforced bridge abutments with a segmental evaluate the performance of Abutment B. Sill beam loading block facing were constructed to support end spans directly occurred during January 1994. The maximum reinforcement for a major bridge in New South Wales, Australia, in 1994. tension at Level 1 approached 33 kN/m and occurred toward The bridge consisted of a nine-span superstructure over the the back of the reinforced soil block. The maximum rein- Tweed River. The abutments were up to 10 m high, con- forcement tension at Level 2 was 21 kN/m and occurred structed in a terraced arrangement, as shown in Figure 2-2. toward the back of the reinforced soil block. At Level 3 rein- The facing comprised "Keystone" segmental concrete forcement, the effect of sill beam loading was evident with a blocks that were partially voided internally, and aggregates maximum reinforcement tension of 22 kN/m occurring under were used to fill the block during construction. High-strength the sill beam region. The maximum strain in the geogrid was 1.6 percent, occurring at Level 1. The maximum settlement p = 60 kN/m 2 was 80 mm. Lateral movements of the reinforced soil structure deduced from wall survey and inclinometers I1 and I2 (see 0.9 m 1.45 m 2.6 m Figure 2-2) were 10 mm up to the completion of the abutment and 26 mm post construction movements for the lowest-tiered 0.9 m Sleeper wall. Subsequent site investigations of the loose upper silty sand layer indicated the presence of thin discontinuous seams of medium stiff silty clay, which could have contributed to the deformation response at the base of the structure. 2.1 m 0.3 m Case A3: Black Hawk Bridge Abutments, 63 6% Colorado (Wu et al., 2001) 1.7 m Two rock-faced GRS abutments were constructed to sup- Figure 2-1. Cross-section of the Vienna railroad port the Bobtail Road Bridge, a 36-m-span steel arched embankment, Austria (Mannsbart and Kropik, 1996). bridge in Black Hawk, Colorado (see Figures 2-3 and 2-4).

OCR for page 11
13 Bridge Sill Beam Tensar SR80 2.5 m Inclinometer I1 5m 5m 2m Geogrid Spacing Vertical Borehole 11 m 0.6 m Inclinometer I1, I2 2m Level 3 11 m 2m Level 2 "Keystone" Blocks 13 m Inclinometer I2 2m Level 1 Geogrid Spacing 7m 0.4 m 15 m Tensar SR110 Compacted Sand Fill Figure 2-2. Cross-section of the New South Wales GRS bridge abutments, Australia (Won et al., 1996). Figure 2-3. Cross-section of the Black Hawk bridge abutments (Wu et al., 2001).

OCR for page 11
14 Figure 2-4. Footings and foundations of the Black Hawk bridge abutments (Wu et al., 2001). Each GRS abutment comprised a two-tier GRS mass with ground portion of the abutment had different heights, vary- two square footings on the lower tier and a strip footing on ing from 1.0 m to 2.7 m for the West abutment; and from the upper tier. The square footings on the West abutment are 1.0 m to 5.4 m for the East abutment. The thickness of the referred to as Footings #1 and #4, and the square footings on upper tier reinforced soil mass was 1.8 m for both abutments. the East abutment are referred to as Footings #2 and #3. The The upper tier reinforced soil mass was built to support the GRS bridge abutments were constructed on a stiff soil. strip footing and the approach ramp. The thicknesses of the lower tier reinforced soil mass under The abutments were constructed with the onsite soil, clas- Footings #1 and #4 were, respectively, 4.5 m and 1.5 m; and sified as SM-SC per ASTM D2487, and reinforced with lay- 7.5 m and 1.5 m under Footings #2 and #3, respectively. The ers of a woven geotextile at vertical spacing of 0.3 m. The lower part of the GRS abutment was embedded in the polypropylene woven geotextile (Amoco 2044) had a wide- ground, while the upper part was above ground. Only the part width tensile strength of 70 kN/m in both machine and cross- above ground was constructed with rock facing. The above machine directions at 18 percent strain, per ASTM D4595

OCR for page 11
15 (the wide-width strip method). The backfill had 12 percent The maximum strain mobilized in the reinforcement of fines (passing sieve No. 200). The backfill material was was very small (less than 0.2 percent at 80 kPa). compacted to 91 percent relative compaction per AASHTO Preloading reduced creep strains in the reinforced struc- T-99 (the moisture-density relation of soil was determined by ture and the geotextile reinforcement. using a 2.5 kg rammer with a 305 mm drop), having a dry unit weight of 15.8 kN/m3 at a water content of 12.2 percent. The measured friction angle and cohesion, as determined Case A4: Founders/Meadows Bridge Abutments, from the CD triaxial compression tests, were 31 deg and Colorado (Abu-Hejleh et al., 2000) 34 kPa, respectively. For each square footing, a vertical pressure of 245 kPa A replacement bridge was constructed over Interstate (1.6 times the design load of 150 kPa) was applied and sus- Highway 25 at Founders/Meadows Parkways near Castle tained for 100 minutes, then unloaded to zero. Three loading- Rock, Colorado, in 1999. In this bridge abutment, both unloading cycles were applied following the first loading- the bridge and the approaching roadway were supported unloading cycle. In the reloading cycles, the typical applied by a system of GRS segmental retaining walls. The front pressure was the design load (150 kPa). For the strip footing, GRS wall supports the bridge superstructure, which extents the vertical load was increased incrementally to 80 kPa around a 90-deg curve into a lower GRS wall supporting the (2 times the design load of 40 kPa), sustained for 120 minutes, wing wall and a second tier, the upper GRS wall. The GRS and then unloaded to zero. The vertical load applied in the abutment was constructed on the native claystone or sand- reloading cycle was 40 kPa (the design load). The load was stone bedrock. The plan view of the structure is shown in Fig- maintained for 120 minutes before unloading. At the design ure 2-5. Each span of the bridge was 34.5 m long and 34.5 m load of 150 kPa in the preloading cycle, the average settle- wide. The design of the abutment followed the AASHTO ments were 13.3 mm, 6.4 mm, 28 mm, and 4.9 mm for Foot- (1997) guidelines. ings #1 through #4, respectively. At 150 kPa in the first Figure 2-6 shows the typical cross-section of the abutment. reloading cycle, the average settlements were reduced to For the reinforced soil zone behind and below the bridge abut- 2.5 mm, 3.8 mm, 4.5 mm, and 3.3 mm for Footings #1 through ment, a trapezoid-shaped reinforcement was adopted, in #4 respectively. Further reduction in the settlement was neg- which reinforcement increased linearly from 8.0 m at the bot- ligible in the subsequent reloading cycle. Preloading reduced tom with 1H:1V slope toward the top. The reinforcement the maximum lateral movement at 150 kPa loading pressure length for the abutment wall was 11 m to 13 m. The center- from 1.5 mm to 0.6 mm in Footing #1, and from 13.2 mm to line of the bridge abutment wall and edge of the foundation 4.5 mm in Footing #3. were 3.1 m and 1.35 m from the front of the wall face. Dry- In the preloading cycle, under a load of 245 kPa sustained stacked hollow-cored concrete blocks were used as the fac- for 60 minutes, the vertical creep displacements of Footings ing. The lower wall had a maximum height of 4.5 m to 5.9 m #1 to #4 were, respectively, 6.7 mm, 4.0 mm, 7.2 mm, and and the upper wall had a maximum height of 3.0 m for the 2.1 mm. In the reloading cycle, under the sustained load of West abutment and 3.2 m for the East abutment. The lower 150 kPa, the vertical and lateral creep deformations were wall had a minimum embedment of 0.45 m. The abutment insignificant. was constructed in two phases to accommodate traffic needs. At 80 kPa in the preloading cycle, the maximum strains Three grades of geogrid reinforcement were used: UX6 in layers A, B, and C were 0.18 percent, 0.04 percent, and with an ultimate strength of 157.3 kN/m used below the foun- 0.06 percent, respectively. At a sustained load of 80 kPa in dation, UX3 and UX2 with ultimate strengths of 64.2 kN/m the preloading cycle, the creep strains in layers A, B, and C and 39.3 kN/m, respectively, per ASTM D4595, used behind were 0.032 percent, 0.009 percent, and 0.003 percent, respec- the abutment wall. The ultimate strength of the geogrids was tively. Locations of layers A, B, and C are shown in Figure measured in accordance with the ASTM D4595 test method. 2-3. The creep strains were negligible at the sustained load The reinforcement spacing was 0.4 m. The backfill soil was of 40 kPa in the reloading cycle. a mixture of gravel (35 percent), sand (54 percent) and fines Based on the measured data, the following findings and (11 percent). The average unit weight and dry unit weight of conclusions were made: the compacted fill were 22.1 kN/m3 and 21 kN/m3 (95 per- cent of AASHTO T-180, the moisture-density relation being By preloading the reinforced soil mass to 245kPa, the determined by using a 4.54 kg rammer with a 457 mm drop), settlement at the design load of 150 kPa was reduced by respectively. The average placement moisture content was a factor of 1.5 to 6 for the four square footings. 5.6 percent. Preloading also reduced the lateral movement of the Field monitoring was performed with various instruments GRS abutments. The lateral movement was reduced by during and after the construction of the structure. The mea- a factor of 2.5 to 3 at 150 kPa. sured vertical stresses did not differ significantly from the After the first reloading cycles, there was no significant static states calculated as z = z + q + z, where q is the reduction of lateral and vertical displacements of GRS uniform surcharge and z is the increase in vertical stress abutments in the subsequent reloading cycles. caused by concentrated surcharge loads assuming 2V:1H

OCR for page 11
16 Figure 2-5. Plan view of the Founders/Meadows bridge-supporting structure (Abu-Hejleh et al., 2000). pressure distribution. The horizontal stresses measured on Eighteen months after opening to traffic, the maximum the facing at the end of construction, however, were much outward displacement of the front wall facing and the smaller than the Rankine active earth pressures. The mea- maximum settlement of the bridge abutment footing sured geogrid strains at the end of construction were very were 13 mm (0.22 percent of wall height) and 11 mm low, on the order of 0.1 percent. (0.18 percent of wall height), respectively. The maxi- The measured outward movement of the GRS wall face mum outward displacement of the front wall facing was also very small. The maximum outward movement occurred at the elevation directly below the bridge sill. experienced along Section 400 during the construction of the Movement of the leveling pad (located at the base of the front GRS wall up to the bridge foundation elevation was GRS structure) was negligible, and the outward wall dis- about 9 mm. The maximum outward movements experi- placement tended to decrease toward the leveling pad. enced during placement of the bridge superstructure were on Both the rates of wall movements and the strain of the order of 7 mm to 9 mm. The field measurements also geogrid reinforcements decreased with time. indicated the sill settled about 13 mm because of the loads of Outward wall displacement as inferred by the integra- the bridge and the approaching roadway structures. Along tion of the strain distribution curve with respect to the Section 400 (see Figure 2-5), the leveling pad settled verti- reinforcement length matched closely with that deter- cally almost 5 mm during the construction of the front GRS mined from surveying. This implies that little slippage wall up to the bridge foundation elevation and settled another between the soil and reinforcement had occurred. 6 mm when the bridge and approaching roadway structures Probable causes for post-construction movements were were placed. traffic load, deformation under sustained load (creep), Post-construction performance of the Founders/Meadows and seasonal variation. bridge abutment was evaluated by Abu-Hejleh et al. (2002), The GRS bridge abutment shows no sign of the "bridge with the following findings: bump" problem. The Founders/Meadows GRS bridge

OCR for page 11
17 Width of the Reinforced Soil Zone, 11 m for Section 200, Abutment Wall (0.76 m wide) 12.97 m for Sections 400 and 800 Bridge Deck (0.13 m high) Approach Slab (3.72 m x 0.3 m) Roadway (0.35 m high) Girder (0.89 m high) 0.4 m high UX2 2m Sleeper Foundation UX3 Membrane & Collector Pipe UX3 Slope paving Cap Unit (0.1 m high) 75 mm Expanded Polystyrene UX3 Geogrid 1.35 m UX6 Geogrid Foundation (3.81 m x 0.61 m) 1.755 m 2.055 m 0.4 m 29 Rows for Section 400, 800 (5.9 m high) 22 Rows for Section 200 (4.5 m high) 0.3 m limit of 19 mm max. size crushed stone UX6 Geogrid Front MSE Wall CDOT Class 1 Backfill Drainage Blanket with Pipe Drains Connector The geogrid reinforcement length Block Unit (0.2 m high) increases linearly from 8 m at the bottom with one to one slope toward the top Geogrid 1st layer Embedment Length is 8 m Embedment 0.45 m Min. 7.8 m Leveling Pad (0.15 m high) Bedrock Figure 2-6. Typical cross-section of front and abutment walls, the Founders/Meadows bridge abutments (Abu-Hejleh et al., 2000). abutment has exhibited excellent short- and long-term The Colorado DOT provided the following guidelines for performance characteristics. design and construction of GRS abutments: Abu-Hejleh et al. (2003) also indicated that the rate of 1. The foundation soil for these abutments should be firm creep reinforcement strain under service load decreased with enough to limit the post-construction settlement of the time: a maximum increase of strain of 0.09 percent during the bridge sill to 75 mm. first year, a maximum increase of strain of 0.04 percent dur- 2. The designer should plan for a bridge sill settlement of ing the second year, and a maximum increase of strain of at least 25 mm caused by the bridge superstructure 0.02 percent during the third year in service. The largest rein- loads. forcement strain occurred directly beneath the bridge sill. 3. The maximum tension line needed in the internal sta- The maximum reinforcement strain after about 33 months in bility analysis should be assumed bilinear, starting at service was 0.27 percent. the toe of the wall and extending through a straight line The Colorado DOT concluded that the general layout and to the back edge of the bridge sill at the mid height of design of future GRS abutments should follow those in the the wall, and from there extending vertically to the Founders/Meadows abutment. The GRS abutments work back edge of the bridge sill. well for multiple span bridges, have the potential for elimi- 4. Ideally, construction should take place during the warm nating the "bump at the bridge" problem, avoid disadvan- and dry seasons. tages associated with the use of deep foundations, and allow 5. The backfill behind the abutment wall should be placed for construction in stages and within a smaller working area. before the girders.

OCR for page 11
18 Case A5: Feather Falls Trail Bridge Abutments, abutment, and each abutment had an embedment depth of California (Keller and Devin, 2003) 0.6 m to offer scour protection against a possible debris slide in the drainage. A 12-m-long trail bridge was constructed in 1999 on the The entire construction of the bridge took about 2 weeks Feather Falls Trail in the Plumas National Forest in northern with a crew of two people. The GRS abutments have per- California. Because the project site was remote and without formed well since the bridge was put in service. road access, the bridge materials had to be flown in with a heli- copter. Because of the deeply incised and narrow channel, the abutments were placed well above the channel high-water Case A6: Alaska Bridge Abutments, Alaska level. GRS abutments were selected for this project because (Keller and Devin, 2003) they use small, lightweight materials and are easy to construct. Two GRS abutments, constructed in 1992, support a Figure 2-7 shows the cross-section of the GRS abutments. 15.1-m-long precast, double-tee concrete bridge in the The two abutments were 1.5 and 2.4 m high, and the wall Tongass National Forest in southeast Alaska. Because trans- facing comprised 0.15 m by 0.15 m treated timbers. Two portation and construction costs are high in this area, the polyester woven geotextiles of different strengths were used bridge and abutment designs had to be economical and easy for the reinforcement. The top four layers of the reinforce- to construct, without the need for specialized equipment. ments had an ultimate strength of 70 kN/m, while the remain- Because the bridge is located in the tidal-influence zone, ing reinforcements had an ultimate strength of 52 kN/m, per there were concerns about corrosion loss, so GRS abutments ASTM D4595. The vertical reinforcement spacing was 0.15 m, were selected over hot-dipped galvanized welded-wire walls, and the average reinforcement length was 2.0 m. Most of the which commonly had been adopted in the area. The GRS reinforcements were sandwiched and nailed between the fac- abutments were 3.7 m high and had three vertical faces: a ing timbers, but the top four reinforcements were wrapped front wall paralleled to the stream alignment and two wing around the outside of the facing timbers and covered with walls oriented at 90 deg and 77 deg relative to the front-face timber boards to ensure maximum connection strength and wall. The distance between the front wall face to the toe of to protect the geotextiles. the sill was 0.9 m, and the distance between the centerline of Onsite rocky soil was used as the backfill and was com- the bearing of the bridge to the front wall face was 1.5 m. The pacted to 95 percent of its maximum dry density per AASHTO combination of dead and live design loads caused by bridge T-99. A geocomposite drain was placed behind each GRS superstructure was limited to 240 kPa. Figure 2-7. Cross-section of the Feather Falls Trail bridge abutments, California (Keller and Devin, 2003).

OCR for page 11
19 HDPE geogrids were used as the reinforcements. The base- Case B1: Garden Experimental Embankment, to-height ratio for the front facing wall was 1:1 and was 0.7:1 France (Gotteland et al., 1997) for the two adjoining wing walls. The vertical spacing for the geogrid was 0.3 m near the base of the wall and 0.15 m near A full-scale experiment was conducted in 1994 to investi- the top of the wall. Geogrids were wrapped around the timber gate the failure behavior of GRS structures as bridge abut- facing, and 19 mm rebar drift pins were driven into pre-bored ments, referred to as the "Garden" program (Geotextile: holes to hold the timber facing together. The full height of the Application en Reinforcement, experimentation et Normali- wall was covered with 50-mm-thick treated timber board to sation). A 4.35-m-high embankment was constructed for the protect the geogrids against UV degradation and floating experiment, as shown in Figure 2-8. The test embankment debris. A free-draining granular material with maximum par- was divided into two symmetrical parts corresponding to two ticle size of 25 mm was used as the backfill. The GRS abut- different embankment profiles: the NW wall and the W wall. ments have performed well since construction. A fine sand used as backfill was compacted at its maximum standard Proctor density. The backfill had a dry unit weight of 16.6 kN/m3, friction angle of 30 deg, and cohesion of Field Experiments of Bridge-Supporting 2 kPa. Segmental concrete blocks were used as the facing. Structures The NW wall was reinforced by a nonwoven geotextile with a tensile strength of 25 kN/m at 30 percent strain. The W wall The test conditions and measured performance of six field was reinforced with a knitted woven geotextile with a tensile experiments of GRS bridge abutments and piers are described strength of 44 kN/m at 15 percent strain. The reinforcement below. The six field experiments are the Garden experi- spacing was 29 cm. mental embankment in France (Gotteland et al., 1997), the The reinforced embankment was loaded in the same way FHWA Turner-Fairbank GRS bridge pier in Virginia (Adams, as a bridge deck through a foundation slab. The 1.0-m-wide 1997), the Havana Yard GRS bridge pier and abutment in foundation was 1.5 m from the edge of facing. The embank- Colorado (Ketchart and Wu, 1997), the Fiber Reinforced ment was loaded by a beam acted on by two thrust rams, each Plastic (FRP) geogrid-reinforced retaining wall in Japan restrained by four tie-bars anchored into the embankment (Miyata and Kawasaki, 1994), the Chemie Linz full-scale foundation. GRS embankment in Austria (Werner and Resl, 1986), and The test embankment was instrumented to monitor the the Trento test wall in Italy (Benigni et al., 1996). performance of the embankment during loading. Two months 4m 5m 3m 5m 4m 4m 4.35 m NW Wall W Wall Geomembrane HDPE 38.41 m Intermediary Zone 1m 1.5 m 1m 1.5 m 2 sheets length = 3.6 m 2 sheets length = 3.6 m 15 X 0.29 m = 4.35 m NW W Figure 2-8. The Garden experimental embankment, France (Gotteland et al., 1997).

OCR for page 11
20 after the construction of the reinforced embankment, a load base and three layers of biaxial geogrid reinforcement, was applied to the foundation on the top of the structure until spaced 0.3 m apart. The RSF was 1.2 m deep, over an area failure occurred. The load was applied over 2 days. The of 7.3 m by 7.5 m. Figure 2-9 shows the cross-section of the experiment was terminated when a permissible facing dis- GRS bridge pier. placement was reached (0.20 m maximum horizontal dis- The pier was constructed with modular concrete blocks as placement for the NW wall and 0.15 m for the W wall). The the facing and was reinforced with a polypropylene woven structure was examined layer by layer by careful excavation geotextile, Amoco 2044, at vertical spacing of 0.2 m. Because at the end of the experiment. the geotextile was stronger in the cross-machine direction A localized failure was noted to have occurred at the upper (38 kN/m at 5 percent strain) than in the machine direction layers of the NW wall (with a nonwoven geotextile), whereas (21 kN/m at 5 percent strain), the width and length direc- the W wall (with woven geotextile) experienced a deeper tions were alternated between layers. The backfill was clas- failure with a downstream tilting effect giving rise to a wide sified as a well-graded gravel. The maximum dry unit surface crack at the upstream end of the geotextile sheets. weight was 24 kN/m3, per AASHTO T-99, with the opti- However, the main deformation occurred at the upper layers mum moisture content being 5.0 percent. The average com- for both walls. The load-settlement curves for two walls paction in the field was about 95 percent of the maximum showed a distinct break point that corresponds to two distinct dry density. slopes of the curves. The "critical loads" at the break point The FHWA Turner-Fairbank pier was load-tested by for the NW wall and the W wall were quite large, 140 kN/m applying vertical loads on top of the backfill in two loading and 123 kN/m respectively. The corresponding settlements cycles. The first loading cycle was performed when the pier were 36 mm and 33 mm, respectively. The lower "critical height was 3.0 m. The 3.0-m-high pier was loaded to about load" of the W wall can be attributed to its shorter "interme- 600 kPa. The settlement varied roughly linearly with the diate" reinforcement (see Figure 2-8), even though the re- applied load. At 200 kPa, the settlement was about 13 mm, inforcement had higher strength than that of the NW wall. and at 600 kPa, the settlement was about 34 mm. The maxi- mum lateral displacements at 200 kPa and 600 kPa were about 6 mm and 20 mm, respectively. Case B2: FHWA Turner-Fairbank GRS Bridge The second loading cycle was performed when the pier Pier, Virginia (Adams, 1997) was at its full height. The second loading cycle was con- ducted in three parts: A full-scale bridge pier was constructed and load tested at the Turner-Fairbank Highway Research Center, FHWA, 1. The pier was incrementally loaded to 415 kPa and then in McLean, Virginia. The pier was 5.4 m high and 3.6 m by held for 100 minutes; 4.8 m at its base. The pier was supported on a reinforced soil 2. The load was then ramped up to 900 kPa and held for foundation (RSF). The RSF comprised compacted road 150 minutes and unloaded; and 3.8 m Compacted Road Base 0.2 m lifts (Crushed Diabase) Reinforcement (Woven Polypropylene Fabric) Dry Stacked Modular Blocks (Split Face Cinder Blocks) 5.4 m Fabric Tail Between Blocks Facing Block Leveling Pad Reinforced Soil Foundation 1.2 m Figure 2-9. Cross-section of the FHWA Turner-Fairbank GRS bridge pier (Adams, 1997).

OCR for page 11
21 3. The pier was reloaded to 415 kPa and then held for At 200 kPa, creep was not a concern in a closely spaced 100 minutes. reinforced soil system with a well-compacted granular backfill. At 415 kPa pressure, the pier settled about 25 mm; at 900 kPa, the settlement was about 70 mm. During the reload cycle, settlement was roughly reduced by a factor of two. Case B3: Havana Yard GRS Bridge Pier and Abutment, Denver, Colorado (Ketchart and Wu, At 200 kPa, the pier deformed laterally less than 3 mm. The 1997) maximum strain in the reinforcement was recorded near the middle of the pier and was about 2.3 percent. The reinforce- The Havana Yard GRS bridge supporting structures, con- ment strain in the first loading cycle at 400 kPa was about sisting of two piers and one abutment, were constructed 0.5 percent. inside a 3.5-m-deep pit. The outer pier and the abutment were Based on the measured results, the following conclusions 7.6 m tall, and the center pier was 7.3 m tall. The center pier were made: and the abutment were of a rectangular shape, and the outer pier was of an oval shape. The base of the outer pier, the cen- At 200 kPa of loading, the GRS pier performed very sat- ter pier, and the abutment were, respectively, 2.4 m by 5.2 m isfactorily. The maximum strain in the reinforcement (major and minor axes), 2.1 m by 4.8 m, and 3.6 m by 5.2 m. was 0.25 percent. The maximum lateral displacement Segmental concrete blocks, each 0.2 m in height, were used was 3 mm, yet no cracks occurred in the facing blocks. as the facing element for all three structures. On the east face For the full-height load tests, the vertical settlement was of the abutment, the facing assumed a 13 percent "negative about 15 mm in the initial load cycles and about 5 mm batter" up to a height of 3.5 m. From 3.5 m to the top of the during the reload cycles. abutment were walking steps. Figure 2-10 shows the cross- Preloading reduced vertical settlement by about 50 per- section of the Havana Yard GRS bridge pier and abutment. cent and limited the vertical creep deformation. Pre- The backfill was a "road base" material containing 13 per- loading did not reduce lateral deformation. cent of fines. The maximum dry unit weight was 21.2 kN/m3, Concrete Blocks Girder 0.74 m 2.44 m 5.61 m Steps 3% Reinforcement spacing 0.2 m Reinforcement spacing 0.2 m Reinforcement spacing 0.2 m 5% 5% 4% 4% 3% 7.62 m 7.31 m 7.62 m 3.53 m 13% 0.9 m Reinforcement spacing 0.3 m Outer Pier Center Pier Abutment Figure 2-10. Cross-section of the Havana Yard GRS bridge pier and abutment, Denver, Colorado, (Ketchart and Wu, 1997).

OCR for page 11
22 and the optimum moisture content was 6.7 percent, per magnitude (0.2 percent to 0.4 percent). This suggests AASHTO T-180. The field measured average dry unit that the lateral movements of these piers are comparable. weight was 19.1 kN/m3 (or 90 percent relative compaction) Under a sustained load of 2,340 kN for 70 days, the in the center pier and the abutment. The average placement creep displacements in both vertical and lateral direc- moisture was 2.5 percent for the center pier and 1.6 percent tions of the outer pier were about 4 times larger than for the abutment. The fill density of the outer pier was those in the abutment because of lower compaction believed to be significantly lower than these measured val- effort of the outer pier. The maximum vertical creep dis- ues because a lighter compaction plant was employed for the placement was 61.6 mm in the outer pier and 18.3 mm outer pier because of its size and shape. in the abutment. The maximum lateral creep displace- The reinforcement for all three structures was a woven ment was 59.5 mm in the outer pier and 14.3 mm in the polypropylene geotextile (Amoco 2044) with a wide width abutment. tensile resistance at 5 percent strain in the machine and cross- A significant part of the maximum vertical and lateral machine directions of 38 kN/m and 21 kN/m, respectively. creep displacements of the pier and the abutment The vertical spacing of the geotextile reinforcement was occurred in the first 15 days. At 15 days, the maxi- 0.2 m. The top four layers of the reinforcement in the abut- mum vertical and lateral creep displacements were about ment employed a wrapped-around procedure behind the 70 percent to 75 percent of the creep displacements at facing block. A center geotextile "tail," 1.2 m in length, was 70 days in respective directions. placed between each of these four layers to connect the back- Creep deformation of the structures decreased with time. fill to the facing blocks. The vertical creep rates reduced nearly linearly (on log- On top of the outer pier and the abutment were 0.3-m-thick log scale) with time. The creep rate of the outer pier concrete pads to support steel bridge girders. The concrete (7.5 mm/day after 3 days and 0.1 mm/day after 70 days) pads were 0.9 m wide and 3.1 m long for the piers and 2.4 m was higher than that of the abutment (2.2 mm/day after wide and 3.7 m long for the abutment. The clearance distance 3 days and 0.03 mm/day after 70 days). of the concrete pad was 0.2 m from the back face of the fac- Hairline cracks of the facing blocks occurred in the ing blocks. outer pier and the abutment because of the lateral bulging For loading tests, three steel bridge girders were placed and the down-drag force because of the friction between over the top concrete pads of the outer pier and the abutment. the backfill and the facing blocks. Installing flexible Each girder was supported by steel bearing plates resting on material (i.e., cushion) between vertically adjacent blocks the concrete pads. The steel bearing plates were located may have alleviated this problem. along the centerline of the top concrete pads. The span of the The maximum strains in the reinforcement were less girders was 10.4 m. A total of 124 concrete blocks was than 1.0 percent. Compared with the rupture strain of the placed on the girders. The total load was 2,340 kN, corre- reinforcement of 18 percent, the safety margin against sponding to an applied pressure of 232 kPa and 130 kPa on rupture of reinforcement appeared to be very high. the outer pier and abutment, respectively. The calculated lateral displacements from the reinforce- The findings and conclusions of this project as summa- ment strain distribution were in very good agreement rized by the authors are as follows: with the measured lateral displacements. With the less stringent construction condition (using a The displacements of the pier and the abutment were lightweight vibrating compaction plate), the outer pier comparable at applied load of 2,340 kN. The maximum showed about 1.5 times larger vertical displacement-to- vertical displacement was slightly higher in the outer height ratio than the Turner-Fairbank pier; whereas the pier than in the abutment. The maximum vertical dis- lateral displacements were similar. placements were 27.1 mm in the abutment and 36.6 mm in the outer pier, corresponding to 0.35 percent and 0.48 percent of the structure height. The maximum lateral Case B4: FRP Geogrid-Reinforced Retaining displacement in the abutment was somewhat higher Wall, Japan (Miyata and Kawasaki, 1994) than that in the outer pier. The maximum lateral elon- gation of the perimeter was 4.3 mm in the abutment and The FRP geogrid-reinforced test embankment consisted of 12.7 mm in the outer pier. three types of GRS retaining walls, referred to as Types A, The ratio of the vertical movement to the structure height B, and C (see Figure 2-11). The test embankment had a at 232 kPa of the outer pier (0.48 percent) was higher height of 5.0 m with a 0.3H:1V slope. Type A had a soft than that of the FHWA Turner-Fairbank pier (0.30 per- wall face with gabions only, Type B had a cement-treated cent). This may be attributed to the much lower com- wall face, and Type C was a gravity retaining wall made paction effort on the outer pier. The reinforcement of cement-treated soil. An FRP geogrid, having a tensile strains in the fill direction of the outer pier and the strength of 49 kN/m at 2 percent strain, was used as re- FHWA Turner-Fairbank pier, however, were of similar inforcement.

OCR for page 11
23 Type A Type B Type C 1m 2.3 m Loading Loading Loading 1.5 m Reinforcements for Separator bearing capacity (GB10) Gabion 1:0.3 1:0.3 1:0.3 5m 2m FRP Geogrid (GB5) FRP Geogrid (GB5) FRP Geogrid (GB5) 1.5 m Cement-treated soil Cement-treated soil Figure 2-11. Cross-section of the FRP geogrid-reinforced retaining wall, Japan (Miyata and Kawasaki, 1994). To perform the load test, a 2.3-m-wide loading frame was 510 kN or 130 kN/m2. Without rupture or critical defor- first placed on the top of the embankment at 1.0 m from the mation occurring, the load of 130 kN/m2 corresponded to front wall face. Loads were then applied by inserting loading 1.7 times the theoretical breaking load, as determined from steel plates into the loading frame in steps up to a total weight Bishop's lamella circular sliding surfaces method. The mea- of 590 kN (q = 127 kPa). The lateral displacement of Type A sured maximum vertical settlement and lateral displacement was much larger than those of Types B and C. The maximum of the embankment face were about 16 cm and 11 cm, values were about 40 mm, 25 mm, and 20 mm for Types A, B, respectively. and C, respectively. In addition, the deformation mode of Type A differed from Types B and C. Type A, with a soft wall face, showed a swell-out mode with the maximum lateral movement Case B6: Trento Test Wall, Italy (Benigni et al., occurring at about the mid-height of the wall. Types B and C, 1996) with a rigid wall surface, showed a forward fall-down mode with the maximum movement occurring at the top of the wall. A 5-m-high test wall, referred to as the Trento test wall, was constructed in Northern Italy. A well-graded cohesion- less sandy gravelly soil, with shear strength parameters c = Case B5: Chemie Linz Full-Scale GRS 100 kPa and = 40 determined from the CD triaxial tests, Embankment, Austria (Werner and Resl, 1986) was used as backfill. It had a dry unit weight of 19.6 to 20.4 kN/m3 with in situ water content of 2.4 to 5.5 percent. A multi-layered geotextile reinforced embankment was Wrapped wall face was adopted for the wall. A geocompos- built in 1981 and had been exposed to 3 years of extreme cli- ite, with tensile strength of 27 kN/m at 16 percent strain per matic fluctuations and environmental influences by the time it DIN EN ISO 10319 (German Standards: geotextiles wide- was loaded in 1984. The height of the embankment was 2.4 m. width tensile test), was used as reinforcement. The reinforced Figure 2-12 shows the geometry and loading scheme of the section of the wall was constructed in lifts separated by geo- Chemie Linz full-scale GRS embankment. A silty gravelly composite layers and with the final spacing being 0.5 m. The sand was used as backfill. The design shear strength parame- reinforcement length was 2.0 m (40 percent of the wall ters of the backfill were = 21, c = 20 kN/m2, and the bulk height). Figure 2-13 shows the cross-section of the test wall. unit weight was 19.3 kN/m3. A polypropylene needle-punched During construction, the wall face was supported by 1-m- nonwoven geotextile, with tear strength = 16 kN/m, grab high wooden forms, assembled with wide long boards nailed strength = 1,200 N, and elongation at failure = 80 percent, as to brackets, which were wedged against a temporary scaffold. determined with DIN 53815 (German Standards: testing of On completion of each lift, the underlying geosynthetic was textiles) was used as reinforcement. The vertical reinforce- wrapped around at the face and extended 2 m inside the back- ment spacing was 0.35 m. Wrapped facing was adopted for the fill. A new reinforced layer was then unrolled parallel to the structure. wall face and positioned so that a 0.5-m-long tail rested on top Seven steel slabs, each measuring 3 m by 1.3 m by 0.2 m, of the one already wrapped around, while the remaining and two steel cylinders of 0.8 m in diameter were used to 2.5-m-long part draped over the wooden form. No windrows load the GRS embankment, which produced a total load of were used to anchor the reinforcement in the backfill.

OCR for page 11
24 3m Steel Cylinders Slabs (7~0.35 m) 2.4 m 9.8 m (a) Front View 1.3 m =85 1.4 m 2.4 m 60 1m = 2.5 m (b) Side View Figure 2-12. Geometry and loading scheme of the Chemie Linz full-scale GRS embankment, Austria (Werner and Resl, 1986). The loading test was performed by the weight of the summarized in Table 2-1. The performance characteristics stacked iron ingots evenly distributed over two 3 m by 3 m include wall height, backfill, reinforcement type, reinforce- wide loading platforms placed on top of the wall. The maxi- ment spacing, facing type and connection, ratio of reinforce- mum surcharge loading, reached after 51 hours, was esti- ment length to wall height, maximum settlement of loading mated at 84 kPa. The wall did not collapse under the applied slab, maximum lateral movement of the wall face, maximum load, although somewhat large movements were recorded. In reinforcement strain, and failure pressure. addition, although most of the horizontal and vertical dis- Based on the measured performance, the following obser- placements were not recovered on unloading, it appeared that vations are made in relation to performance, design, and con- the wall had sustained almost no damage. struction of GRS bridge-supporting structures: GRS bridge abutments with flexible facings are indeed a Synthesis of Performance Characteristics viable alternative to conventional bridge abutments. All six in-service GRS bridge abutments (Cases A1 through The main performance characteristics of the 12 case his- A6) exhibited satisfactory performance characteristics tories reviewed in this study, including six in-service GRS under service loads. The maximum settlements and max- bridge abutments and six full-scale field experiments, are imum lateral displacements for all the abutments were