U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
An empty FRP panel, used previously to determine flexure stiffness (panel #3), was selected to fabricate the railing post-FRP deck connection. The FRP panel consisted of seven FRP tubes with the narrow side of the panel facing the rail. A steel railing post was mounted over a high-density polyethylene (HDPE) pad and attached to the FRP deck using four 1-inch threaded anchors, connecting the railing post to a steel base plate on the bottom face of the FRP deck. Figure 75 shows details of the different components of the railing post specimen. The steel base plate had four bolts welded onto the plate; the anchor rods were bolted to the rail base until they felt "snug-tight."
Figure 75. Photos and sketch. Railing post specimen.
Figure 76 shows details of the specimen test setup. Two steel beams (W6×20) were used to attach the FRP panel to a structural frame; their distance to the railing post base as well as to each other was based on the full-scale prototype. A hydraulic ram (22-ton capacity and 6-inch stroke) was used to apply force on the railing post at 25 inches from the base of the deck panel. The FRP deck was connected to the steel beams using hollo-bolts, flat-clips, and washers (see figure 77) similar to those that will be used in the full-scale prototype. The hollo-bolts were tight using the "turn-of-the-nut" procedure: the nuts were turned 1/3 after snug-tight conditions. After the bolts were tensioned, the torque in each bolt torque was verified using a calibrated torque wrench, and it was found to be in the order of 65 to 75 lb-ft.
|1. Side View||1. Top View|
Figure 76. Photo and diagrams. Railing post test setup.
To secure the W6×20 girders to the structural frame, steel braces, A490 bolts, and dywidag bars were used. Neither the FRP panel nor the braces were directly attached to the floor.
(a Photograph showing bracing system
(b) Bracing system & hollo bolt details
Figure 77. Photo and diagram. Detailing of steel beam-structural frame connection.
To measure deflections, six dial gages and four string pots were attached to the FRP deck, railing post and W6×20 girders. In addition, digital image correlation was used to monitor the movements of few of the bolts on the back side of the panel. Figure 78 shows photographs and location of all these transducers; their orientation with respect to a Cartesian axis of reference is also indicated in these images.
|(a) Profile view of gage locations||(b) View of hydraulic ram at 32 kip-failure|
c) Back view of gage locations
(d) Photograph of panel back view
(e) Top View of railing post gages
(f) Top view of post & L-brackets – Test 1
Figure 78. Photos and diagrams. Location of transducers used during the test.
The first test (Test 1) was completed on July 16, 2012. A 7-inch-diameter steel plate was placed on top of the hydraulic jack to distribute the load; see figure 78f. The specimen was incrementally loaded, and measurements were taken at 0, 3, 6, 9, 12, 13.5, and 15 tons. At 15-ton (30-kip) load, the steel plate sat at a very pronounced angle, so the test was stopped to prevent instability in case of failure. Measurements were not taken during the unloading part.
With a roller support on the ram to increase stability (figure 78b), Test 2 was completed on July 17, 2012. For this test, it was decided to load the specimen to AASHTO TL-2 load level (27 kips), unload incrementally to obtain more complete data, then load the specimen again up to failure (Test 2-Failure) and back down. Table 29 shows the data points recorded for the end rail displacement measured by string pot 2 (see figure 78a) for all tests; figure 79 shows a plot of these data. For Test 2, due to the possibility of large deformations and brittle failures, some transducers were removed (string pots 1 and 3 and the digital image correlation camera); in addition, Dial_X3 and Dial_Y3 replaced two LVDTs to record the data at the same locations.
|Load (kip)||Deflection (in.)|
Max 30 kip
Max 27 kip
Max 32 kip
Figure 79. Graph. Load-displacement behavior of railing post end (string pot 2).
No cracking sound was heard during Test 1 or Test 2 up to 27 kips. When the specimen was reloaded again (Test 2-Failure), cracking sounds were heard around 32 kips. It was also observed that, under increasing displacement, no further load increase was obtained. At a displacement of 5.2 inches, registered by string pot 2 (see figure 78b), the railing post-FRP deck connection failed when a corner of the back steel plate was pulled (wedged) into the FRP deck, as shown in figure 80a. As noted, the significant deformation of the rail base and HDPE pad is believed to be the reason for the increasing pulling forces on the bolts at the bottom of the steel base plate. The steel plate was bent as a result of this pulling effect; figure 80b shows the extent of the deformation. The wedging effect caused a local shear failure of the deck through the first layer of the tubes (aka punching shear), shown in detail in figures 80c and 80d. Post-test visual observations of all bolt holes in the FRP panel both at the railing post and hollo-bolt connections revealed no signs of bearing failure. Figure 79 shows that non-linear deformation (most likely coming from the failure of the connection) started between 30 and 32 kips. Permanent deformation registered by string pot 2 was in the order of 1 inch.
|(a) Steel plate wedging failure||(b) Bent on steel plate|
|(c) FRP deck panel failure-close-up||(d) FRP deck failure-photograph post-test|
Figure 80. Photos. Failure mode of the railing post-FRP deck panel connection.
In Test 1, three string pots were attached to the railing post to measure its total vertical displacement relative to the ground. String pots 1 and 3 were used to monitor any possible rotation of the railing post during loading, whereas string pot 2 was placed at the railing post end; see figure 81. Results indicate that the railing post was loaded symmetrically across its width.
|(a) Railing post load-displacement behavior||(b) Photograph of 3 string pots used|
Figure 81. Graph and photo. Vertical displacement of the railing post, Test 1.
The large amount of vertical displacement registered by the string pots indicates that, in addition to the railing post deformation, other sources of deformation also contributed to this displacement: 1) the rotation and movement of the anchor bolts connecting the post to the FRP deck panel, which in turn allowed for railing post base and HDPE rotation, 2) the HDPE pad deformation, 3) the FRP deck panel deformation, 4) the possible movement of the hollo bolt connections, and 5) movement of the bracing system used to secure the steel beams to the structural frame. Figure 82 shows a sketch illustrating these possible sources of deformation. In the next sections, they will be described in more detail.
Figure 82. Diagram. Railing post connection (deformed shape, not to scale).
The total vertical displacement at the end of the railing post at 27 kips (AASHTO TL-2), and failure (32 kips) was 3.11 inches and 5.21 inches, respectively. The deformation of the HDPE pad was found to be a significant contributor of this vertical displacement. (For reference, the AASHTO Load and Resistance Factor Design Bridge Design Specification for TL-3 and TL-4 uses 54 kips; rail heights are also higher.)
The four corners of the HDPE pad closest to the railing post base were measured against the FRP deck face using calipers. Table 30 shows values of the change in distance from the original position at zero-load for Test 2. Corners A and B are located toward the edge of the FRP deck panel, whereas corners C and D are on the opposite side, see figure 83.
The measurements show that the HDPE pad at corners A and B was compressed under increasing load, whereas corners C and D were stretched. A straight bar was used during testing to confirm that the HDPE pad was bending under increasing load; therefore, it cannot be assumed that it deforms as a rigid plate. After failure of the steel plate and unloading of the specimen, permanent deformation of the HDPE was also measured at the corners.
To estimate the contribution of the HDPE deformation to the vertical railing post displacement, its rotation was calculated based on the deformations recorded in table 30. At 27 kips of applied load, it was estimated that the vertical deflection at the end of the railing post corresponding to the HDPE rotation was in the order of 40 percent of the total value measured by string pot 2. It is also expected that the HDPE deformation would have greatly increased the tension forces in the bottom two bolts of the railing post connection. The gap between the railing post base and the HDPE pad was measured in the vicinity of the maximum load of 32 kips and found to be in the order of 0.125 inches.
|Load (kip)||Horizontal Displacement from FRP deck surface (in.)|
|Corner A||Corner B||Corner C||Corner D|
|(a) Sketch showing HDPE pad corners||(b)HDPE deformation at 27 kip|
|(c)HDPE deformation at max load (32 kip)||(d)HDPE deformation post steel plate failure|
Figure 83. Photos. HDPE deformation at different loading levels, Test 2.
Movement of the FRP deck panel in two directions was recorded using LVDTs and dial gages. Displacement in the direction perpendicular to the applied load (x-direction) was estimated from dial gages Dial_X1, Dial_X2, and Dial_X3. Dial_X2 and Dial_X1 measure the FRP deck movement with respect to the top and bottom steel girder, whereas Dial_X3 measures the total movement of the top of the panel relative to the structural frame (attached to the floor). Figure 84 shows these deformations for Test 2. As expected, the deformation of the FRP deck panel is larger toward the cantilever edge and smaller near the steel girder support. At the location near the second (bottom) steel girder, the FRP panel experiences a reversed curvature, thus deforms in the opposite direction than the cantilever edge. The curves on figure 84 show that the FRP deforms linearly up to 30 kips; beyond that load level, non-linear deformations appear, possibly associated with the damage of the railing post connection. At 27 kips, the displacements registered by Dial_X3, Dial_X2 and Dial_X1 are 0.195, 0.188, and -0.233 inches, respectively.
Figure 84. Graph. Horizontal deflection (x-direction), Test 2.
Figure 85 shows the vertical displacements (in the y-direction) recorded by the three dial gages during Test 2. Dial_Y1 and Dial_Y2 measure the vertical movement of the FRP panel relative to the adjacent steel girders while Dial_Y3 measures the overall movement of the top of the FRP panel (with respect to the structural frame, attached to the floor). As expected, the displacement of the top of the FRP panel is larger than those near the steel girders. The behavior of Dial_Y3 shows that the panel does not return to its initial position after the first unloading. After failure of the railing post connection, the specimen was unloaded; the permanent displacement of the FRP deck at the top was found to be 0.058 inches.
Figure 85. Graph. Vertical displacement (y-direction), Test 2.
The vertical displacement at the top of the FRP panel (Dial_Y3) is a combination of the relative movement of the steel girders with respect to the structural frame, as well as the true "elongation" of the panel under loading. To evaluate the magnitude of the FRP elongation, the deflection of the bottom girder relative to the floor was measured by string pot 4 (see figure 86). At 27 kip of loading, string pot 4 registered 0.082 inches. This small deflection suggests that the brace and bolt system used to hold the steel girders and FRP deck panel system in place, performed well. Overall, this global movement was limited to about 3 percent of the total movement railing post end displacement.
At 27 kip of loading, the top of the FRP deck panel displaced 0.195 inches (captured by Dial_Y3). Assuming that the first steel girder deformed in similar magnitude as the second girder (measured by string pot 4), it can be assumed that the FRP elongated 0.113 inches at this loading (4 percent of the vertical deflection measured by string pot 2, discussed previously).
Figure 86. Graph. String pot 4 movement, Test 2.
The digital image correlation technique was used to monitor the movement of three bolts in the back of the FRP deck panel during Test 1 (loading up to 30 kips and unloading): two bolts on the steel base plate and one on the steel fastening clip. To protect the digital equipment, digital image correlation was not used during Test 2 to failure. Figure 87 shows the location of the monitored bolts as well as their movements, with increasing loading with respect to the Cartesian coordinate system used in this report (y, z axis).
Bolts 1 and 2 were expected to have similar movements with respect to each other since they were both attached to the steel base plate. They both show similar displacements (0.17 inches upward) in the y axis, but their displacements in the z axis are in the opposite directions. Bolt 3 (expansion bolt connecting the FRP panel to the first steel girder) showed smaller displacement values in the y direction (0.10 inch), but similar in magnitude to the global displacement measured by string pot 4.
|(a) Photograph of the back of the FRP deck; area used for digital image correlation is highlighted|
|(b) Bolt 1 displacement||(c) Bolt locations|
|(d) Bolt 2 displacement||(e) Bolt 3 displacement|
Figure 87. Photos and graphs. Movement of three bolts on FRP back panel using digital image correlation, Test 1.
As indicated before, the FRP deck was connected to the steel beams using 5/8-inch-diameter expansion (hollo) ebolts, flat-clips, and 3/8-inch-diameter washers (see figure 88a). The expansion bolts were tightened using the “turn-of-the-nut” procedure. The bolt length (3 inches) allowed them to pass through two walls of the FRP tubes, as shown in figure 88b. During the testing of the railing post, the movement of one of the bolts was monitored using digital image correlation, showing very small displacement values that can be related to the movement of the steel girders. After the test was concluded, visual inspection of all the bolt holes in the FRP panel revealed no signs of bearing failure. It can be inferred that the bolted connection was able to effectively hold the FRP panel in place, up to the load and displacement levels when the railing post connection failed.
|(a) Sketch of bolt connection||(b) Cross section of FRP tube & connection|
|(c) Profile view of bolted connection||(d) Typical FRP deck hole (post test)|
Figure 88. Diagram and photos. Expansion bolt connection.