STRESS IN ORTHOTROPIC STEEL DECK COMPONENTS DUE TO VEHICULAR LOADS WanChun Jen1 and Ben T.

Yen2
1

ATLSS Research Center, Lehigh University, 117 ATLSS Drive, Bethlehem, PA; PH (610) 758-5613; FAX (610) 758-6568; email: jew2@lehigh.edu 2 ATLSS Research Center, Lehigh University, 117 ATLSS Drive, Bethlehem, PA; PH (610) 758-5553; FAX (610) 758-6568; email: bty0@lehigh.edu Abstract Laboratory measurement of local stresses was made on components of a full scale model of orthotropic deck panel of Bronx-Whitestone Bridge. Simulated wheel load of trucks was placed on the deck at various locations along longitudinal stiffening ribs of trapezoidal shape. The loads induced local stresses and local bending of diaphragm web plates and rib walls. The local stresses were moderately high in magnitude in all components of the model deck. Introduction Orthotropic steel decks with longitudinal, closed rib stiffeners serve the dual function of being the upper flange of the box girder, real or equivalent, and being the member to transfer vehicular loads to other parts of the bridges. Stresses induced by vehicular loads are the primary cause of fatigue cracks in decks. Some analytical and experimental studies have been conducted to examine the local stresses in deck components in order to alleviate fatigue cracking at connections between longitudinal ribs and transverse diaphragms (Connor 2001, Tsakopoulos 2002, Ye 2004). After the fatigue testing in laboratory of a full scale model deck of the BronxWhitestone Bridge (BWB) in New York City, stresses at various components of the specimen were measured for examination of the regional effect of wheel loads of trucks. This paper briefly summarizes some of the results. Test Specimen and instrumentation The model deck of BWB was 48 feet (14.63 m) long and 37 feet (11.28 m) wide, 1

as shown schematically in Figure 1. It consisted of 2 panels of continuous deck supported by three floorbeams. The floorbeams were each supported by a wall column, and by a stiffening girder at the other end, Figure 2. The test deck modeled half of the bridge width with the test deck plate connected to the wall longitudinally simulating continuity at the center line of the bridge.

REACTION WALL
Rib No. 16 15 14 13 12 11

Line 5, 6 Line 1 ...

Floor Beam A Diaphragm A1
1 2

10 9 8 7 6 5 4 3 2 1

Floor Beam B
3 4

Diaphragm B1

Floor Beam C

LOADIND POSITION

Figure 1 Top View of Specimen Setup

Figure 2 Elevation of Specimen The orthotropic deck had 5/8 in. (16 mm) thick deck plate, two longitudinal plate stiffeners and fourteen longitudinal trough stiffeners with 5/16 in. (8 mm) thick wall. The diaphragm web was inch (13 mm) thick. 2

The trapezoidal ribs were 13 in. (330 mm) wide at the top, 5 in. (127 mm) wide at the bottom and the inclined walls were 14 in. (357 mm) deep. The spacing between ribs is also 13 in. (330 mm), so the deck plate is supported uniformly for most of its width. A large number of strain gages and displacement transducers (LVDTS) were placed between Diaphragm A1 and B1. The emphasis was on measuring strains in the deck components at Diaphragm B. Figure 3 shows schematically the strain gage locations around Rib 6 at Diaphragm B. Loading Procedure Hydraulic actuators applied vertical loads though rubber pads (footprints) to the deck to simulate wheel loads of HS 25 trucks. Since linear behavior at details was observed from strain reading during loading, each applied load was increased from 20K to 80K to exaggerate the strains in the components for easy comparison.

Figure 3 Strain Gage Location and Loading Lines, Rib 6 at Diaphragm B

Figure 4 Loading Positions The loads were applied individually along six lines, as depicted in Figure 3. These lines simulate the truck wheels directly over the connections of ribs wall to the deck plate, and in between. 3

Four load positions were used in each line: above Diaphragm A1, halfway between A1 and B, above Diaphragm B and halfway between B and B1. These positions are indicated in Figure 4. Results Linear Elastic Behavior The test deck behaved linear elastically under the applied load. Figure 5 shows as examples the load versus deflection relationship at nine LVDTs throughout the test deck. The load was on Line 1 at Position 2. Figure 6 shows the load versus strain relationship of four strain gages at Diaphragm B1. In all cases, the deflection and strain increased linearly with the applied load, and return to the same original value when the loads were removed. Strains in the Deck Pate The strain distributions on the bottom of deck plate along Diaphragm A1 when the simulated load was at different positions on Line 6 are plotted in Figure 7. Line 6 was along the mid-width of Rib 8. With 80K applied at Position 1 directly over Diaphragm A1, the maximum stress on the bottom of the deck at Rib 8 was not the highest. The highest stress of about 7 Ksi (230 in/in strain) occurred when the load was at Position 2 between Diaphragm A1 and B. When the 80K load was at Position 4 between two diaphragms, the bottom of the deck plate was in low tension at the junction with the rib wall at Ribs 8, 9 and 10.
120

100

80 Load (kips)

LVDT_1 LVDT_2 LVDT_3 LVDT_4 LVDT_5 LVDT_6 LVDT_7 LVDT_8 LVDT_9

60

40

20

0 -0.05 0 0.05 0.1 Deflection (inch) 0.15 0.2 0.25

Figure 5 Load vs Deflection (Line 1, Loading Position 2) 4

400

200

0 Load (kips) 0 -200 20 40 60 80 100 120

-400

Strain Gage 46 Strain Gage 47

-600

Strain Gage 13 Strain Gage 15

-800 Strain (in./in.)

Figure 6 Load vs. Strain at Diaphragm B1 (Line 1, Loading Position 2)

100

P = 80 K

50

0 -10 Strain (in./in.) -50 0 10 20

Loading Position 1 Loading Position 3

30

40
Loading Position 2 Loading Position 4

50

60

-100

Loading Position 3+ 4

Loading Position 2+ 3+ 4

-150

-200

-250 Distance (inch)

Figure 7 Strains on the Bottom of Deck Plate along Diaphragm A1 (Loads on Line 6) Figure 8 shows the strain distribution on the bottom of deck plate along diaphragm B under the same loading positions of Figure 7. All stresses were low under a 80K load, with a maximum of less than 10 Ksi (3450 in/in strain) when the 5

applied load was between diaphragms. Under a wheel load of HS 25, the maximum live load stress under the deck would be less than 3 Ksi. The strain distributions along Diaphragm B in Figure 8 indicate that the regional influence of loads on deck plate stresses was confined to only the adjacent one or two ribs. 600 P = 80 K
500

Loading Position 1 Loading Position 2 Loading Position 3 Loading Position 4

400

300 Strain (in./in.)

Loading Position 3+ 4 Loading Position 2+ 3+ 4

200

100

0 -10 -100 0 10 20 30 40 50 60 70

-200 Distance (inch)

Figure 8 Strains on Bottom of Deck Plate along Diaphragm B (Loads on Line 6) Stresses in Diaphragm at Cutout The longitudinal stiffening ribs passed though diaphragm webs at cutouts, as shown in Figure 2 and 3. The geometry of the cutout was determined by analysis and was one of the main reasons of fatigue testing the model deck. In the static testing of the deck model to examine the regional effects of loads, strains on diaphragm webs at cutouts were measured when loads were applied at various positions. Example results are presented in Figures 9 to 10. In Figure 9, the strains on Diaphragm B at the top of cutout for Ribs 6, 7, 8, and 9 are presented. The loads were applied along Line 3 over the connection between the deck and a web of Rib 6. When the 80K load was at Position 2 between diaphragms, the highest strain of about 380 in/in was induced at the cutout at the other web of Rib 6. That is corresponding to less than 3 Ksi under a wheel load of HS 25. Again, the influence of a wheel load on local stresses is limited to one adjacent 6

rib only, as depicted by the strains in Figure 9. P = 80 K 200

100

Observed strains at Cutout

0 -10 -100 Loading Position 1 -200 Loading Position 2 Loading Position 3 Loading Position 4 -400 Loading Position 3+ 4 -500 Loading Position 2+ 3+ 4 10 30 50 70 90 110

-300

-600 Distance (inch)

Figure 9 Strains at Cutout on South Face of Diaphragm B (Loads on Line 3) P = 80 K

0 10 20 30 40 50 60 70 80 90 100

100

0 -10 -100 Strains (in./in.)

-200

Line 1 Line 2

-300 Line 3 -400 Line 4 Line 5 -500 Distance (inch)

Figure 10 Strains on the South Face of Diaphragm B at Cutout (Loading Position 2)

The variation of stresses on Diaphragm B at cutouts of Rib 6 to 9 for loading Position 2 of Loading Line 1 to 5, are presented in Figure 10. Position 2 is between diaphragms. As the loading line moved away from Rib 6 (to Lines 4 and 5), the strains at the cutout of Rib 6 decreased. The region of effect of only one adjacent rib is again obvious. When the applied load was on Line 5 over Rib 8, relatively high strains developed in Rib 8. There was a connection of the diaphragm between Rib 8 and Rib 9 (as shown in Figure 2), which affected the behavior of Rib 8. Influence Line of Strains The strain diagrams presented so far, Figure 7 to 10, provide information on the distribution of stresses in deck components near the point of loading. The subsequent diagrams show the magnitude of strain at specific points as a load was applied at different locations nearby. The results are essentially Influence Lines. Figure 11 shows the horizontal and vertical strains on the web of Diaphragm B at the connection of deck plate and Rib 6. The strain gage locations are given in Figure 3. Both the horizontal and vertical strains (at gage 10x and 12y) were highest when the 80K load was directly above. The magnitude of strains at these gages decreased as the load was placed away. Similarly, the horizontal and vertical strains at gage 20(x) and 22(y) were the highest when the load was directly above. When the load was at Line 5 over Rib 7, the strain at all four gages at Rib 6 were near zero. Line 1 2 3 4 5
400 300 200 100 0 -100 -200 -300 -400 0 5 10 15 20 25 30 35 40 45 50

Strain (in./in.)

Gage 10 x Gage 20 x Gage 12 y Gage 22 y

Loading Line Distance (inch)

Figure 11 Influence Lines of Horizontal (x) and Vertical (y) Strains on Diaphragm B at Connection of Deck and Rib 6 (Loading Position 2)

Line 1
600

Gage 128 N 400 Gage 129 S Gage 134 N 200 Strain (in./in.) Gage 135 S

-200

-400

-600 0 5 10 15 20 25 30 35 40 45 50

Loading Line Distance (inch)

Figure 12 Influence Lines of Strains on North and South Faces of Diaphragm B at Top of Cutout, Rib 6 (Loading Position 2) On Diaphragm B at the cutout of Rib 6, strains on the South surface of the diaphragm web at the top of cutout (Figure 12) were slightly higher when the load was at Line 2 between the rib walls than when the load was directly over the walls. The difference in strains at back to back strain gages (128/129,134/135) indicates that the diaphragm web was subjected to local bending. In this case, the magnitude of bending was about the same on the two sides of Rib 6. On the other hand, the diaphragm web plate local bending was not prominent at the lower corners of the cutout, as the strains at back to back strain gages (130/131, and 132/133 in Figure 13) increased or decreased similarly. All four gages had the highest strain when the applied load was directly above, and had almost no strain when the load was one rib away. The rib walls were also subjected to local bending when the applied load was nearby. The Influence Lines for strain gage pairs on the walls of Rib 6 are given in Figure 14. The difference in strain between the gages of each pair (137/138, 139/140) signifies local plate bending. The maximum difference was about 350 in/in in the 5/16 in. thick plate, comparing to about 400 in/in in the in. web plate of Diaphragm B in Figure 12.

Line 1
600

Gage 130 N 400 Gage 131 S Gage 132 N Gage 133 S Strain (in./in.) 0

200

-200

-400

-600

-800 0 5 10 15 20 25 30 35 40 45 50

Loading Line Distance (inch)

Figure 13 Influence Lines of Strains on Diaphragm B at Lower Corner of Cutout, Rib 6 (Loading Position 2) Line 1
600

400

200 Strain (in./in.)

-200

Gage 137 N Gage 138 S

-400

Gage 139 N Gage 140 S

-600 0 5 10 15 20 25 30 35 40 45 50

Loading Line Distance (inch)

Figure 14 Influence Lines of Strains on Web of Rib 6 (Loading Position 2) Figure 15 shows the Influence Line of strain on the wall of Rib 6 between Diaphragm B and B1. When the applied load was at Position 4 directly over the cross section of the rib, the bottom of the rib had the highest strain whether the load was over the rib wall or in between. The shape of these Influence Lines is typical for a 10

continuous beam. The strains on the web of the rib about an inch from the deck and directly below the applied load along Line 3, however, decreased without changing sign when the load was moved away. For this Loading Line, there was practically no local strain at the point on the opposite web of the rib.
1200 Rib wall East Line 2 Rib bottom Line 2 Rib wall West Line 2 Rib wall East Line 3 St rain (in./in.) Rib bottom Line 3 Rib wall West Line 3 600 400 200 0 -200 -180 -160 -140 -120 -100 -80 -60 -40 -20 0 -200 -400 -600 -800 Distance (inch) 1000 800

Figure 15 Influence Lines of Strain on Rib 6 between Diaphragm B and B1 Discussions and Conclusions The strains in components were measured when a simulated wheel load of trucks was placed at various locations on the deck. Results indicate that the local stresses induced by the wheel load were essentially zero when the wheel load was one rib away in the transverse direction of the orthotropic deck. By considering the configuration of trucks on the bridge, it can be concluded that wheel loads of parallel trucks have little effect on the local stresses in deck plate, diaphragms and rib walls. In the longitudinal direction, multiple simulated wheel loads were applied during testing. Because of the difference in stiffness of the diaphragms with or without floorbeams, the effect of load position on local stresses was strongly influence by the relative position of the loads to the diaphragms. Results of strain measurement confirmed the linear elastic behavior of the deck and thus the adequacy of 11

superposition of multiple loads. Computer analysis using a finite element model also confirmed the measured strains when one or multiple simulated wheel loads were applied. A limited parametric study of the deck component dimensions is being conducted. Preliminary results indicate that were the deck plate thickness be reduced to in. with all other thickness and dimensions being the same, the stresses in diaphragms would increase slightly under the simulated wheel load while the stresses in the rib walls would increase more. Obviously the relative dimensions of components have strong effect on local stresses. For the specific replacement orthotropic deck of BWB, it can be calculated that vehicle induced local stresses in the components are moderately high but well within permissible values. Acknowledgments The funding for this study was from Pennsylvania Infrastructure Technology Alliance (PITA) and the model deck was provided by Triborough Bridge and Tunnel Authority of New York City (TBTA). The prototype deck panel was designed by Weidlinger Associate Inc. of New York City and manufactured by Leonard Kunkin Associates of Line Lexington, PA. Testing was conducted at ATLSS Research Center of Lehigh University. References Connor, R. J. and Fisher, J. W. (2001), Results of Field Measurements on the Williamsburg Bridge Orthotropic Deck," ATLSS Report 01-01. Tsakopoulos, P. T., and Fisher, J. W. (1999), Williamsburg Bridge, Replacement Orthotropic Deck, As-Built Full-Scale Fatigue Test," ATLSS Report 99-02. Tsakopoulos, P. T., and Fisher, J. W. (2002), Fatigue Resistance Investigation for the Orthotropic Deck on the Bronx-Whitestone Bridge," ATLSS Report 02-05. Ye, Q., and Fanjiang, G. N. (2004). Analysis and Design of Steel Orthotropic Decks, IABSE, Shanghai, China, Sep. 2004. 222-223

12