Load Testing and Load Distribution Response of

Paper No. 06-1679

Duplication for publication or sale is strictly prohibited without prior written permission of the Transportation Research Board.

Title:

Load Testing and Load Distribution Response of Missouri Bridges Retrofitted with Various FRP Systems Using a Non-Contact Optical Measurement System

Author:

Wesley J. Merkle BLN, L.L.C. Indianapolis, IN John J. Myers, Ph.D., P.E. The University of Missouri at Rolla Center for Infrastructures Engineering Studies 325 Butler-Carlton CE Hall Rolla, Missouri, USA 65409-0030 Tel: 573-341-6618 Fax: 573-341-6215 Email: [email protected]

Transportation Research Board 85th Annual Meeting January 22nd - 26th, 2006 Washington, D.C

TRB 2006 Annual Meeting CD-ROM

Paper revised from original submittal.

Merkle, W. and Myers, J.J.

2

Load Testing and Load Distribution Response of Missouri Bridges Retrofitted with Various FRP Systems Using a Non-Contact Optical Measurement System Wesley J. Merkle, Design Engineer John J. Myers, Ph.D., P.E., Associate Professor of Civil, Architectural and Envir. Engineering The University of Missouri at Rolla Center for Infrastructures Engineering Studies 325 Butler-Carlton CE Hall Rolla, Missouri, USA 65409-0030 Email: [email protected]

Abstract. Strengthening structures with Fiber-Reinforced Polymer (FRP) composite systems has been growing in popularity over recent years for the many benefits that the technology offers. A project entitled “Preservation of Missouri Infrastructure: Validation of FRP Composite Technology through Field Testing” was undertaken in Missouri which was designed to push forward composite strengthening schemes for use on real structures. Using five deficient bridge structures demanded that the strengthened structures be monitored for performance to verify that the composites were working and that they were not degrading over time. Monitoring the structures meant scheduling load tests for all five bridges. Difficulties in using traditional monitoring equipment due to site constraints, like Linear Variable Displacement Transducer (LVDT) systems, on these structures required the search for a better monitoring system. This paper presents high-precision Surveying Equipment as a new serviceability monitoring system for load testing; the materials, procedures, and data analysis techniques are discussed. This paper also presents the results of the load testing and the load distribution. All five strengthened bridges are compared in terms of serviceability before and after strengthening results to show that the FRP strengthening systems are performing acceptably. FEM modeling is performed to compare analytical predictions to measured values. Keywords. Bridge Strengthening, FRP Retrofit, Load Distribution, Load Testing, Non-Contact Measurements INTRODUCTION The transportation infrastructure in the United States is often taken for granted by most citizens; it can be considered the backbone of America because without it, our nation’s economy as well as our lifestyle would simply cease to exist. Bridges are a vital part of this infrastructure; as many structures reach a critical age and become deficient or obsolete, they must be repaired or replaced [13]. Over 40 percent of the nation’s bridges are estimated to be in need of repair or replacement [20]. The Federal Highway Administration (FHwA) claims that 25 percent of our nation’s bridges are either deficient or functionally obsolete; in Missouri, that claim jumps to 35 percent [17]. Many states, including Missouri, cannot afford to tend to all the repair and replacement needs immediately, so the bridge owners post load restrictions as a temporary solution to the problem. Bridges become deficient or obsolete for a variety of reasons. Bridge decks typically require major repair or replacement every 15 to 20 years, while the substructure and superstructure tend to last 40 years or more [13]. Deicing salt applications eat away concrete while water seeping through cracks in the concrete corrode unprotected reinforcing steel. Over the past decades, standard truck design weights have increased; as a result, older bridges that are still in good condition do not have the load carrying capacity required for today’s traffic needs. Traffic volumes have increased dramatically in recent years. Safety considerations have changed as well – most roadways today are designed to have 12-foot wide lanes. Both changing traffic demands and an already decaying infrastructure require engineers to find alternative solutions on a slim budget. Fiber Reinforced Polymer (FRP) composites have enjoyed growing popularity over the past decade in repair and strengthening applications of existing bridges and buildings. Use of FRP technologies for such applications was well known for benefits including low cost, fast installation, minimal traffic disruption, and anticipated long term durability. With the benefits of such a temporary strengthening system in mind, engineers are looking to push using these technologies on real structures. In the case of bridge rehabilitation with new FRP strengthening technologies, diagnostic load testing is required to verify the effects of remediation on the bridge [7]; load testing is an effective means of evaluating the




3

performance of a structure [4]. Protocol for load testing should be carefully planned; however, no guidelines for load testing currently exist – testing procedures are to be determined by the organization performing the test [4]. LOAD TESTING OVERVIEW Load testing is observing and measuring the response of a structure subjected to controlled loads in the elastic range [5]. Load testing has many purposes and benefits to match; it allows engineers to determine unknown factors such as composite action and unintended fixity levels, continuity, and participation of other members. Load testing is proven to be an effective way for structural engineers to evaluate the structural performance of a bridge [4]. Several researchers have investigated in-situ load testing of bridges [10, 21, 23]. Load testing has two main varieties – diagnostic testing and proof testing. A diagnostic load test is chosen to find the response characteristics of a structure, including load response, load distribution, validating analytical models, and evaluating the effectiveness of structural repairs and upgrades where the structure may not be accurately load rated [5]. The Five-Bridges Project utilizes diagnostic load testing to validate that the FRP strengthening is working effectively. Alternatively, the proof load test is designed to establish the safe load carrying capacity of the structure while loading the structure within the elastic range [5]. Determining the load rating for a bridge could be evaluated from a proof load test. Diagnostic and proof load tests can be performed using static or dynamic loads, or both. Static loads are stationary loads; they induce no vibrations. The location and magnitude of the loads may change during testing. Using standard loaded trucks to park on the structure tends to be easiest for testing real structures. Dynamic load tests use moving loads to induce vibrations and find a very rapid loading response of the structure. These tests allow for the analysis of dynamic load allowance, frequency, vibration, and a stress range for fatigue evaluation [5]. Dynamic loading is again easiest done with loaded trucks and driving them over pre-determined locations at given speeds. The following section describes the five bridges that were strengthened with various FRP systems and the load testing techniques. BRIDGE CHARACTERISTICS, STRENGTHENING, AND MONITORING DESCRIPTIONS Bridge Descriptions The following sub-sections briefly describe the five Missouri Dept. of Transportation (MoDOT) Bridges that were strengthened with FRP and subsequently load tested. Table 1 provides a summary of the strengthening type and location for the respective bridges. Additional details about the bridges and specifics on the design and strengthening systems used may be found in two detailed reports: “Preservation of Missouri Transportation Infrastructure: Validation of FRP Composite Technology through Field Testing – Volume I” [22] and a report by Merkle and Myers [15]. Only a brief overview of the bridges are presented herein as a lead into the load testing and monitoring of the bridges. Structural analysis was performed using HS20-44 and 3S2 truck loading cases following AASHTO specifications for each bridge to determine the appropriate retrofit requirement (see Table 1). Strengthening was designed following ACI 440.2R-02 [8] guidelines for each of the bridges. Bridge X-596. Bridge X-596 carried Route C over Lander’s Fork Creek in Morgan County, Missouri (MoDOT District 5). The bridge average daily traffic (ADT) was 2000; it was load posted for trucks over eighteen tons to travel fifteen miles per hour over the bridge. The bridge was built in 1946 as a Deck Girder/Reinforced Concrete Tee Beam Structure. The bridge had three spans: 42.5 feet, 52.5 feet, and 42.5 feet totaling 137.5 feet with no skew angle. All spans were simply – supported three span deck girders, or three Tee Beams spaced 9 feet on center, with diaphragms at mid span. The deck slab was monolithically cast 6 inches deep and 23.6 feet wide; the roadway was 20 feet wide and striped as a one – lane bridge (see Fig. 1). The bridge’s visual inspection rating was as follows: deck 6, superstructure 5, and substructure 5. Concrete cracking, exposed reinforcement, and concrete deterioration were found in various locations throughout the deck, superstructure, and substructure. Rusted steel bearing plates were found on the end bents. Concrete and steel samples were taken from the bridges and tested. The concrete compressive strength was found to be 6,000 psi and the yield strength of the reinforcing steel was found to be 40 ksi. The girders were strengthened for shear as needed. The bents were also strengthened for shear and flexure. Bridge T-530. Bridge T-530 carried Route M over Crooked Creek in Crawford County, Missouri (MoDOT District 9). The bridge ADT was 200; it was load posted for trucks over twenty-one tons to travel fifteen miles per hour over the bridge.




4

The bridge was built in 1939 as a Deck Girder/Reinforced Concrete Tee Beam Structure. The bridge had five spans each 47.5 feet long, totaling 237.5 feet on a 30 degree skew. All spans were simply – supported four Tee Beams spaced 6.5 feet on center, with diaphragms at mid span. The deck slab was monolithically cast 6 inches deep and 23 feet wide; the roadway was 20 feet wide and striped as a one – lane bridge (see Fig. 2). The bridge’s visual inspection rating was as follows: deck 5, superstructure 5, and substructure 5. Some concrete deterioration and minor concrete cracking throughout the structure; the bent caps had cracking from steel corrosion. The concrete compressive strength was found to be 6,250 psi and the yield strength of the reinforcing steel was found to be 40 ksi. Flexural strengthening schemes were implemented for the deck and girders. The girders were not strengthened for shear; however, FRP U-Wraps were installed to help anchor the flexural reinforcement. The bents were not strengthened. Bridge X-495. Bridge X-495 carried Route C over Crane Pond Creek in Iron County, Missouri (MoDOT District 9). The bridge average daily traffic (ADT) was 300; it was load posted for trucks over nineteen tons to travel fifteen miles per hour over the bridge. The bridge was built in 1948 as a Deck Girder/Reinforced Concrete Tee Beam Structure. The bridge had three spans: 42.5 feet, 52.5 feet, and 42.5 feet totaling 137.5 feet with a 30 degree skew. All spans were simply – supported three span deck girders, or three Tee Beams spaced 9 feet on center, with diaphragms at mid span. The deck slab was monolithically cast 6 inches deep and 23 feet – 7 inches wide; the roadway was 20 feet wide and striped as a one – lane bridge (see Fig. 3). The bridge’s visual inspection rating was as follows: deck 6, superstructure 6, and substructure 7. The concrete and all of the bridge’s main structural components were found to be in good condition. The concrete compressive strength was found to be 5,450 psi and the yield strength of the reinforcing steel was found to be 40 ksi. Flexural strengthening schemes were implemented for the deck and girders. The girders were strengthened for shear as needed. The bents were also strengthened for shear and flexure. Bridge P-962. Bridge P-962 carried Route B over Dousinbury Creek in Dallas County, Missouri (MoDOT District 8). The bridge ADT was 350; it was load posted for trucks over eighteen tons to travel fifteen miles per hour over the bridge. The bridge was built in 1956 as a Deck Girder/Reinforced Concrete Tee Beam Structure. The bridge had three spans each 42.5 feet long, totaling 127.5 feet on a 15 degree skew. All spans were simply – supported four Tee Beams spaced 9 feet on center, with diaphragms at mid span. The deck slab was monolithically cast 6 inches deep and 23.7 feet wide; the roadway was 20 feet wide and striped as a two – lane bridge (see Fig. 4). The bridge’s visual inspection rating was as follows: deck 7, superstructure 6, and substructure 6. The concrete and all of the bridge’s main structural components were found to be in good condition. The concrete compressive strength was found to be 6,850 psi and the yield strength of the reinforcing steel was found to be 40 ksi. Flexural strengthening schemes were implemented for the deck and girders. The girders were strengthened for shear as needed. The bents were also strengthened for shear and flexure. Bridge Y-298. Bridge Y-298 carried Route U over Crews Branch Creek in Pulaski County, Missouri (MoDOT District 9). The bridge ADT was 1,100; it was load posted for trucks over eighteen tons to travel fifteen miles per hour over the bridge. The bridge was built in 1937 as a Reinforced Concrete Slab Structure. The bridge had two continuous spans each 15 feet long, totaling 30 feet on a 45 degree skew. The deck slab was monolithically cast 7 inches deep and 27.2 feet wide; the roadway was 24 feet wide and striped as a two – lane roadway (see Fig. 5). The bridge’s visual inspection rating was as follows: deck 5, superstructure 5, and substructure 5. The concrete in the East span was in poor condition, with evidence of water seepage and some permanent deformation. The concrete in the West Span was in more sound condition. Longitudinal cracking and exposed reinforcement was found in various locations. The concrete compressive strength was assumed to be 4,000 psi and the yield strength of the reinforcing steel was assumed to be 40 ksi. Flexural strengthening schemes were implemented for the deck. The strengthening was oriented in the same direction as the existing reinforcing steel. The supports were not strengthened. Instrumentation Program This overall research project retrofits five structurally deficient Missouri bridges with Fiber Reinforced Polymer (FRP) composites as the primary strengthening material. Load testing was to be carried out in effort to verify that the FRP strengthening system was performing as expected. All five bridges were scheduled for load testing biennially for five years following strengthening. Because of existing site conditions at four of the five bridges,




5

traditional serviceability monitoring techniques, like using LVDT systems, would be extremely difficult; the search for a better system was initiated. A variety of high-precision optical, laser methods and other surveying tools are commonly used to measure deflection on bridges [5]. Traditional surveying equipment, similar to older generations of total stations, was tried in the past to measure deflection, but the equipment was ultimately not precise enough to measure deflection on stiff structures with short spans [23]. In the last few years surveying technology has made many technological advances. These advances in technology have made it possible to use surveying equipment like the total station and GPS to monitor deformation in structural monitoring applications [19]. The Leica TCA 2003 total station measures with much more accuracy than its predecessors [14]. Robotics further increase the accuracy of this instrument. This model is commonly used in absolute monitoring applications [14]. Traditional land surveying techniques utilized many components, with the total station as the principal device. The total station (see Fig. 6) was set atop a secure tripod in a location with an unobstructed view of the field targets. Reference points (see Fig. 7), or backsites, were set in place to transfer a horizontal angle or an elevation from the reference point to the total station, and then from the total station to the targets (see Fig. 8). The reference points also served to check that the total station had not moved. The reference point was a prism mounted atop a secure tripod; the targets were prisms fixed atop a metal rod (range pole). In the case of load testing, the prisms were fixed to points on the tested structure. Prior to implementing the use of this system several case studies were investigated in the laboratory and in-situ to compare its performance to other traditional systems. The performance of the total station in these preliminary validation studies was found to be acceptable [16]. After one year of using the Surveying Equipment for load testing bridges, the accuracy is generally better than 0.005 inches. The major drawback is its limitation to monitor dynamic load tests and the time required to read points and acquire data. Load Testing Program The monitoring plan called for load testing the critical locations for each structure. Unless the bridge had identical span lengths or utilized varied strengthening systems on a particular span, the longest span was selected for load testing as it would yield the highest stress and deflection values. In most cases the total station was the only viable system to measure deflection due to girder height and or water restrictions below the structure. Table 2 summarizes the bridge designation with the respective span length that was load tested. All bridges were load tested prior to strengthening and biennially thereafter for a contract period of 5 years (see Table 3). Herein, the first three load tests are discussed. Each bridge was load tested using two H20 configuration dump trucks provided by MoDOT. The trucks were loaded up to 60 kips each, with most of the weight carried by the rear axles. The trucks were parked in five separate locations on the selected span; the locations were determined to maximize stresses in the structure. Stop 1 produced maximum shear loading condition, where the trucks were parked side-by-side with their rear axles near one end of the supports. The distance from the support was chosen based on placing the back-most wheels a minimum distance equal to the depth of the section. This was assumed to ensure that the trucks’ weight would develop shearing stresses in the span instead of only loading the support underneath. Stop 2 produced maximum moment loading condition, where the trucks were parked side-by-side with their rear axles centered at midspan. Stop 3 produced maximum shear loading at the opposite end of the span condition, similar to Stop 1. Stop 4 produced an overload condition on fully-monitored exterior girder, where the trucks were parked back-to-back closest to the barrier wall. Stop 5 produced a similar overload condition on the interior girder, where the trucks were parked backto-back with their rear axles centered at the bridge’s centerline. Fig. 9 shows the procedure for Bridge X-596 as a representative example. This paper will report the results from the high moment Load Stops 2, 4 and 5. DATA ANALYSIS Basic Calculations. For every set of readings, an averaged elevation was listed to each point number. The sets were the first control readings, truck stop readings, and the last control readings. A control set was used to establish a baseline, or a plane of reference for the truck stop sets. Deflection was found by subtracting the baseline reading from the truck stop reading (see Fig. 10). Adjusting the Baseline. The initial purpose of the two control sets was to verify that the individual points on the bridge span had not undesirably moved; plotting the control set difference showed this. However, when considering thermal effects this was not the case. Many load tests took place on sunny or partly sunny days, and a changing temperature gradient




6

during the day induced a net camber (upwards deflection) in the tested span. The problem was determining how this gradient induced a change in camber during testing. All of the load tests were performed from around 10 AM to 2 PM; this was the part of the day when ambient temperatures were rising. Thermal effects were monitored during every load test by measuring the top and bottom surface temperatures of the bridge deck and girders. The gradient was calculated to be the average difference between top and bottom surface temperatures (an approximation rather than the exact gradient). To better address this issue, internal thermocouples would help characterize the exact thermal gradient occurring between load tests, but this was not feasible for in-service structures. The change in this gradient was computed twice for each load test: between the start and half-way, and between half-way and the finish. The camber was assumed to change as the measured temperature gradient changed. Therefore, if the change in the temperature gradient was the same both times then change in camber was changing linearly, and the baseline for all truck stop sets was computed the same as previously. If the change in the temperature gradient was not linear, then the baselines were adjusted accordingly using a similar procedure. Fig. 11 and Fig. 12 show the control set difference where thermal effects were significant (sunny conditions during testing) and where thermal effects were insignificant (cloudy conditions during testing). Adjusting Individual Data Sets. The purpose of measuring the reference points before and after every set of readings was to verify that the total station was not moving. The measured elevation of each reference point was compared from one set of readings to the next. It was noted that movement of the reference points was acceptable if the movement was accounted for by the internal error of the instrument. The purpose of using three reference points was justified by the following: if one reference point had moved, and only one other reference point was used, then there was no way to know which reference point was good. While more redundancy (more than three reference points) was acceptable, it was not performed because of time constraints in measuring extra points in every data set. An adjustment method was developed for the case that all three reference points appeared unstable and may be referenced [15]. This however was not required in most cases as the reference points did not settle or move. ANALYTICAL MODELING Three (3) Mass-section Tee-Beam models and two (2) Finite Element Modeling (FEM) were created in effort to accurately predict the deflections produced from load testing and to subsequently be used to investigate using load testing as an inspection rating tool. This section discusses the concepts, approaches, assumptions, and procedures for each model. The models used basic calculations to find pre-strengthened deflection measurements. The loads used were taken from the truck weight tickets collected during testing. H20 truck geometry and truck stop diagrams (see Fig. 9) were used to locate the wheel loads, which were assumed to act as point loads. The axle loads for each truck at each test are shown in Table 4. For every structure except Bridge Y-298, the skew angle was ignored because it was less than thirty degrees [4]. The measured in-situ material properties (except for Bridge Y-298) were used in the bridge modeling. Reinforcement quantities were used based on values reported in the contract documents. Mass-Section Tee-Beam Modeling. This scheme was only used to model Deck Girder/Reinforced Concrete Tee Beam Structures. The entire crosssection of the bridge was taken as one Tee-Beam; the entire deck was treated as the flange and all three girders were treated as one (sum the widths and reinforcements). The truck loads were therefore not distributed to the individual beams. While this model could not differentiate between interior and exterior girder deflections, it would set a general value. The model allowed for the concrete barrier walls to contribute to stiffness – these barrier walls were ignored for design; however, they were known to contribute to the structure’s overall stiffness [12]. Using the dimensions shown in the plans for each bridge, a basic contributing shape was developed for each wall. For bridges X-596, T-530, and X-495, the contributing section was simplified to 13 inches tall by 20 inches wide. The remaining portions were ignored as they were very slim in the cross-section and did not contribute to the overall stiffness in a significant fashion. For Bridge P-962, the contributing section was 8 inches tall by 20 inches wide, set atop another 20 inches tall by 8 inches wide block. The barrier walls were assumed to be built from the same concrete as the rest of the structure; a perfect bond was assumed between the wall and deck structure. To see the barrier walls’ effects, the resulting Tee-Beam’s Gross Moment of Inertia was computed with the barrier walls set to zero, fifty, and one-hundred percent effective. Again, each structure was assumed to be uncracked.




7

Finite Element Modeling. FEM was performed using the Ansys Classic 7.1 software package [9]. While FEM was the most complex way to estimate deflection, using the computer and software package made FEM a desirable method for this analysis. Several assumptions were made in using FEM. The bridges were assumed to have consistent material properties in all locations. Their elements behaved as a linear – elastic and isotropic material. The Modulus of Elasticity for concrete (Ec) was defined by Equation 1 [6]. Poisson’s Ratio (v) was assumed to be 0.20 [6]. Ec = 57,000 * √ f’c (PSI) Equation 1 The element SOLID65 was chosen to model three – dimensional reinforced concrete with or without reinforcing steel bars. Flexural reinforcement was added to the Tee-Beams for all of the bridges except Bridge Y298 where the deck reinforcement was added. The exact bridge geometry was used. Two models were developed for each bridge – one with and one without the barrier wall. Each structure was modeled using a cracked and uncracked section. Flexural cracks were simulated in the beams by changing the Modulus of Elasticity to near zero for a block of elements near midspan at a depth approximated by cracking moment calculations. To model each bridge with FEM, first the inputs were set. The bridge span was then drawn using the known geometries. Next, the model was meshed into “brick” elements no more than six inches on one side; the structure was broken into approximately 100,000 nodes (elements). The supports were then set to accurately represent the simply – supported bridge span. Loads were next applied according to which truck stop was modeled. With all inputs set, the model was solved and vertical deformation was plotted on a three – dimensional contour map. Fig. 13 shows an example of the output from Ansys. THEORY AND EXPERIMENTAL This section compares the theoretical models defined in the previous section with the measurements obtained in the field. Only before-strengthening testing was used as the models were developed considering the original unstrengthened structure. The results are discussed and compared with visual inspection ratings of each structure. Comparison Methods. For comparison, only the mid-span deflections of the interior and exterior girders (monitored with nine targets per girder) were used. The calculated deflections for each bridge and each loading scenario were taken from all three models discussed in the previous section. Only the maximum flexural testing (truck stops 2, 4 and 5) data was used because the deflection for maximum shear loading was minimal. Shear contribution to deflection was also assumed insignificant in comparison to the flexural contribution. For each theoretical value, a percent error was computed; this was the difference between theoretical and experimental values divided by the experimental value. A positive error value meant that the theory over-estimated the deflection; negative error value meant that the theory underestimated the deflection. The comparison was done for before-strengthening load testing only. Results of Analysis. Tables 5 through 8 show the results of theoretical and (pre-strengthened) experimental deflection comparisons for each bridge except Bridge Y-298. A percent error relative to the measured value was listed for every model. For each given truck stop, consistencies were noticed both between the models and between the bridges. Differences in percent error from one model to another were similar from one bridge to the next. When the barrier wall was considered, the stiffness increased, and as a result the estimated deflection decreased. This was true for the MassSection Models and the FEM Models. To quantify a health rating of the bridge, the percent errors were averaged into one number for each bridge. Initially it was noted that for Bridge P-962, the average percent error of the models was significantly different than that of the other three bridges (X-596, T-530, and X-495). In fact, the theoretical deflection was typically much less than the measured deflection – especially with both FEM models. This indicated that a higher level of softening (i.e. cracking) had occurred with this structure. Although there were no obvious visual deficiencies in the structure, the FEM models were run again for this bridge with cracked sections instead of uncracked sections. The results were shown in Table 8; the percent error values for this bridge were then more consistent with the other three bridges. Only FEM models were produced for Bridge Y-298. Modeling schemes included both continuous and simply supported span structures as well as with and without the barrier walls. The length-to-width ratio of each span was found to be about 3, making each span one-way by definition [6]. As expected, the barrier walls were




8

found to be insignificant. The simply-supported model more accurately resembled the field measurements. This was valid because the condition of this structure was relatively poor, so the concrete was most likely cracked, perhaps due to overloaded conditions, and the actual continuity of the spans would therefore be small. The resulting FEM models were compared to the load testing results, shown in Figs. 14 and 15. Average theoretical error values were computed similarly to the other four bridges, except only one model (shown in Figs. 14 and 15) was used. For Bridge Y-298, the complex geometry made drawing conclusions from theoretical and experimental results difficult. For both tests on the East Span, the predicted deflection was less than the actual deflection from span width 20 to 28 feet, whereas the opposite was observed from span width 2 to 20 feet. This finding was consistent with an area of damaged concrete found underneath this span of the structure, span width 25-35 feet, which could weaken the structure and reduce its stiffness. Inspection Ratings Comparison. As discussed in the previous section, the percent error from theory was used to quantify the health of the structure. The models were created for ideal structures, meaning that any deficiencies would not be accounted for. The limitation of these models to predict actual bridge deflections under load should be noted. Deficiencies in the structure would most likely reduce the structure’s stiffness increasing the deflection. Using this knowledge, an attempt was made to correlate the average percent error from theory to the MoDOT [11] visual inspection ratings shown previously. The deck and superstructure ratings were averaged and plotted against the average theoretical error (see Fig. 16). A linear trend line was then plotted; the equation of this trend line (Equation 2) was used to compute a load test rating (see Table 9). (Rating) = 5.33 + 1.19 * (Theoretical Percent Error)

Equation 2

While the results suggest that there may likely be a relationship between load testing performance (theoretical percent error) and visual inspection ratings, many more similar structures should be load tested and analyzed to prove or disprove this approach. Validating this relationship would give bridge engineers a new tool to quantify their visual inspection ratings as well as show the real condition and the load-carrying capacity of the structure. Load testing new bridges prior to opening the bridge to service would better establish a benchmark for future load tests, supplementing the analytical predictions which often yield high variability. The benchmark would be compared to future load test data to show any real degradation in the structure over time, comparing that to lowered visual inspection ratings. STRENGTHENING EFFECTS In the 5-Bridges Project load testing, one major variable was the changing truck weight. Target truck weights were not requested during the first and second rounds of testing; as expected, truck weights varied – sometimes significantly. To find out how a changed truck weights affected deflection at each monitored location (load distribution), Finite Element Models were run for each bridge and each truck stop (without barrier walls, discussed in the previous section). Each girder’s deflection at midspan was recorded from the model. To compare the measured deflection values from the second load test to the first, the second measured value was multiplied by the ratio of the first theoretical value to the second theoretical value. This was used to account for changing truck weights. To compare one load test to another, average change in deflection was measured by taking the difference between before and after-strengthening measurements and dividing this value by the original or before-strengthening deflection. For each bridge, the data points were combined into a weighted average – weighted by the magnitude of the original measured deflection. Table 10 shows the results of the analysis. For every bridge, the averages varied from one point to another by ±10 percent. There were no obvious consistencies between Truck Stops; however, some consistencies existed between exterior and interior girders. Bridge T-530 and P-962 showed an expected small increase in apparent stiffness (decrease in deflections). The term apparent stiffness is used by the authors because the actual contribution of FRP to the section properties is negligible. Much of the increase in apparent stiffness for these strengthened structures is due to the FRP’s ability to reduce crack opening under load and thereby appear to enhance the stiffness of the structure. This increase in apparent stiffness does indicate that the FRP strengthening is engaged. For Bridge P-962, the apparent stiffness increase was the same for all three spans. The results of load tests one (before strengthening), two and three (afterstrengthening) for Bridge P-962 were shown in Figs. 17 and 18. The graphs were normalized for the variables




9

previously discussed. For both the longitudinal and transverse deflection plots, the consistent increase in apparent stiffness was nearly consistent for the entire span. The third test showed that the increased stiffness was holding since the second load test. The third round of load testing was presented in the next section. Bridge X-495 posted a loss in apparent stiffness (increase in deflection). FRP installation most likely did not degrade the stiffness of the structure, so this can most likely be explained by either an incorrectly-reported truck weight, or an invalid assumption regarding thermal effects. Recall that for the first round of load tests, the bridges’ temperatures were not monitored. Bridges X-596 and X-495 were expected to post smaller increases in apparent stiffness because the tested spans were only strengthened on the deck and interior girders. Bridge X-495 showed almost no loss in apparent stiffness for the interior girder with most of the loss coming from the exterior girders – indicating the presence of strengthening on the interior girder. However, Bridge X-596 did not show a difference in changed apparent stiffness between girders. Bridge Y-298 posted a strong increase in overall apparent stiffness. Here, the change in truck weights was almost fifty percent (see Table 4). For all five bridges, further testing would help prove and disprove the values in Table 10. LOAD TEST COMPARISON AND DISCUSIONS Table 11 shows the change in deflection from Load Test 1 (before-strengthening) to both Test 2 and Test 3 (afterstrengthening). In comparing Test 1 to Test 3, each bridge posted a reduction in deflection values, or an increase in stiffness. Bridges X-596, T-530, and X-495 posted a significant increase in apparent stiffness, especially in comparison to the Test 2 results. However, Bridge P-962 posted the same increase in apparent stiffness (see Figs. 17 and 18). Bridge Y-298 posted similar results as well. The significant increase in apparent stiffness of Bridge T-530 was shown in Figs. 19 and 20. The graphs were normalized for the variables discussed in the previous section. The serviceability of these bridges will continue to be monitored over the remaining 4 years of load testing. LOAD DISTRIBUTION Live load distribution factors were attained from AASHTO LRFD Section 4.6.2.2 [4]. Table 12 lists the calculated distribution factors for both interior and exterior girders. The values represent the maximum response of any individual girder for a particular loading case. The case of two lanes loaded would be compared to Stop 2 of the load testing (also Stops 6 and 7 for Bridge P-962) while one lane loaded would be compared to Stops 4 and 5 of the load testing. For every load test at each bridge except Bridge Y-298, the individual girder response was taken from midspan. The deflections were normalized for the variables previously discussed. Again, only Stops 2, 4, and 5 (plus 6 and 7 for Bridge P-962) were considered in the analysis. To examine the load distribution to the bridge girders, the individual girder deflection was divided by the sum of the deflections for all three girders; the result was a fraction of the load that the individual girder carries [23]. The calculated responses due to the load testing are called the load distribution coefficients. The responses used for calculation may be strain or deflection [18]. The beforestrengthening responses for each bridge and each girder are listed in Table 13. The load distribution coefficients ranged from 50 to 95 percent of the values computed from AASHTO; this implies a level of conservatism in the AASHTO LRFD distribution factors. The truck stops were set to maximize the responses in different locations instead of locating the trucks in the AASHTO design lanes. The conservative nature of the AASHTO design specification can be seen in comparing the data in Tables 12 and 13. Changes in the computed after-strengthening load distribution coefficients can be seen in Tables 14 and 15. Deck strengthening and flexural strengthening of different girders would produce changes in the load distribution coefficients. However, FRP strengthening only slightly changed the overall stiffness of each bridge. The small magnitude in changes from before to after strengthening, the changes in values from Tables 14 and 15, and the inconsistencies at each bridge made drawing conclusions difficult. From the above data, it can be concluded that FRP strengthening has very little or no effect on a bridge’s load distribution coefficients. LOAD TESTING SENSITIVITY Accurately placing the H20 trucks in a load test was critical to the test’s outcome. Therefore, a sensitivity study was undertaken to examine the influence of axle placement and load accuracy. The trucks wheels were assumed to be placed within 12 inches of the desired location. To find the effect of this variation, the data was modeled for Truck Stop 2 on Bridge X-596 using Finite Element Modeling (FEM), modeled without barrier walls and both trucks weighing 55 kips each. The first model was run with the loads placed at the desired locations. The second model was run with one of the two trucks simulated as being misplaced 12 inches transversally. The third model was run




10

with one of the two trucks simulated as being misplaced 12 inches longitudinally. Finally, the forth model was run with one truck weighing 5 kips less than in the first model; this would simulate an error in reporting the truck weight. Truck weight was considered most at risk for error as it was not overseen by load testing personnel. Other possibilities for error included the accuracy of the Surveying Equipment (± 0.005 inches), unknown thermal and rebounding effects, different truck axle configuration dimensions, and shifted loads in the truck bed during transport. The percent changes in deflection were reported in Table 16. The comparison points were on each girder at midspan. The case study for Bridge X-596 shows the sensitivity involved in load testing for the Five Bridges Project. The errors discussed may account for the variations seen in the results previously discussed. CONCLUSIONS The following conclusions were drawn from the results of these investigations: 1. Where site problems make LVDT and Dial Gage systems very difficult to set up and operate successfully, the Surveying Equipment is an ideal alternative. 2. Surveying Equipment can compete with traditional systems to monitor serviceability of structures during static load testing. 3. The Surveying Equipment currently cannot display real-time results during a load test. The Surveying Equipment cannot be used to monitor serviceability of structures during dynamic load testing in an effective manner. 4. Bridges P-962, T-530, and Y-298 posted expected increases in apparent stiffness after the second load test. 5. All five bridges posted increases in post-strengthening apparent stiffness after the third load test. 6. Thermal effects were significant in load testing for the Five Bridges Project. 7. Rebounding or relaxation effects were negligible in load testing for the Five Bridges Project. 8. All five bridges posted five to twenty-five percent increases in apparent stiffness from the first load test to the third load test. The strengthening systems appear to be effective with no loss in serviceability observed after one year. This “apparent” change in stiffness is not attributed to a significant change in the section modulus of structure after strengthening, but rather the engagement of the FRP strengthening to reduce crack width opening under load. 9. FRP Strengthening has little or no effect on load distribution. 10. Load Distribution Coefficients were less than the AASHTO LRFD Load Distribution Values. 11. Even minor misplacement of axles or misreported vehicle loads can result in poor correlation of measured deflections values with predicted values for short span bridges. ACKNOWLEDGEMENTS The authors would like to acknowledge that support for this project was provided by MoDOT and the UTC at UMR. MoDOT also provided the H20 dump trucks and traffic control for load testing. John Clark of Leica Geosystems and Kevin Lindsey of Laser Specialists provided technical support with the Leica TCA 2003 Total Station and surveying equipment. The authors would also like to thank Antonio Nanni, Alexis Lopez-Inojosa, Juan La Gamma-Gianelli, Nestore Galati, Andrea Rizzo, Paolo Casadei, Ursula Deza, and Kah Tan for their assistance during the case studies. The authors would also like to acknowledge and thank the TRB reviewers for their constructive feedback and comments. REFERENCES 1. 2. 3. 4. 5. 6.

AASTHO Guide Specifications. “Thermal Effects in Concrete Bridge Superstructures,” American Association of State Highway and Transportation Officials, Washington, D.C., 1989. AASHTO. “Guide Specifications for Distribution of Loads for Highway Bridges,” American Association of State Highway and Transportation Officials, Washington, D.C., 1994. AASHTO. “LRFD Bridge Design Specifications,” American Association of State Highway and Transportation Officials, Washington, D.C., 1998. AASHTO. “Standard Specification for Highway Bridges,” 17th Ed., American Association of State Highway and Transportation Officials, Washington, D.C., 2002. AASHTO. “Manual for Condition Evaluation and Load Resistance Factor Rating of Highway Bridges,” 2nd Ed., American Association of State Highway and Transportation Officials, Washington, D.C., 2003. ACI Committee 318-02. “Building Code Requirements for Structural Concrete and Commentary (ACI 318R02),” American Concrete Institute, Farmington Hills, MI, 2002.



Merkle, W. and Myers, J.J. 7. 8. 9. 10.

11. 12. 13.

14. 15. 16.

17. 18. 19. 20. 21.

22. 23.

11

ACI 437R-03. “Strength Evaluation of Existing Concrete Buildings,” American Concrete Institute, Farmington Hills, MI, 2002. ACI 440.2R-02. “Guide for the Design and Construction of Externally Bonded FRP Systems for Strengthening Concrete Structures,” American Concrete Institute, Farmington Hills, MI, 2002. ANSYS University Advanced Version 7.1. Finite Element Modeling Software, www.ansys.com, May 20, 2004. Benmokrane, B. Masmoudi, R. Chekired, M. Rahman, H. Debbache, Z. and Tadros, G. “Design, Construction, and Monitoring of Fiber Reinforced Polymer Reinforced Concrete Bridge Deck” ACI Special Publication, Vol. 188, August 1999. “Bridge Inspection Rating Manual,” Missouri Department of Transportation, www.modot.org, May 20, 2004. Galati, N., Casadei, P., and Nanni, A. “Strengthening of Martin Springs Outer Road Bridge” Center for Infrastructure Engineering Studies, University of Missouri at Rolla, No. 04-47, 2003. Koenigsfeld, D., and Myers, J. J. “Secondary Reinforcement for Fiber Reinforced Polymers Reinforced Concrete Panels,” Center for Infrastructure Engineering Studies, University of Missouri at Rolla, No. 03-45, 2003. Leica Geosystems. Leica TCA 2003 Total Station, www.leica-geosystems.com, May 20, 2004. Merkle, W.J. and Myers, J.J. “Load Distribution Response of Bridges Retrofitted with Fiber Reinforced Systems,” Center for Infrastructure Engineering Studies, University of Missouri at Rolla, 2004a. Merkle, W.J., and Myers, J.J., “Use of the Total Station for Load Testing of Retrofitted Bridges with Limited Access,” Proceedings of SPIE – Smart Structures and Materials: Sensors and Smart Technologies for Civil, Mechanical, and Aerospace Systems, Vol. 5391 pp.687-694, 2004b. “National Bridge Inventory,” U.S. Department of Transportation, Federal Highway Administration, Washington, D.C., 2003. Neely, W.D. “Evaluation of In-Service Performance of the Tom’s Creek Bridge,” MS Thesis, Virginia Polytechnic Institute and State University, 2000. Roberts, G.W., Meng, X. and Dodson, A.H. “Integrating a Global Positioning System and Accelerometers to Monitor the Deflection of Bridges,” ASCE Journal of Surveying Engineering, May 2004. Stone, D.K., Tumialan, J.G., Parretti, R., and Nanni, A. “Near-Surface Mounted FRP Reinforcement: Application of an Emerging Technology,” Concrete UK, Vol. 36, No. 5, 2002. Vurpillot, S., Krueger, G., Benouaich, D., Clement, D. and Inaudi, D. “Vertical Deflection of a Pre-Stressed Concrete Bridge Obtained Using Deformation Sensors and Inclinometer Measurements,” ACI Structural Journal, Vol. 95, No. 5, September 1998. “Preservation of Missouri Transportation Infrastructure: Validation of FRP Composite Technology Through Field Testing – Volume I”, www.cies.umr.edu/reports, May 2004. Yang, Y. and Myers, J.J. “Live Load Test Results of Missouri’s First High Performance Concrete Superstructure Bridge,” Proceedings for the Transportation Research Board 82nd Annual Meeting, Washington, D.C., 2003.

TABLES AND FIGURES List of Tables: Table 1 – FRP bridge flexural and shear strengthening types and location with analytical capacity improvement. Table 2 – Bridge designation and span length load testing information. Table 3 – Load testing schedule. Table 4 – Truck axle loads. Table 5 – Theoretical and experimental deflections for Bridge X-596. Table 6 – Theoretical and experimental deflections for Bridge T-530. Table 7 – Theoretical and experimental deflections for Bridge X-495. Table 8 – Theoretical and experimental deflections for Bridge P-962. Table 9 – MoDOT vs. load test ratings. Table 10 – Avg. deflection change from load test 1 to 2. Table 11 – Avg. deflection change from load test 1 to 2 and 1 to 3. Table 12 – AASHTO LRFD load distribution factors [4]. Table 13 – Load distribution coefficients – before strengthening results.




12

Table 14 – Percent change in load distribution coefficients from load test 1 to 2. Table 15 – Percent change in load distribution coefficients from load test 1 to 3. Table 16 – Load testing sensitivity for an individual girder. List of Figures: Fig. 1 – Tested Span of Bridge X-596. Fig. 2 – Tested Span of Bridge T-530. Fig. 3 – Tested Span of Bridge X-495. Fig. 4 – Side view of Bridge P-962. Fig. 5 – Side view of Bridge Y-298. Fig. 6 – Leica TCA 2003. Fig. 7 – Reference point. Fig. 8 – Target (Prism) on a bridge. Fig. 9 – Truck stops – Bridge X-596 Fig. 10 – Illustration of basic calculations. Fig. 11 – Example of significant thermal effects during testing. (Bridge X-596 tested on Aug. 19, 2003 – sunny day – change in gradient: 10°F) Fig. 12 – Example of non-significant thermal effects during testing. (Bridge X-596 tested on Nov. 10, 2003 – cloudy day – change in gradient: 1°F) Fig. 13 – Example deformation contour plot. Fig. 14 – Bridge Y-298 west span. Fig. 15 – Bridge Y-298 east span. Fig. 16 – MoDOT inspection rating vs. percent error from theory. Fig. 17 – Bridge P-962 longitudinal deflection – interior girder – span 1 – truck stop 2. Fig. 18 – Bridge P-962 transverse deflection at midspan – span 1 – truck stop 2. Fig. 19 – Bridge T-530 longitudinal deflection – interior girder – truck stop 2. Fig. 20 – Bridge T-530 transverse deflection at midspan – truck stop 2.




13

Table 1 – FRP bridge flexural and shear strengthening types and location with analytical capacity improvement. Span #

Girder

Flexural Reinforcing Description (Girder)

Capacity Increase

X-596

g

2

Interior

ML: 4 Plies 20" Wide; NSM Bars: 4 total

42%

X-596

g

2

Exterior

None

NA

X-596

g

1, 3

Interior


44%

X-596

g

1, 3

Exterior

ML: 2 Plies 16" Wide

16%

g

1, 3, 5

Interior


29%

T-530g

1, 3, 5

Exterior


15%

T-530

g

2, 4

Interior

1 Laminate Plate: 12" Wide

29%

T-530

g

2, 4

Exterior

1 Laminate Plate: 12" Wide

15%

X-495

g

2

Interior


40%

X-495

g

2

Exterior

None

NA

X-495

g

1, 3

Interior


44%

X-495

g

1, 3

Exterior


16%

P-962

g

1, 2

Interior

ML: 5 Plies 16" Wide plus 4 NSM Bars

56%

P-962

g

1, 2

Exterior


25%

P-962

g

3

Interior

SRP 3X2: 3 Plies 16" Wide

54%

P-962

g

3

Exterior

SRP 3X2: 2 Plies 16" Wide

49%

Span #

Type

Flexural Reinforcing Description (Slab)

Capacity Increase

2

NSM Tape

2 per Groove at 12" O/C

78%

1, 3 1, 3, 5

Manual Layup Manual Layup

1 ply 6" Wide at 15" O/C 1 Ply 9" Wide at 15" O/C

61% 141%

Bridge ID

T-530

Bridge ID X-596

s

X-596

s

T-530

s

T-530s

2, 4

Laminate Plates

1 Plate 3" Wide at 15" O/C

143%

X-495

s

1, 2, 3

Manual Layup

1 ply 6" Wide at 14" O/C

65%

Y-298

s

1, 2

Manual Layup

2 Plies 8" Wide at 12" O/C

23%

P-962

s

3

SRP 3X2

1 Ply 4" Wide at 20" O/C

62%

P-962

s

1, 2

Manual Layup

1 Ply 6" Wide at 14" O/C

64%

Span #

Girder

Shear Reinforcing Description (Girder)

Capacity Increase

2

Interior

ML: 1 Ply Continuous U-Wrap

26%

Bridge ID X-596

g

X-596

g

2

Exterior

None

NA

X-596g

1, 3

Interior


52%

X-596g

1, 3

Exterior

None

NA

X-495g

2

Interior


30%

X-495g

2

Exterior

None

N/A

1, 3

Interior

ML: 2 Plies Continuous U-Wrap

51%

1, 3

Exterior

ML: 1 Ply 12" Wide U-Wrap at 24" O/C

18%

1, 2

Interior

ML: 4 Plies Continuous U-Wrap

64%

1, 2

Exterior


24%

X-495

g

X-495

g

P-962

g

P-962g P-962g

3 Interior SRP 3SX: 3 Plies Continuous U-Wrap 63% P-962g 3 Exterior SRP 3SX: 1 Ply Continuous U-Wrap 36% s g Key: – slab/deck strengthening; – girder strengthening; ML – manual layup; NSM – near surface mounted bar; SRP – steel reinforced polymer;




14

Table 2 – Bridge designation and span length load testing information. Bridge ID

Load Testing Span (total # spans)

Span #1

Span #2

Span #3

Span #4

Span #5

X-596*

Span #2 (3)

42.5

52.5

42.5

NA

NA

T-530*

Span #3 (5)

47.5

47.5

47.5

47.5

47.5

X-495*

Span #2 (3)

42.5

52.5

42.5

NA

NA

Y-298*

Spans #1 & 2 (2)

15

15

NA

NA

NA

P-962*

Span #2 (3)

Span Length (feet)

42.5 42.5 NA NA 42.5 NA - Not Applicable; Conversion factor: 1 m = 3.281 feet; * - simply supported bridge.

Table 3 – Load testing schedule. Load Test 1

Load Test 2

Load Test 3

Before Strengthening

After Strengthening

After Strengthening

X-596

8/19/2003

11/10/2003

5/06/2004

T-530

7/09/2003

11/03/2003

5/12/2004

X-495

7/08/2003

10/29/2003

5/04/2004

P-962

8/14/2003

11/17/2003

5/05/2004

Y-298

10/15/2003

11/12/2003

5/13/2004

Bridge ID

Table 4 – Truck axle loads.

After Strengthening

Before Strengthening

Bridge ID X-596 T-530 X-495 P-962 Y-298 X-596 T-530 X-495 P-962 Y-298

Truck 1 (kips) Truck 2 (kips) Rear Axles Front Axle Total Rear Axles Front Axle 41.03 17.59 58.62 40.01 17.15 40.90 16.80 57.70 38.32 16.40 42.32 12.76 55.08 42.48 12.72 38.72 13.36 52.08 43.04 17.22 25.48 11.35 36.83 25.48 11.35 42.26 21.50 63.76 45.58 19.54 39.95 16.27 56.22 38.07 16.31 38.15 16.35 54.50 34.83 14.93 40.72 17.16 57.88 41.72 12.72 12.78 35.04 47.82 15.58 39.60 Conversions Units: 1 kip = 453.6 kg


Total 57.16 54.72 55.20 60.26 36.83 65.12 54.38 49.76 54.44 55.18



15

Table 5 – Theoretical and experimental deflections for Bridge X-596. Truck Stop Girder Measured Deflection Mass-Section Analysis Barrier Wall 0% Mass-Section Analysis Barrier Wall 50% Mass-Section Analysis Barrier Wall 100% FEM No Barrier Wall FEM with Barrier Wall

2 2 Exterior Interior 0.125 0.147 0.201 48% 0.164 21% 0.138 1% 0.195 0.229 56% 56% 0.095 0.102 -24% -31%

4 4 Exterior Interior 0.151 0.125 0.176 28% 0.143 4% 0.121 -12% 0.279 0.195 85% 56% 0.109 0.084 -28% -33%


Table 6 – Theoretical and experimental deflections for Bridge T-530. Truck Stop Girder Measured Deflection Mass-Section Analysis Barrier Wall 0% Mass-Section Analysis Barrier Wall 50% Mass-Section Analysis Barrier Wall 100% FEM No Barrier Wall FEM with Barrier Wall

2 2 Exterior Interior 0.089 0.109 0.145 46% 0.117 18% 0.097 -2% 0.145 0.165 63% 51% 0.093 0.135 4% 24%

4 4 Exterior Interior 0.113 0.114 0.124 9% 0.099 -13% 0.083 -27% 0.218 0.19 93% 67% 0.14 0.145 24% 27%

5 5 Exterior Interior 0.075 0.112 0.124 33% 0.099 6% 0.087 -7% 0.11 0.161 47% 44% 0.075 0.135 0% 21%

Table 7 – Theoretical and experimental deflections for Bridge X-495. Truck Stop Girder Measured Deflection Mass-Section Analysis Barrier Wall 0% Mass-Section Analysis Barrier Wall 50% Mass-Section Analysis Barrier Wall 100% FEM No Barrier Wall FEM with Barrier Wall


2 Exterior 0.088

2 Interior 0.105

0.2 107% 0.164 70% 0.138 43% 0.201 128% 0.09 2%

0.242 130% 0.096 -9%





16

Table 8 – Theoretical and experimental deflections for Bridge P-962. Truck Stop Girder Measured Deflection Mass-Section Analysis Barrier Wall 0% Mass-Section Analysis Barrier Wall 50% Mass-Section Analysis Barrier Wall 100% FEM No Barrier Wall


FEM with Barrier Wall


5 5 Exterior Interior 0.109 0.197 0.288 88% 0.114 -25% 0.071 -54% 0.197 0.278 81% 41% 0.095 0.203 -13% 3%

Table 9 – MoDOT vs. load test ratings. Bridge ID

MoDOT Rating

Load Test Rating

X-596

5.5

5.5

T-530

5.0

5.6

X-495

6.0

5.9

P-962

6.5

5.6

Y-298

5.0

5.0

Table 10 – Average deflection change from load test 1 to 2. Deflection Change Bridge ID

Test 1 to Test 2

X-596

-2%

T-530

-6%

X-495

+ 8%

P-962

-6%

Y-298

-21%

Table 11 – Average deflection change from load test 1 to 2 and 1 to 3. Deflection Change


Bridge ID

Test 1 to 2

Test 1 to 3

X-596

-2 %

-15 %

T-530

-6 %

-16 %

X-495

8%

-16 %

P-962

-6 %

-7 %

Y-298

-21 %

-25 %



17

Table 12 – AASHTO LRFD load distribution factors [4].

X-596 T-530 X-495 P-962

Exterior Girder

Interior Girder

1 Lane Loaded

0.48

0.60

2 Lanes Loaded

0.64

0.81

1 Lane Loaded

0.35

0.49

2 Lanes Loaded

0.46

0.64

1 Lane Loaded

0.48

0.60

2 Lanes Loaded

0.64

0.81

1 Lane Loaded

0.51

0.64

2 Lanes Loaded

0.68

0.86

Table 13 – Load distribution coefficients – before strengthening results.

X-596

T-530

X-495

P-962

Girder:

A

B

C

D

Stop 2

0.30

0.38

0.32

Stop 4

0.44

0.37

0.19

Stop 5

0.29

0.38

0.33

Stop 2

0.19

0.25

0.29

0.26

Stop 4

0.30

0.29

0.27

0.13

Stop 5

0.18

0.28

0.32

0.22

Stop 2

0.31

0.34

0.34

Stop 4

0.47

0.35

0.19

Stop 5

0.29

0.38

0.34

Stop 2

0.28

0.40

0.33

Stop 4

0.40

0.40

0.20

Stop 5

0.24

0.44

0.32

Stop 6

0.31

0.41

0.28

Stop 7

0.31

0.41

0.28

Table 14 – Percent change in load distribution coefficients from load test 1 to 2. Girder

A

B

C

Bridge X-596

-2

4

-1

Bridge T-530

-1

-3

2

Bridge X-495

1

-3

5

Bridge P-962

3

1

-5

D 2

Table 15 – Percent change in load distribution coefficients from load test 1 to 3. Girder

A

B

C

Bridge X-596

1

-3

3

Bridge T-530

-1

-5

5

Bridge X-495

-4

4

1

Bridge P-962

-2

3

-4


D 4



18

Table 16 – Load testing sensitivity for an individual girder. Error

Deflection Change

Truck Placed ± 12" Transversally

10%

Truck Placed ± 12" Longitudinally

5%

Truck Reported Weight ± 5 kips Actual Weight

7%

Surveying Equipment

4%

Fig. 1 –. Tested Span of Bridge X-596

Fig. 2 – Tested Span of Bridge T-530.

Fig. 3 – Tested Span of Bridge X-495.

Fig. 4 – Side view of Bridge P-962.

Fig. 5 – Side view of Bridge Y-298.




Fig. 6 – Leica TCA 2003.

19

Fig. 7 – Reference point.

Fig. 8 – Target (Prism) on a bridge.

Fig. 9 – Truck stops – Bridge X-596




20

99.636 99.635 99.639 99.381 99.381 99.384

Control Readings Test Readings

99.637

99.382

Deflection = 0.255 Fig. 10 – Illustration of basic calculations. 0.15

0.08 0.06

Deflection (in)

Deflection (in)

0.12

0.09

0.06

0.03

0.04 0.02 0.00 -0.02

0.00

-0.04 0

5

10

15

20

25

30

35

40

45

50

Span Length (ft) Exterior Girder (A)

Interior Girder (B)

0

5

10

15

20

25

30

35

40

45

50

Span Length (ft) Exterior Girder (A)

Interior Girder (B)

Conversion Units: 1 in = 25.4 mm; 1 ft = 0.3048 m


Fig. 11 – Example of significant thermal effects during testing. (Bridge X-596 tested on Aug. 19, 2003 – sunny day – change in gradient: 10°F)

Fig. 12 – Example of non-significant thermal effects during testing. (Bridge X-596 tested on Nov. 10, 2003 – cloudy day – change in gradient: 1°F)

Fig. 13 – Example deformation contour plot.



Deflection (in)

Deflection (in)

0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040

0.000 0.010 0.020 0.030 0.040 0

0

10

20

30

10

40

Load Test Stop 2

Theory - Stop 1

30

40

Span Width (ft)

Span Width (ft) Load Test Stop 1

20

Theory - Stop 2

Load Test Stop 1

Load Test Stop 2

Theory - Stop 1

Theory - Stop 2



Fig. 14 – Bridge Y-298 west span.

Fig. 15 – Bridge Y-298 east span.

7.5

MoDOT Rating

7.0 6.5 6.0 5.5 5.0 4.5 4.0 -0.20

0.00

0.20

0.40

0.60

% Error Theory (1/100)

Fig. 16 – MoDOT inspection rating vs. percent error from theory. 0.00

Deflection (in)

0.03 0.06 0.09 0.12 0.15 0.18 0.21 0

5

10

15

20

25

30

35

40

Span Length (ft)

Test 1

Test 2

Test 3


Fig. 17 – Bridge P-962 longitudinal deflection – interior girder – span 1 – truck stop 2.




22

0.00

Deflection (in)

0.03 0.06 0.09 0.12 0.15 0.18 0.21 0

3

6

9

12

15

18

21

24

Span Width (ft)

Test 1

Test 2

Test 3


Fig. 18 – Bridge P-962 transverse deflection at midspan – span 1 – truck stop 2.

Deflection (in)

0.00 0.03 0.06 0.09 0.12 0.15 0

5

10

15

20

25

30

35

40

45

Span Length (ft)

Test 1

Test 2

Test 3


Fig. 19 – Bridge T-530 longitudinal deflection – interior girder – truck stop 2.

Deflection (in)

0.00 0.03 0.06 0.09 0.12 0.15 0

3

6

9

12

15

18

21

24

Span Width (ft)

Test 1

Test 2

Test 3


Fig. 20 – Bridge T-530 transverse deflection at midspan – truck stop 2.

