FINAL CONTRACT REPORT VTRC 09-CR8 LIVE LOAD TEST AND FAILURE ANALYSIS FOR THE STEEL DECK TRUSS BRIDGE OVER THE NEW RIVER IN VIRGINIA LUCAS HICKEY Graduate Research Engineer CARIN ROBERTS-WOLLMANN, Ph.D., P.E. Associate Professor THOMAS COUSINS, Ph.D., P.E. Professor ELISA SOTELINO, Ph.D. Professor W. SAMUEL EASTERLING, Ph.D., P.E. Professor Via Department of Civil and Environmental Engineering Virginia Polytechnic Institute & State University http://www.virginiadot.org/vtrc/main/online_reports/pdf/09-cr8.pdf Standard Title Page - Report on Federally Funded Project 1. Report No.: 2. Government Accession No.: 3. Recipient’s Catalog No.: FHWA/VTRC 09-CR8 4. Title and Subtitle: 5. Report Date: Live Load Test and Failure Analysis for the Steel Deck Truss Bridge Over the New May 2009 River in Virginia 6. Performing Organization Code: 7. Author(s): 8. Performing Organization Report No.: Lucas Hickey, Carin Roberts-Wollmann, Thomas Cousins, Elisa Sotelino, and W. VTRC 09-CR8 Samuel Easterling 9. Performing Organization and Address: 10. Work Unit No. (TRAIS): Virginia Transportation Research Council 530 Edgemont Road 11. Contract or Grant No.: Charlottesville, VA 22903 87975 12. Sponsoring Agencies’ Name and Address: 13. Type of Report and Period Covered: Virginia Department of Transportation Federal Highway Administration Final Contract 1401 E. Broad Street 400 North 8th Street, Room 750 14. Sponsoring Agency Code: Richmond, VA 23219 Richmond, VA 23219-4825 15. Supplementary Notes: This project was financed with federal grant funds at an estimated cost of $89,000. 16. Abstract: This report presents the methods used to model a steel deck truss bridge over the New River in Hillsville, Virginia. These methods were evaluated by comparing analytical results with data recorded from 14 members during live load testing. The research presented herein is part of a larger endeavor to understand the structural behavior and collapse mechanism of the erstwhile I-35W bridge in Minneapolis, Minnesota, that collapsed on August 1, 2007. Objectives accomplished toward this end include investigation of lacing effects on built-up member strain measurement, live load testing of a steel truss bridge, and evaluation of modeling techniques in comparison to recorded data. The most accurate model was used to conduct a failure analysis with the intent of then loading the steel truss bridge to failure. Before any live load testing could be performed, it was necessary to confirm an acceptable strain gage layout for measuring member strains. The effect of riveted lacing in built-up members was investigated by constructing a two-thirds mockup of a typical bridge member. The mockup was instrumented with strain gages and subjected to known loads to determine the most effective strain gage arrangement. The results of the testing analysis showed that for a built-up member consisting of laced channels, one strain gage installed on the middle of the extreme fiber of each channel’s flanges was sufficient. Thus, laced members on the bridge were mounted with four strain gages each. Data from live loads were obtained by loading two trucks to 25 tons each. Trucks were positioned at eight locations on the bridge in four different relative truck positions. Data were recorded continuously and reduced to member forces for model validation comparisons. Deflections at selected truss nodes were also recorded for model validation purposes. The model validation process began by developing four simple truss models, each reflecting different expected restraint conditions, in the hopes of bracketing data from recorded results. The models included a simple truss model, a frame model with only the truss members, and a frame model that included the stringers. The final, most accurate model was selected and used for a failure analysis. This model showed where the minimum amount of load could be applied to learn about the bridge’s failure behavior and was to be used for a test to be conducted at a later time. Unfortunately, the project was terminated because of a lack of funding before the actual test to failure of the steel truss bridge was conducted. Nevertheless, findings from the study led to two important recommendations: 1. When instrumenting a steel truss bridge for load testing by placing strain gages on built-up members, four gages, one placed on each flange of each channel, should be used. 2. When modeling deck truss bridges, the system should be considered to be a frame and should include the stringers in the model. 17 Key Words: 18. Distribution Statement: Steel deck truss bridge, live load tests, laced members No restrictions. This document is available to the public through NTIS, Springfield, VA 22161. 19. Security Classif. (of this report): 20. Security Classif. (of this page): 21. No. of Pages: 22. Price: Unclassified Unclassified 97 Form DOT F 1700.7 (8-72) Reproduction of completed page authorized DISCLAIMER The project that is the subject of this report was done under contract for the Virginia Department of Transportation, Virginia Transportation Research Council. The contents of this report reflect the views of the author(s), who is responsible for the facts and the accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the Virginia Department of Transportation, the Commonwealth Transportation Board, or the Federal Highway Administration. This report does not constitute a standard, specification, or regulation. Any inclusion of manufacturer names, trade names, or trademarks is for identification purposes only and is not to be considered an endorsement. Each contract report is peer reviewed and accepted for publication by Research Council staff with expertise in related technical areas. Final editing and proofreading of the report are performed by the contractor. Copyright 2009 by the Commonwealth of Virginia. All rights reserved. i i ABSTRACT This report presents the methods used to model a steel deck truss bridge over the New River in Hillsville, Virginia. These methods were evaluated by comparing analytical results with data recorded from 14 members during live load testing. The research presented herein is part of a larger endeavor to understand the structural behavior and collapse mechanism of the erstwhile I-35W bridge in Minneapolis, Minnesota, that collapsed on August 1, 2007. Objectives accomplished toward this end include investigation of lacing effects on built-up member strain measurement, live load testing of a steel truss bridge, and evaluation of modeling techniques in comparison to recorded data. The most accurate model was used to conduct a failure analysis with the intent of then loading the steel truss bridge to failure. Before any live load testing could be performed, it was necessary to confirm an acceptable strain gage layout for measuring member strains. The effect of riveted lacing in built- up members was investigated by constructing a two-thirds mockup of a typical bridge member. The mockup was instrumented with strain gages and subjected to known loads to determine the most effective strain gage arrangement. The results of the testing analysis showed that for a built-up member consisting of laced channels, one strain gage installed on the middle of the extreme fiber of each channel’s flanges was sufficient. Thus, laced members on the bridge were mounted with four strain gages each. Data from live loads were obtained by loading two trucks to 25 tons each. Trucks were positioned at eight locations on the bridge in four different relative truck positions. Data were recorded continuously and reduced to member forces for model validation comparisons. Deflections at selected truss nodes were also recorded for model validation purposes. The model validation process began by developing four simple truss models, each reflecting different expected restraint conditions, in the hopes of bracketing data from recorded results. The models included a simple truss model, a frame model with only the truss members, and a frame model that included the stringers. The final, most accurate model was selected and used for a failure analysis. This model showed where the minimum amount of load could be applied to learn about the bridge’s failure behavior and was to be used for a test to be conducted at a later time. Unfortunately, the project was terminated because of a lack of funding before the actual test to failure of the steel truss bridge was conducted. Nevertheless, findings from the study led to two important recommendations: 1. When instrumenting a steel truss bridge for load testing by placing strain gages on built-up members, four gages, one placed on each flange of each channel, should be used. 2. When modeling deck truss bridges, the system should be considered to be a frame and should include the stringers in the model. ii i FINAL CONTRACT REPORT LIVE LOAD TEST AND FAILURE ANALYSIS FOR THE STEEL DECK TRUSS BRIDGE OVER THE NEW RIVER IN VIRGINIA Lucas Hickey Graduate Research Engineer Carin Roberts-Wollmann, Ph.D., P.E. Associate Professor Thomas Cousins, Ph.D., P.E. Professor Elisa Sotelino, Ph.D. Professor W. Samuel Easterling, Ph.D., P.E. Professor Via Department of Civil and Environmental Engineering Virginia Polytechnic Institute & State University INTRODUCTION On Wednesday, August 1, 2007, the I-35W bridge in Minneapolis, Minnesota, collapsed into the Mississippi River. The bridge was an eight-lane, arched deck truss bridge made of steel. It had opened in 1967 with a total length of 1,907 ft and a maximum span of 456 ft. Figure 1 shows the steel arch and concrete deck spanning the Mississippi. The collapse occurred while the bridge was carrying standard vehicular loads on a warm August afternoon. Steel deck truss bridges are abundant in the bridge inventories of almost all state departments of transportation; hence, it is important to understand the causes of this failure. Virginia Route 100 crosses the New River between Pulaski and Hillsville. The bridge over the river is a steel deck truss bridge, constructed in 1941 with only two lanes. Its total length is 846 ft, 300 ft of which is spanned by the deck truss. Figure 2 shows the Hillsville truss. The Federal Highway Administration (FHWA) and the Virginia Department of Transportation (VDOT) believed that results from tests of this bridge would be useful for understanding the behavior of older steel deck truss bridges and would provide insight into the cause of the collapse of the I-35W bridge. The research presented herein is a concerted attempt to approximate mathematically the behavior of the Hillsdale Bridge under load so that a plan for a test to failure could be developed. This imposed failure was expected to provide valuable information to transportation authorities with regard to the evaluation and future design of steel deck truss bridges. Figure 1. I-35W Truss Figure 2. Hillsville Truss 2 PURPOSE AND SCOPE The purpose of this study was to conduct a field test and detailed structural analysis of a steel truss bridge over the New River in Hillsville, Virginia, to allow a better understanding of the structural behavior and collapse mechanism of the I-35W Bridge in Minneapolis, Minnesota. The results of the field test were used to calibrate structural models of the steel truss bridge. Once calibrated, the model was used to conduct a failure analysis. The objective was then to test the Hillsdale Bridge to failure; however, funding was not available and the research efforts were terminated. The scope of this report is limited to the field test results, model development and calibration, and failure analysis of the structural model. To understand the bridge’s behavior, the responses of the trusses to live load at various locations needed to be recorded and compared to mathematical models that could approximate the bridge’s behavior. This was accomplished by parking loaded trucks of known weights and dimensions on the bridge and recording strains in selected members. The data obtained from live load testing were used to develop an accurate structural model. This model was then used to find the location and magnitude of load that must be applied to yield a truss member. All models investigated were two dimensional with linear-elastic constitutive laws. A non-linear, three- dimensional model was considered outside the scope of this research. Thus, the objectives of this research were the following: 1. Determine the best method to evaluate member forces in situ. 2. Perform live load tests and gather data. 3. Develop models for comparison to data. 4. Using the best model, determine the load and load position to cause failure. METHODS Laced Member Strain Gaging Axial forces and nodal deflections were used to compare and evaluate structural models. In order to measure the axial force in a bridge member accurately, the average strain must be accurately determined. All the top chord, bottom chord, and diagonal members in the deck trusses are composed of built-up members. These are channel sections of known sizes and properties that are laced together, either singly or doubly, with bars of known size attached with hot rivets. As designed, the channels resist all axial forces while the lacing provides stability and maintains spacing between the channels. In reality, it is important to understand differences in strains read at lacing connection points compared to strains read between them. Knowing which strain gage attachment location on the channels yields the most accurate average strain results is crucial to proper truss instrumentation and force calculation. To accomplish this objective, a two-thirds size mockup of a typical built-up section was tested in the Virginia Tech Structures Lab. It was built of two 4-ft sections of C8 by 13.75 channels doubly 3 laced at five spacings with bars measuring 2 in by ¼ in. The assembly was capped by welding a plate with the dimensions of 12 in by 16 in by ¾ in on each end. Figure 3 presents the tested member and gage installation locations. The column was instrumented with an array of gages. One gage was installed at each web centroid at locations where bolts attached the lacing bars. At sections equidistant between the lacing connections, six gages were attached as shown in Figure 3, Sections A and C. Gages used in this setup were adhered to the steel with a cyanoacrylate adhesive. Gages used for live load testing in the field however, were welded to members, as this is a faster installation. Prior to gages being welded in the field, one was welded to the built-up column to confirm its accuracy. The column was then placed in a SATEC testing machine to apply load and monitor deflection. Figure 3. Column Gaging Diagram Data gathered from column testing were used to investigate three issues associated with attaching strain gages to aging laced structural members: 1. The strain gages mounted to the webs of each section were compared to determine if lacing attachment points affected strain measurements. 4 2. The average strain value taken from the four flange gages in a section were compared to the average strain value of all six gages in a section to determine relative accuracies of the four flange gages to all six. 3. Both of these layouts were compared to the expected strain based on the SATEC load record to determine relative accuracies measured from the strain gages to expected values determined from the SATEC load cell. Truss Instrumentation and Data Collection The result of Item 1 in the preceding list was used to determine the necessary layout of strain gages to be used on each member. Once this was selected, the members to be monitored were chosen. Initial reports after the I-35W Bridge collapse indicated the second top chord truss connection out from the center pier was a crucial failure location. This was confirmed by a preliminary truss analysis. Thus, all members framing into Node U6 were selected for monitoring. On the truss designated “Far” (farther from the new replacement bridge, on the south/downstream side), seven members were instrumented with four strain gages each. Figure 4 presents a plan view of the old bridge and shows its position relative to its replacement structure. The New River and the bridge’s orientation between Pulaski and Hillsville are also shown. Figure 5 shows the member instrumentation for the far truss; all highlighted members have four gages. On the truss designated “Near” (nearer to the new replacement bridge, on the north/upstream side), two members were instrumented with four gages each, and the remaining Figure 4. Plan and Profile View of Bridge 5 Figure 5. Instrumentation of Far Truss. All monitored members have 4 gages. five members were instrumented with one gage each. Figure 6 shows the member instrumentation for the near truss. Boldly highlighted members have four gages each; faintly highlighted members have only one gage each. This was done to save time on gage installation but still provide rough data on the second truss for comparison. On each truss, only members on the Pulaski side of the central pier were instrumented. The assumption of perfect structural symmetry between the Pulaski and Hillsville truss halves was made. Models used to reflect bridge behavior were developed as complete, two-span trusses. Deflectometers, referred to as twangers, were used to measure vertical deflection at Nodes L1 and L6 on each truss as well. These were calibrated at the Virginia Tech Structures Lab and were installed on the day of testing. Figure 7 shows the installation locations for twangers on both trusses. After instrumentation, two trucks of known weights and dimensions were used to impose live load on the bridge. Four load regimes were imposed on the bridge, each moved incrementally from Nodes U1 to U8. The first regime was one truck traveling down the center; the second was one truck traveling down the left (far truss) lane; the third was two trucks side by side; and the fourth was two trucks rear to rear traveling down the left lane. Figures 8 and 9 illustrate the four load regimes. Figure 6. Instrumentation of Near Truss. Only end members have 4 gages. 6