Evaluation of stability and integrity of a steel truss bridge in a forensic investigation

The studies presented in this Thesis have been developed in the frame of the forensic investigation into the causes of the collapse of the I-35West Bridge (I-35W) in Minneapolis, Minnesota, USA that occurred on August 1 st, 2007. The failure of the I-35W represents a major case-study for the evaluation of stability and integrity of a steel truss bridge. The Thesis has been developed at Columbia University and at the engineering firm Thornton Tomasetti (TT) which was hired by a national law firm, Robins, Kaplan, Miller & Ciresi, to perform a forensic investigation into the cause of the catastrophic collapse. According to the findings of the forensic investigation, the collapse was triggered by the buckling of an element of the main truss bottom chord in the main span close to the pier. The Thesis focused on technical aspects and did not attempt to assign responsibility among the involved parties. In the first part of the thesis, the background and motivation for the forensic investigation are presented together with a description of the I-35W Bridge. The definition of bridge safety and related classifications are given. The concept of structural stability and integrity of steel structures are discussed. The nature of structures and their complexity are considered as well as the methodologies used to study them. An extensive description of the structural decomposition method is presented and detailed for the case study. In this work, using the framework of a multilevel approach, the structural system has been broken down in order to perform a detailed analysis and evaluate the system performances at macro (global) and micro (local) levels. The effect of boundary conditions, thermal loads on the global system and post buckling capacity of the main truss bottom chord built up member on a local level have been studied. First, a 3D finite element model has been developed in SAP2000 using frame elements. This global-level model reproduces the entire bridge based on original drawings, design and construction specifications. The model has been verified by comparing results with the available original design calculations. Member forces and reactions based on the asdesigned conditions with the specified design loads have been confirmed. The model served to investigate the elastic behavior of the bridge and its overall response to various loading and boundary conditions. In particular, from the global model it has been possible to evaluate the static stress condition on the bridge showing how some of the temperature changes and the possible deterioration of the designed supports could affect the demand on the load carrying members. A specific lower chord member was identified as a critical member for temperature loading in particular. Second, a 3D solid element model of the recognized critical load bearing member comprised of a welded built up section with perforated cover plates was built in Abaqus. This local-level model provided information on the post buckling behavior and capacity of the load bearing member. The effects of the perforations and boundary conditions have been outlined. Furthermore, the results have been compared against hand calculations following the provisions of the Code of Standard Practice for Structural Steel Buildings and Bridges (AISC, 2005) for built up members and the Timoshenko plate theory for columns with perforated cover plates.