skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Thermal-Structural Analysis of the MacArthur Maze Freeway Collapse

Conference ·

At approximately 3:41 AM on the morning of April 29, 2007, a tractor-trailer rig carrying 8,600 gallons (32.6 m{sup 3}) of fuel overturned on Interstate 880 in Oakland, CA. The resultant fire weakened the surrounding steel superstructure and caused a 50-yard (45.7 m) long section of the above connecting ramp from Interstate 80 to Interstate 580 to fail in approximately 18 minutes. In this study, we performed a loosely-coupled thermal-structural finite element analysis of the freeway using the LLNL Engineering codes NIKE3D, DYNA3D and TOPAZ3D. First, we applied an implicit structural code to statically initialize the stresses and displacements in the roadway at ambient conditions due to gravity loading. Next, we performed a thermal analysis by approximating the tanker fire as a moving box region of uniform temperature. This approach allowed for feasible calculation of the fire-to-structure radiative view factors and convective heat transport. We used a mass scaling methodology in the thermal analysis to reduce the overall simulation time so an explicit structural analysis could be used, which provided a more computationally efficient simulation of structural failure. Our approach showed structural failure of both spans due to thermal softening under gravity loading at approximately 20 minutes for a fixed fire temperature of 1200 C and fixed thermal properties. When temperature-dependent thermal properties were applied, the south and north spans collapsed at approximately 10 minutes and 16 minutes, respectively. Finally, we performed a preliminary fully-coupled analysis of the system using the new LLNL implicit multi-mechanics code Diablo. Our investigation shows that our approach provides a reasonable first-order analysis of the system, but improved modeling of the transport properties and the girder-box beam connections is required for more accurate predictions.

Research Organization:
Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
Sponsoring Organization:
USDOE
DOE Contract Number:
W-7405-ENG-48
OSTI ID:
944376
Report Number(s):
LLNL-PROC-401748; TRN: US200902%%735
Resource Relation:
Conference: Presented at: ASME 2008 Summer Heat Transfer Conference, Jacksonville, FL, United States, Aug 10 - Aug 14, 2008
Country of Publication:
United States
Language:
English