Recent bridge failures have shoved bridge design and analysis to the forefront.
Engineers historically “overdesigned” bridges using substantial margins of safety to compensate for unknown forces that could affect a structure over time. To help ensure future bridges will be economical, as well as reliable and easy to maintain, researchers at Penn State University use Abaqus/Standard software from Simulia in Providence, R.I., to model and analyze virtual bridges. This lets them solve design problems before any concrete is poured or spans erected.
“Until recently, civil engineering relied on linear-elastic, small-deflection FEA as the backbone of bridge analysis,” says Civil and Engineering Associate Professor Daniel Linzell. “But, in many instances, it is an approximation. It doesn’t capture the full range of real-world nonlinear responses in large and complicated structures. So higher-order FEA software is necessary.”
Such software, for example, incorporates responses of steel and concrete to the weight of the bridge, as well as to wind, water, traffic, corrosion, temperature, and even time. (Concrete and steel creep, resulting in long-term deformation). Of course the software also accounts for traditional parameters such as material and geometric nonlinearities.
The software closely predicts nonlinear deformation of bridge materials. For example, steel has yield stresses of 100 ksi, almost three times that of steel 10 to 15 years ago. The new, stronger steel is lighter, but also more flexible.
Linzell says his group uses Abaqus to model existing designs, which can be in CAD formats, on paper, or based on as-built measurements taken from structures under construction. They then choose the material model, say, concrete or steel, and an element type, such as 3D triangles or rectangles, depending on the material model and geometry.
Next, the group establishes boundary conditions. “This entails selecting how the area of the bridge under consideration is restrained, whether with a contact condition or discrete restraint, and how friction is represented,” says Linzell. “We apply forces to the bridge deck to represent vehicle loads and to beam faces to represent wind loads. We also account for thermal loads such as temperature changes during construction and use the Abaqus creep module for time-dependent factors. The software calculates stresses at the nodes or in the elements.”
The group modifies bridge models where loads are excessive and repeats the process until an acceptable configuration is reached. “The software can also evaluate residual life of structures which are still standing but may be cracked,” says Linzell. “And it can pinpoint reasons for collapse as well.”