Pipeline Response to Ground Oscillation During Earthquakes

Laura Barton

 

            During earthquakes, pipelines can be affected by both permanent ground deformation (PGD) and transient ground deformation (TGD).  PGD, one of the most prominent causes of pipeline damage, can be the result of the effects of liquefied soil.    Liquefaction can also promote transient effects that may result in ground deformation and pipeline strains substantially larger than those generated by traveling ground waves (TGW). 

            An important source of liquefaction-induced TGD is referred to as ground oscillation.  The Committee on Earthquake Engineering (1985) identified ground oscillation as "a source of TGD associated with liquefaction in areas of virtually level ground where near-surface soils oscillate on top of an underlying liquefied layer."

            Problems arise at the interface between liquefied and non-liquefied soils.  When liquefiable soils are bounded by non-liquefiable soils, the resulting area may be described as a basin.  Within the basin, lateral surface displacements are relatively constant, and consequentially stresses and strains on buried pipelines are also small.  At the interface with a non-liquefied soil, however, the displacements are often larger, creating damaging zones of compression and tension on pipelines.

            The liquefiable soil oscillates below the water table, consequentially oscillating the ground above it.  The ground behaves as though it is floating in the liquefiable soil, and it is being pushed into and pulled away from the rest of the ground that is not under the liquefiable soil. 

            The purpose of this research is to characterize pipeline response to ground oscillation and to use these characterizations to aid in the design of pipeline systems. 

 

Research Methods:

            A basin was modeled in ABAQUS, a finite element software package.  Pipe elements were used to model the pipe, and spring slider elements were used to model the soil-pipe interaction.  A schematic of this model is shown below.  When the soil around a pipe moves, the pipe resists the movement.  This mobilizes the force around the pipe by local soil deformations.  These local soil deformations are what are being modeled by the springs.  The nodes on the spring slider elements were given a far field displacement to represent the first wave of an oscillation.  This far field displacement is independent of the pipe structure. 

            The model was simulated under four conditions: elastic pipe material, bilinear shear transfer; elastic pipe material, perfectly plastic shear transfer; plastic pipe material, bilinear shear transfer; and plastic pipe material, perfectly plastic shear transfer.

            A closed form solution was derived for the ideal case of elastic material and perfectly plastic shear transfer.  Based on this closed form solution and the ABAQUS results for more realistic solutions, it is possible to start to see how the real situations differ from ideal and what adjustments may have to be made in order to design for real life situations. 

            The motivation for these studies is to ultimately provide a design tool for engineers to use to predict the demands that will be based upon pipeline systems.  This problem has not been extensively studied.  In the big picture, these preliminary studies may be used to determine mathematical relationships between all the parameters of this study.  This will enable engineers to predict actual values of strains on pipes by knowing the strain for an ideal condition and adjusting it to suit a real condition by various factors that will come out of these mathematical relationships.