Dynamic Soil-Structure Interaction Analysis Using Ls-Dyna: Dr. Navin Peiris Ove Arup & Partners, London
Dynamic Soil-Structure Interaction Analysis Using Ls-Dyna: Dr. Navin Peiris Ove Arup & Partners, London
Dynamic Soil-Structure Interaction Analysis Using Ls-Dyna: Dr. Navin Peiris Ove Arup & Partners, London
LS-DYNA
Dr. Navin Peiris
Ove Arup & Partners, London
Abstract
A dynamic soil-structure interaction analysis was carried out for the seismic design of a pile foundation, for a
light railway structure at the JFK Airport in New York. The analysis was carried out using the explicit non-linear
finite element program, LS-DYNA. This summary briefly describes the methodology of dynamic soil-structure
interaction and discusses the results of the analysis.
Introduction
The support structure for the light railway system (LRS)
shown in Figure 1 was to be founded on 19.8m long and J ! J'" = J
J
0.46m diameter steel monotube piles. A seismic hazard : i ; ] i
assessment gave a peak ground acceleration of 0.12g for
the project site, which has a 10% probability of exceedence
in a 50 year design life. This meant that the pile
foundation, support structure and the super-structure had to
be designed for seismic loads. The site consists of a 4m
thick layer of hydraulic fill overlying a thin layer of silt,
which is underlain by medium to fine glacial sand, clay
and sand to the bedrock at about 240m below the ground
surface. A liquefaction assessment carried out for the site
revealed that the hydraulic fill layer below the water table
(at 2m below the ground surface) has liquefaction potential
during the design earthquake. For the design of the piles,
Figure 1 LRS support structure
the effect of the presence of a liquefied hydraulic fill layer
was considered since this is likely to be conservative.
Pseudo-static pile analyses were carried out using the program Oasys ALP to determine the bending moment and
shear force distributions for the piles. This was carried out for various combinations of base shear from the
superstructure at the pile head and the kinematic load from the soil (soil displacements) derived from a site
response analysis using the design earthquake. The maximum pile bending moment obtained was 187kNm at the
pile cap level, which is close to the capacity of the pile at 200kNm. The maximum bending moment was
obtained when the base shear and the soil displacements are acting in the same direction on the pile. In order to
model the soil-pile-structure interaction more accurately, a dynamic soil-structure interaction analysis was
carried out.
Monotube
Piles
.000000000
.000000000
Figure 2 LS-DYNA 3D half model Figure 3 Detailed view of the pile foundation
7
Downloaded by [ NEW YORK UNIVERSITY] on [24/03/17]. Copyright © ICE Publishing, all rights reserved.
The soil model used in LS-DYNA describes the hysteretic behviour of soils to cyclic loading with strain
dependent damping (hysteretic). The piles and the supporting column shown in Figure 3 were modelled using
beam elements that can form plastic hinges at their ends. The piles were connected to the soil elements via
transverse and vertical kinematic elastic-plastic springs. The limiting strength of the transverse and vertical
springs represents the passive resistance and the skin friction at the pile/soil interface, respectively. The piles
were fixed rigidly to the pile cap modelled as a rigid material. The elastomeric bearing isolator was modelled as
a linear spring and a dashpot.
The earthquake motion was applied at 29.5m below the ground surface. The ground motion at this level was
obtained from a site response analysis of the entire soil deposit above the bedrock using the program Oasys
SIREN. In addition, free field ground motions derived from this site response analysis were applied at the end
and far side boundaries of the LS-DYNA model. The use of LS-DYNA for DSSI analysis are described in
Lubkowski (1996) and Willford et al., (1996).
Figure 4 Deformation of the soil block Figure 5 Deformation of the piles and the LRS
during the analysis mass during the analysis
of the piles. The comparison of the test results and the Average force pei trailing pile, test results
Average force pei loading pile, sialic push-over analysi:
Oasys T/HIS
Force pet trailing pile, static push-over analysis
analysis are shown in Figure 6 where the solid lines
represent the load test results for the leading and trailing
edge piles. They show good correspondence between the Figure 6 Comparison of load test and analysis
load test and the analysis, which verifies the material results
models adopted in the LS-DYNA dynamic soil-structure
interaction analysis.
References
Lubkowski, Z. A. (1996) - "Verification of Oasys DYNA3D for the analysis of dynamic soil-structure
interaction", MSc Thesis, University of Bristol.
Willford, M. R., Piepenbrock, T. F., Izatt, C. B. and Lubkowski, Z. A. (1996) - "Non-linear time history seismic
th
structural analysis using LS-DYNA", Proc. 1 1 World Conf. In Earthquake Eng., Paper 558, Acapulco, Mexico.
2
Downloaded by [ NEW YORK UNIVERSITY] on [24/03/17]. Copyright © ICE Publishing, all rights reserved.