Project Title/ID Number Simulation of Soil-Foundation-Structure Interaction of Deep Foundations—2312003
Start/End Dates 10/1/03—9/30/04
Project Leader Ross Boulanger (UCD /F)
Team Members Scott Brandenberg (UCD/GS), Dongdong Chang (UCD/GS), Heidi Feigenbaum (UCD/GS)

F=faculty; GS=graduate student; US=undergraduate student; PD=post-doc; I=industrial collaborator; O=other

Click on images to enlarge in a new window

1. Project Goals/Objectives:

The objectives of this project are:

  1. To perform OpenSees-based simulations of the seismic response of deep foundations in liquefied and laterally spreading ground, using the nonlinear p-y, t-z, and q-z material models developed and implemented in Year 6,
  2. Use the OpenSees-based simulations and available experimental data to investigate pressing issues in the development of performance-based pseudo-static pushover design methodologies, and
  3. Evaluate emerging pushover design methodologies against available physical data in support of the related proposal for “Summary Workshop on Emerging Design Methodologies for Pile Foundations in Liquefied Ground.”

2. Role of this project in supporting PEER’s mission (vision):

This project supports PEER's mission of providing performance based methodologies for bridge and building systems by addressing a problem that is of strong interest to owners and a major source of uncertainty in current design and assessment practices.

3. Methodology Employed:

The OpenSees-based simulations have focused on validation against existing centrifuge model test data for pile foundations in laterally spreading ground. Comparisons of simulations and experimental recordings have identified important physical mechanisms that were subsequently explored further by detailed back-analyses of loading mechanisms from the dense instrumentation arrays in the centrifuge models by Boulanger et al. (2003) and Chang et al. (2004). Attention is being focused on:

  1. The lateral load transfer mechanism between pile groups/caps and laterally spreading non-liquefied crust layers,
  2. The ability of pile foundations to restrain lateral spreading of bridge abutments, which reduces the loads compared to those predicted by uncoupled analyses of the lateral spreading and foundation responses; and
  3. Suitable kinematic and inertial load combinations for performance-based design practice. The plan for the comparative evaluation of emerging design methodologies is described in the accompanying proposal for a summary workshop.

4. Brief Description of past year’s accomplishments (Year 6) & more detail on expected Year 7 accomplishments:

figure 1This project is the natural progression of the work performed in prior years on developing and implementing p-y, t-z, and q-z materials for liquefaction conditions into OpenSees and the subsequent validation efforts. These different material models are now available on the OpenSees website and are being used by other researchers. The models are described in a report (Boulanger et al. 2003) and an upcoming ASCE-GI conference paper (Boulanger et al. 2004).

Simulations of piles in lateral spreads led to a re-evaluation of the load-transfer mechanism between pile caps and non-liquefied surface crust layers. This re-examination showed that the occurrence of liquefaction beneath a surface layer caused the load-versus-relative-displacement relation to be much softer than expected based on the common analogy to the static loading of retaining walls. This photo shows a side view of a centrifuge model pile cap embedded in a clay layer, after a surface layer of dry sand has been removed, that spread laterally over a loose sand layer that liquefied during earthquake shaking. Notice the passive bulge on the uphill/right side of the cap, the gap on the downhill/left side, and the various ground cracks.

figure 2The lateral load on the pile cap during virgin peak relative displacements were plotted versus the relative displacement between the pile cap and the free-field soil to the side of the pile cap (approximating the lateral spreading displacement that would occur away from the influence of the pile cap). The resulting load-transfer relation, shown below, was an order of magnitude softer than expected based on analogy to static load tests on retaining walls (as represented by the C=0.04 line in this figure). These results are presented in a paper at the upcoming WCEE conference by Brandenberg et al. (2004).

Simulations of the dynamic responses of the pile-group-supported structures, like the one shown in the photograph above, are focusing on the timing of inertial and kinematic load combinations. Again, these comparisons led to some more advanced back-analysis efforts on the experimental data, which involved superstructures with fixed-based periods of 0.3 and 0.8 seconds and two very different earthquake ground motion records. These results are showing that the peak lateral loading conditions occur mainly during strong shaking and are very close (within 10%) to the sum of the peak kinematic and peak inertial loads alone. Numerical analyses indicate that the appropriate load combinations are dependent on the pile foundation's lateral stiffness and the period of the structure relative to the dominant period for the laterally spreading crust as it transiently lurches during shaking. Additional analyses are progressing to numerically evaluate these findings over a broader range of ground motions, soil conditions, structural periods, and foundation stiffness.

5. Other Similar Work Being Conducted Within and Outside PEER and How This Project Differs:

Boulanger and Kutter have performed dynamic centrifuge model tests of piles in liquefied and laterally spreading ground with Caltrans funding. Dobry, Abdoun, and O’Rourke, in various collaborations, have also been doing centrifuge studies and developing simplified design methods (pseudo-static), but not nonlinear time-history methods. Ashford and colleagues have been doing field tests of piles in soils liquefied by blasting and have back-calculated p-y behavior. Elgamal, Dobry, & Abdoun are starting additional physical modeling efforts (shake table and centrifuge) in collaboration with Japanese researchers. To my knowledge, none of these other studies are performing dynamic FEM analyses similar to those proposed herein or are addressing the same design issues. The project proposed herein would bring these researchers together in the comparative study and associated workshop.

6. Plans for Year 8 if project is expected to be continued:

It is expected that completion of the above tasks will naturally progress into year 8. The development of performance based design methodologies requires addressing several different unresolved issues, of which only a few are being addressed in year 7. In addition, it is anticipated that PEER lifelines will fund additional centrifuge testing in years 8-10 that will address the interaction between bridge abutments and pile foundations in the presence of liquefaction. One issue being explored in these tests is the ability of pile foundations to restrain lateral spreading of bridge abutments; thereby reducing the loads compared to those predicted by uncoupled analyses of the lateral spreading and foundation responses. OpenSees-based simulations of those centrifuge models are needed to validate the ability of our numerical methods to capture this form of soil-structure-interaction, and subsequently explore a broader range of conditions than can be modeled in centrifuge experiments.

In addition, a continuation of this project would include providing numerical modeling and validation support to the PBEE demonstration projects planned under the Bridge and Transportation Systems area.

7. Describe any actual instances where you are aware your results have been used in industry:


8. Expected Milestones & Deliverables:

The deliverables of this project will include:

  1. A report documenting the recommendations for performance-based design practice as derived from the OpenSees-based simulations, the re-examination of existing dynamic centrifuge model test data, and the comparative evaluations of emerging design methodologies, and
  2. Continuing revisions and upgrades to the p-y, t-z, and q-z material models in OpenSees.
back to top