Experimental P-Y Curves from Centrifuge Tests on Pile Foundations Subjected to Liquefaction and Lateral Spreading-Paper

Paper

Pdf Summary

The study uses five centrifuge tests of pile-supported wharves to assess pile behavior under earthquake shaking, liquefaction, and lateral spreading. Researchers back-calculated static and dynamic p–y curves from measured pile bending moments and displacements, compared them with API sand models, and evaluated the resulting models in LPILE against observed pile bending demands.

Static p–y curves were reasonably represented by API relationships using friction angles of 33° for loose sand, 37° for dense sand, and 45° for rockfill. However, the back-calculated soil stiffnesses were substantially lower than API recommendations: the modulus of subgrade reaction was 3,500 kN/m³ for both loose and dense sand and 5,200 kN/m³ for rockfill.

Dynamic curves showed markedly reduced ultimate soil resistance in partially or fully liquefied ground compared with static curves. Although dynamic resistance could exhibit short-lived increases associated with sand dilation, these were limited. The authors therefore tested a simplified approach that reduces API p–y resistance using a multiplier based on excess pore-water pressure: <strong>Pm = 1 − Ru</strong>, where Ru is the pore-pressure ratio.

LPILE analyses using these multipliers reasonably reproduced measured peak pile bending moments across the centrifuge tests. The findings support using pore-pressure-dependent p-multipliers to represent liquefaction effects in pseudo-static pile analyses, while emphasizing the need to select soil stiffness parameters appropriate to the conditions being modeled.

Keywords

centrifuge tests

pile-supported wharves

earthquake shaking

liquefaction

lateral spreading

p-y curves

pile bending moments

API sand models

LPILE analysis

p-multipliers