Comparison of Predicted Pile Load-Displacement Response Using PISA and API p-y to Static Laterally Loaded Onshore Test Piles in Sand at Cuxhaven-Paper

Paper

Pdf Summary

The paper compares the API sand p–y method with the PISA General Dunkirk Sand Model (GDSM) in predicting the lateral response of a large-diameter monopile. Predictions were assessed against static load-test measurements from an impact-driven, 4.3 m diameter pile at Cuxhaven, Germany, embedded in dense, over-consolidated sand. Its embedment-to-diameter ratio was 4.3—far below the ratios of the slender piles on which the API method was originally based.<br /><br />The PISA rule-based model matched the measured load–displacement response substantially better than API over the tested displacement range, up to approximately 0.03 pile diameters. At that displacement, API predicted a capacity 88% higher than measured, whereas PISA was 8% lower. At a normalized load of 0.4, API predicted roughly one-third of the measured displacement, potentially producing an unsafe serviceability assessment. At larger displacements, the API method also predicted greater capacity than PISA, raising concerns about non-conservative ultimate-limit-state designs, although the test pile was not loaded far enough to verify that behavior directly.<br /><br />Both methods underpredicted initial stiffness. The authors suggest that a site-specific PISA model, calibrated using advanced three-dimensional finite-element analysis, might improve this prediction. They conclude that the traditional API method may not reliably represent the behavior of modern large-diameter, low-embedment-ratio monopiles in sand, while PISA provides more accurate predictions for the tested conditions. Further research is needed to identify when API-based capacity estimates become unconservative.

Keywords

API p–y method

PISA GDSM

lateral response

large-diameter monopile

load-displacement response

Cuxhaven sand

static load test

serviceability assessment

initial stiffness

over-consolidated sand