Improving the Reliability of the Calculated Axial Capacity of Piles in Sand-Paper

Paper

Pdf Summary

The paper evaluates the reliability of methods used to calculate the axial capacity of offshore piles in sand and proposes a more consistent, reliability-based approach to design. It considers the API method and four CPT-based methods, using probabilistic analyses that account for uncertainty in soil properties, loads and calculation methods. Resistance factors are calibrated against annual failure-probability targets of 10⁻³, 10⁻⁴ and 10⁻⁵.<br /><br />The methods show substantially different uncertainty: capacity coefficients of variation range from 17% to 39% in a case study, with API generally having the highest uncertainty and UWA the lowest. For the CPT-based methods, the study’s optimized resistance factors are approximately 1.35, 1.65 and 1.85 for the three respective failure-probability targets, assuming mean soil properties as representative values and specified load factors.<br /><br />The authors also assess a new “Unified” CPT-based method, developed from the strengths of existing approaches and calibrated against a screened database of pile load tests. Its uncertainty is lower than that of the existing methods, and comparisons across six sites indicate that replacing API with the Unified method while retaining current safety factors does not reduce foundation reliability. The method also predicts less sensitivity of capacity to pile diameter than many existing approaches.<br /><br />The Unified method is intended for driven tubular piles in siliceous, free-draining sands and should be used cautiously outside the conditions represented in its calibration data. Further large-diameter pile tests are needed, along with research on other soils, pile installation and capacity changes over time.<br /><br />Overall, the paper recommends reliability analysis as a complement to deterministic design. It helps quantify uncertainty, compare methods and support safer, more cost-effective decisions, while recognizing that calculated failure probabilities may be conservative because they include both natural variability and uncertainty in knowledge and models.

Keywords

offshore piles

axial capacity

sand

reliability analysis

API method

CPT-based methods

resistance factors

failure probability

Unified method

pile design