An uplift pile can continue moving while its applied load stays constant. A recent study in Buildings examines that response in calcareous sand using four instrumented model piles. Its findings make a useful case for paying attention to loading duration, but its experimental limits are equally important when considering transfer to a real foundation.
This is Civil Insight’s editorial review of a peer-reviewed paper by Xiuli Wu, Pengpeng Qu and Bin Tang, published online on 5 October 2026. Civil Insight’s review has not itself undergone peer review. The findings below are the authors’ laboratory results; the practical interpretation is identified separately.
The featured photograph shows a model pile and sand bed during setup for a separate laboratory uplift study. Sakr et al. (2025), Figure 5 · CC BY 4.0.
The research question and test arrangement
The researchers asked how pile-surface roughness and sand gradation affect movement and load transfer during sustained uplift loading. They prepared two calcareous sands: poorly graded Sand A and well-graded Sand B. Particles larger than 2 mm were removed, and the dry sand beds were prepared at a target relative density of 42% ± 2%.
The model piles were PVC tubes, 30 mm in diameter with 900 mm embedment. Three piles in Sand A had different sandpaper finishes representing smooth, semi-rough and rough interfaces. A fourth, semi-rough pile was tested in Sand B. The chamber measured 1,000 × 500 × 1,800 mm. Sand was placed around the piles, rather than installing them by driving or jacking through an already completed bed.
A pulley-and-deadweight system applied four successive load levels: 0.3, 0.5, 0.7 and 0.9 times the separately established ultimate uplift capacity. Each load stage lasted 21 days, with the next stage applied to the same pile. A dial indicator measured head movement, while strain gauges supported calculation of axial force and shaft resistance.
What the experiments demonstrated
Movement increased with holding time. Lower loading produced rapid initial movement followed by a flatter response; higher loading showed a more sustained accumulation after the initial stage. For the smooth pile A-Z1, movement under 210 N increased from 0.209 to 0.244 mm. At its highest sustained load, 630 N, the reported head displacement reached 3.781 mm after 21 days.
At the respective highest load levels, uplift deformation stiffness fell by 43.0%, 33.8%, 28.7% and 27.4% for A-Z1, A-Z2, A-Z3 and B-Z4. The smooth Sand A pile declined from 293 to 167 N/mm; the rough Sand A pile declined from 684 to 488 N/mm. These are model-system measurements at different absolute loads, so the percentages are not interchangeable field design factors.
The rougher Sand A systems accumulated less additional movement at their respective maximum loads. The authors also found time-dependent changes in axial-force distribution, concentrated in the upper and middle parts of the piles. Much of that adjustment occurred in the first three to seven days, while head movement continued accumulating through the 21-day holding periods.
A Burgers creep model fitted the measured displacement curves with coefficients of determination above 0.967. This indicates close agreement with the fitted experimental curves. It does not, by itself, establish predictive accuracy for another soil, installation method or full-scale pile.
Strengths and important experimental limits
The study combines head movement with instrumented load-transfer measurements, giving more information than a single ultimate-capacity value. It also states limitations that materially affect interpretation. There were no replicate tests or separate specimens at each sustained-load level. Consequently, the higher-load results include the history of earlier loading, and repeatability cannot be assessed.
The gradation comparison is also confounded. A-Z2 and B-Z4 had the same sandpaper grit but different normalised roughness because the sands had different median particle sizes. Their differences therefore cannot be attributed to gradation alone. Particle rearrangement, breakage and interface dilation were not directly measured; explanations involving those mechanisms remain interpretations informed by earlier research.
Scale and preparation matter. A 30 mm PVC model in dry, medium-dense sand does not reproduce the stress conditions, installation disturbance, groundwater or variability of every marine foundation. Although the paper adopts a geometric similarity ratio of 1:30, that does not establish complete similarity of long-term field behaviour. Its observed three-to-seven-day redistribution should not become a universal field waiting period.
Civil Insight’s practical interpretation
Where a project’s ground investigation identifies calcareous sand and uplift loading is relevant, this study supports asking how the proposed evidence deals with time-dependent displacement. Useful questions concern loading duration, installation history, soil state, interface properties and whether results have been independently repeated.
For site and project engineers, retaining the loading sequence and measurement history is particularly valuable: a final reading alone cannot explain a staged test. Selection of project testing, acceptance criteria and foundation design still requires the responsible geotechnical and structural specialists. The paper does not demonstrate Australian Standards compliance, field certification or a general allowable movement.
Original paper
Wu, X.; Qu, P.; Tang, B. (2026). Time-Dependent Bearing Behavior of Uplift Piles in Calcareous Sand. Buildings, 16(19), 3947. Published online 5 October 2026. DOI: 10.3390/buildings16193947. Full published article reviewed. Publisher record and correction/retraction searches checked on 9 November 2026; no correction or retraction was identified.
