Random Field Model for Longitudinal Vibration Analysis of Pipe Piles in Bidirectionally Heterogeneous Soil

The inherent spatial variability of soil severely impacts the vibration of pipe piles, undermining the reliability of low‐strain integrity testing. Conventional deterministic models fail to adequately capture these effects. This study proposes a novel random field model to analyze the longitudinal vibration of pipe piles in bidirectionally heterogeneous soil. The model integrates the radial and vertical variability of the external soil with the vertical variability and mass inertial effects of the soil plug. Parametric analysis reveals that the degree of the soil variability in shear wave velocity has a dual impact on the vibration characteristics. For short piles, this variability acts as detrimental interference, increasing the failure probability of pile length detection. Conversely, the effect differs for long piles, which are often characterized by weak signals. For these piles, a moderate degree of soil variability can actually enhance the pile‐toe reflection, serving as a “natural signal enhancer” and significantly reducing the failure probability. Furthermore, a strong correlation was identified between construction disturbance and soil variability. The randomness of the external soil is the dominant factor determining failure modes, rather than the randomness of the soil plug. This research establishes a robust probabilistic framework for reinterpreting signals from low‐strain pile testing in variable soils, providing valuable theoretical guidance for improving the accuracy of pile integrity assessments.