Deterministic and probabilistic analysis of local scour in alluvial bed

Bridge pier safety is governed by the local scour depth, a complex process influenced by flow-structure interactions. Accurate prediction is essential to prevent failures, yet existing empirical methods vary across regions and contain uncertainties. To address these uncertainties, a probabilistic analysis is required to estimate accurate and precise scour depths around the pier. This study performs a deterministic and probabilistic analysis of widely used scour prediction equations in alluvial stream.

The comparative study was performed using 274 field and 458 laboratory data points, Theil's coefficient (U) statistical test was applied to assess model accuracy. The Indian Road Congress method (IRC-78, 2000) performed best for regime channel data (U = 0.18), while the commonly used HEC-18 method also yielded reliable results. The Shen et al. (1969) model, incorporating Reynolds number effects, was most accurate for combined datasets.

A probabilistic analysis using the First Order Reliability Method (FORM) demonstrated that reliability increases with higher safety factors. The effect of variable distributions on pier reliability was significant at high safety factors but negligible at lower values. Sensitivity analysis identified pier diameter as the most influential factor for HEC-18, whereas flow velocity was critical for IRC-78.

Finally, a reliability-based safety factor was proposed for a target reliability index of 3.50. The findings enhance scour depth predictions, improving design safety and cost efficiency for bridge foundations in alluvial environments.