The accumulation of large wood (LW) at bridge structures poses a significant flood hazard, yet the critical role of initial log placement in triggering these blockages remains insufficiently quantified. This study provides a systematic experimental investigation into the parameters governing the accumulation probability (AP) at bridge piers and abutments. It focuses on four key domains: (1) initial log placement conditions (release distance, lateral position, and orientation); (2) approach flow conditions; (3) geometric characteristics of the bridge, particularly the varying distance between the pier and abutment; and (4) LW characteristics, including length, diameter, branching, and transport regime (uncongested vs. congested). Results identified the worst-case scenario for LW accumulation around piers and abutments: logs released 1m upstream of the abutment nose, from the channel centerline, and oriented perpendicular to the flow. Furthermore, the analysis reveals that accumulation probability is predominantly a function of approach flow velocity, log length and diameter, and the distance between the pier and abutment. Overall, these findings provide a quantitative, design-oriented basis for estimating LW accumulation risk and for guiding the layout of pier–abutment systems to enhance resilience against wood-laden floods.
The experimental program was conducted in a large rectangular flume (14 m length, 1 m width, 0.8 m depth) at Semnan University's Hydraulics Laboratory. The setup featured two circular plexiglass piers (diameter D_p = 0.05 m) and a rectangular abutment positioned 9.45 m downstream in the channel centerline. Three distinct flow conditions were examined, with discharges of 0.016, 0.026, and 0.036 m³/s, corresponding to approach velocities (u_0) of 0.16, 0.26, and 0.36 m/s and Froude numbers (Fr_0) of 0.16, 0.26, and 0.36, respectively. The approach flow depth (h_0) was maintained constant at 0.10 m throughout all experiments. Model large wood elements consisted of smooth natural wooden logs with specific lengths of L_Lw = 0.25, 0.3, 0.35, and 0.40 m and corresponding diameters of d_LW = 0.006, 0.012, and 0.025 m, scaled using a factor of λ = 20 based on field observations from the 2021 flood event in the Sajadrood River, Iran.

The experimental methodology employed a novel two-pronged framework. First, critical placement thresholds were identified by testing combinations of key parameters: LW release distances of 1 m, 3 m, and 6 m from the abutment nose; lateral release positions at the left bank (adjacent to the abutment), channel centerline, and right bank; and log position angles of 0°, 45°, and 90° relative to the flow direction. The combination that maximized AP defined the worst-case scenario. Subsequently, this worst-case scenario was utilized to systematically quantify the influence of other governing parameters, including LW characteristics and transport regimes.

The results identify a definitive worst-case scenario where logs released 1 m upstream of the abutment, from the channel centerline, with a perpendicular orientation (θ_LW = 90°) demonstrated dramatically increased accumulation probabilities. This configuration showed up to 86% higher AP compared to longer release distances (6 m) and 375% greater probability compared to parallel orientations (0°). The accumulation probability decreased sharply from p = 45% at x_LW = 1 m to p = 25% at x_LW = 3 m, and further to p = 6% at x_LW = 6 m for 25 cm logs at velocity 0.16 m/s. Similarly, for longer logs (L_LW = 40 cm), the probability of accumulation decreased by 61% as release distance increased from 1 m to 6 m.

Regarding orientation effects, for a 25 cm log at u_0 = 0.16 m/s, the accumulation probability increased progressively from p = 12% at 0° to p = 40% at 45°, and reached p = 45% at 90°. This probability rose dramatically for longer logs, reaching p = 0.93 for a 40 cm log at the perpendicular orientation. Lateral positioning also proved critical, with the highest accumulation probability consistently occurring at the channel centerline across all tested conditions. At velocity 0.16 m/s, accumulation probability for long logs reached 85% at the centerline, compared to 77% and 75% on the left and right banks, respectively. This distinction became more pronounced at velocity 0.26 m/s, where the difference between the centerline (80%) and the right bank (55%) reached 25 percentage points.

Log dimensions significantly influenced accumulation patterns. Longer logs (40 cm) showed dramatically higher accumulation rates than short logs (25 cm), with probability increasing by 365% at velocity 0.16 m/s and surging by 656% at velocity 0.26 m/s. Diameter effects were also substantial, with p = 22% (d_LW = 0.6 cm), 51% (d_LW = 1.2 cm), and 55% (d_LW = 2.5 cm) for L_LW = 25 cm at u_0 = 0.16 m/s. Shape characteristics further modulated accumulation behavior, with branched logs (89%) and coniferous logs (93%) showing significantly higher AP than smooth logs (29%) at low velocity (0.16 m/s), though these differences diminished at higher velocities.

Transport regime emerged as another crucial factor, with congested transport resulting in substantially higher AP compared to uncongested conditions. For L_LW = 25 cm, AP increased by an average of 15.5% for congested wood transport, while for L_LW = 40 cm, the value increased by an average of 82%. Under congested transport with L_LW = 40 cm, p ≥ 90%, exhibiting the highest values observed in this study.

These findings collectively establish initial log placement as a decisive factor in accumulation initiation, offering novel, quantitative insights for predictive models and the design of safer bridge infrastructure in wood-prone rivers. The research demonstrates that how wood arrives at a structure can be more influential than the wood's dimensions alone, providing crucial information for flood hazard assessment and resilient infrastructure design in river systems susceptible to large wood transport during flood events.