Flume Study on the Process of Slender Wood Jamming at Bridge Piers

The accumulation of floating large wood at bridge piers exacerbates flood risk. Climatic change, deforestation, soil erosion and expanding settlements cause increasing loads of wood into rivers. Besides naturally eroded wood, also harvested and treated wood, e.g. wood cut and stored on the floodplain that is mobilized during inundation events, is of concern. Cut wood is typically unbranched, more slender, dryer and often smoother than naturally eroded wood. Understanding how these different wood properties affect the jamming processes and identifying their governing control parameters is key for a bridge design with reduced jamming vulnerability. In this study, we therefore experimentally investigate the initiation and growth of wood jams from slender wood elements.

Flume experiments are conducted in a 1.7 m wide, fixed bottom flume at the TU Wien hydraulics lab. Flow depth was set to 0.30 m at a Froude number of 0.23 and a flow-Reynolds number of 1.16 x 10⁵. A cylindrical pier with 0.1 m of diameter was installed centrally in the flume. Unbranched cylindrical elements of 30, 45 and 60 cm length and 0.4 and 0.6 cm diameter were used to covered high slenderness regimes (l/d) of 50 - 100 and high relative lengths (l/D) of 3 – 6. The elements were produced from waterproofed pine dowels and plastic pipes sealed at both ends yielded elements with relative densities between 0.3 and 0.6 in water. A downward-looking camera recorded the jamming process.

Preliminary experiments focused on phenomenological observations of the jamming process. Approaching elements were only trapped, if their eccentricity (the lateral distance between their center and the center of the pier), was below one third of the element length. Within this range, slender long elements remained trapped for a long time – up to infinity in many cases. This first metastable regime is possible because of stabilizing compensatory movements, including rotational swaying around the bridge pier, vertical dipping and vibrations related to vortex-shedding. Hereby, swaying had the most stabilising effect as it exposed one end of the element into higher flow velocities upstream, thus increasing drag and initiating reverse rotation. The second stage of jam formation was governed by the interaction and collision of additional elements with the first element. At low eccentricity, the colliding element was rotated and attached parallel to the first element. At higher eccentricity, the collision destabilized the first element and rotated both elements. In this case, a third element was required to collide within a critical impact time to stop rotation and dislodgement. Thus, the stabilising mechanism shifted from compensatory movements to compensatory collisions. When collisions caused the trapping of elements, three (or more) elements formed a triangular, scissor-like pattern around the pier. This ‘scissor-pattern’ was a second metastable regime, typical for the tested slender long elements and observed throughout all runs. Experiments indicated, that friction between the elements and the pier surface controls the stability of the ‘scissor-pattern’, which is subject of ongoing analyses.