Investigating the effect of grain composition on the erosion around deepwater foundations with a new simplified scour resistance test

Bridges are critical nodes in the transportation system, especially those crossing wide rivers and seas. Once in service, bridge foundations and the surrounding soils will act as a system together and are subjected to erosion caused by the flows. For more than one century, researchers and engineers have made significant efforts to understand the scour process and its mechanisms to protect bridge foundations from failures. Past studies on bridge scour mainly started from treating scour a hydraulic problem and focused on the flow field due to the interactions between flow and structure. Actually, for a given flow condition, the erodibility of bed materials plays an important role for scour initiation and propagation. For this reason, this study developed a new laboratory apparatus, Simplified Scour Resistance Test (SSRT), that can quickly evaluate the erodibility properties of soil samples and assist in identifying the critical conditions for scour initiation and development. It was firstly validated by Toyoura sand samples with a median particle size of 0.17 mm. Additional tests were then conducted on samples sieved to be populated by single-size particles (d = 0.075 mm, 0.25 mm, 0.5 mm, and 2.0 mm, respectively) and four different soil mixtures under different flow conditions. Results indicate that the grain composition has a significant influence on the scour resistance, while the compaction or density has less influence than the grain size. Granular materials with the same median particle size showed very different scour resistance, i.e., well-graded samples have much higher scour resistance than poorly graded ones. Scour characteristics of the mixed soils can be related to the particle size distribution curve. With this new testing apparatus and the results obtained, a correction factor ks is proposed to improve the traditional predictive methods considering soil composition, which overcomes the limitation of traditional scour prediction methods that often only considered the median particle size, D50.