PSI - Issue 84

Maria Giovanna Tanda et al. / Procedia Structural Integrity 84 (2026) 272–279

276

Figure 2- Dimensionless maximum scour versus dimensionless collar location: a) circular pier with and without collar or hooked collar; b) rectangular rounded nose pier with and without collar. During the tests, the evolution of the maximum scour at the front side ( y s ) of the pier was monitored in time by visual inspection and/or time-lapse photos. Figure 3a shows the dimensionless scour ( y s /b ) versus the dimensionless time  =t∙U / b (Melville and Coleman, 2000) for circular piers. The two time-series of the scour obtained for the pier without a collar—positioned either upstream or downstream within the flume—show a good degree of test repeatability, despite the inherent uncertainties associated with visual evaluation of the depth of erosion. The figures show also that equilibrium scour conditions were not reached in most tests. However, given that the geometric scale of the model ( λ ) reasonably ranges from 1:50 to 1:80, and assuming Froude similarity, the corresponding time scale between model and prototype is √ λ, approximately 1/7 to 1/9. Thus, a 24-hour model test corresponds to 7–9 days at prototype scale, under the assumption of constant and elevated discharges. While longer test durations could have been performed—at the expense of extending the experimental campaign or reducing the number of tests—they are considered not fully representative of realistic hydraulic scenarios. From the figure it is evident that the presence of the collar significantly slows down the progression of the excavation. This effect, already, highlighted by several authors (Mashahir and Zarrati, 2002; Mashahir et al., 2004; Heidarpour and Afzalimehr, 2005), becomes more pronounced as the collar is located deeper into the sandy bed. However, in several tests, it is observed that over time, the maximum erosion tends to approach the value observed without the collar. Experimental observations reveal the following sequence of processes: scour initiates immediately and rapidly develops to the elevation of the collar, provided the collar is embedded below the original bed level. At this stage, the scour depth tends to stabilize, as the horseshoe vortex is not yet sufficiently developed to penetrate beneath the collar. Nevertheless, lateral expansion of the scour hole continues until its width equals or exceeds that of the collar. Once this threshold is surpassed, the resulting complex flow field induces secondary vortices and can erode the bed material beneath the collar, leading to a renewed increase in scour depth. Consequently, it can be inferred that both the position and width of the collar significantly influence the scour development and the final equilibrium value (Kumar et al., 1999). However, the effect of collar width was not investigated in the present test series, as very large collars were deemed structurally impractical due to support constraints associated with the pier. In such cases, alternative countermeasures independent of the pier—such as riprap— or a combination of two types of countermeasures (Gaudio et al., 2012) may offer more viable solutions. Anyway, whereas high flow conditions persist for a not too long period of time, delayed excavation progression provides a notable benefit to the overall stability of the pier. Figure 3b presents the results of the experiments conducted with round-nosed rectangular piers characterized by an aspect ratio L / b =3. As in previous cases, the two tests without a collar exhibit a reasonable degree of repeatability. The test with a collar installed flush with the sand bed shows a reduction in scour depth development over time, consistent with observations for circular piers under similar conditions. However, by the end of the test, which in this case appears to have reached a steady-state condition, the final scour depth is only marginally reduced compared to the no-collar case, with a normalized scour depth of y s / b =1.4 versus 1.6. The results for the test with the collar embedded at a depth of 1 cm below the bed surface ( d / b =0.33) are more interesting. In this configuration, the behavior previously observed for circular piers becomes even more pronounced: once the scour depth reaches the collar level ( y s / b =0.33), it temporarily stabilizes, indicating a suppression of further