PSI - Issue 84

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

274

2. Laboratory setup The experiments were conducted in a flume with glass walls 10 m long, 0.30 m wide and with the bottom slope set at 0.5%. An inlet section with a tube bundle regularizes the entering flow. By adjusting a downstream sluice gate, efforts were made to establish conditions close to uniform flow along the entire flume, with particular attention to the section containing the piers. The bottom was covered with a 15 cm thick layer of siliceous sand with d 50 =0.28 mm and geometric standard deviation (Raudkiwi, 1998) equal to =( 84.1 15.9 ) 0.5 =( 0 0 . . 4 1 3 9 ) 0.5 =1.50 (1) Experiments were conducted in clear water conditions, with a flow velocity slightly below the incipient motion condition, for which no sand recirculation was needed. Circular and round-nosed rectangular piers were used in the tests. Circular piers have a diameter b =0.04 m, while rectangular ones have a thickness b =0.03 m and a length L =0.09 or L =0.12 m ( L /b =3 or 4) and were aligned with the flow. Piers without collars and with collars have been tested. The circular piers were equipped with both hooked and unhooked collars, while a sole unhooked collar was applied to the round-nosed rectangular ones. For circular piers the collars have effective width ( w − b )/2 and thickness equal to 0.02 m (half of the diameter) and 0.003 m respectively; for rectangular piers the width is 0.015 m (half of the width). The hooked collar features a 0.007 m high hook internally (0.01 m externally). In each test, two identical piers were evaluated simultaneously: one without a collar, serving as a reference, and the other fitted with a collar. Even if the experiments were conducted under clear-water conditions, bedforms (sand dunes) downstream of the piers were generated by the vortex systems triggered by the piers themselves. However, it was experimentally verified that the dunes formed immediately behind the upstream pier did not propagate toward the downstream one at any time throughout the testing period (  24 h). The piers were equipped with a metric scale in the upstream and downstream front. During the tests, visual readings were done, and, in some cases, time lapses videos were realized focused on the upstream side of the pier to detect the progression of the scour. At the end of each test, a photogrammetric survey was carried out to detect the dunes originating from the transport induced by the presence of the pier. The study considers two test series, with circular (A) and round-nosed rectangular (B) piers. The main characteristics common to each set of tests are reported in Table 1.

Table 1- Summary of the test characteristics; B is the width of the experimental channel.

Series

w / b ( − )

b / B ( − )

b / d 50 (-)

U / U c ( − ) 0.79

h / b ( − )

b (m)

L (m)

L/b ( − )

w (m)

d 50 10 − 3 (m)

h (m)

U 10 − 2 (m/s)

A 0.04 0.04

1 0.08 2 0.13 0.28 3 0.12 4 0.10 0.28 4 0.15 5 0.10 0.28

143

0.207

22.5

5.18

B1

0.03 0.09

107 0.175/0.207 22.5/26.7 0.79/0.94 5.83/6.90 107 0.176/0.196 23.8/26.5 0.84/0.93 5.87/6.53

B2 0.03 0.12

3. Experimental tests and results 3.1. Maximum scour and time evolution of the erosion

Table 2 and Table 3 summarize the conditions of the carried-out tests and the maximum scour observed at the upstream face of the circular and round-nosed rectangular (type B1 and B2) piers. The mean velocity U in the channel varied between 0.225 m/s and 0.267 m/s. Theoretical values of the threshold velocity for general bed movement, estimated with the application of literature relationships (Raudkiwi, 1990; Neill, 1968; Neill, 1973; Goncarov, 1964; Garde, 1970), vary in the range 0.20  0.34 m/s. Experimentally, it was evaluated as U c  0.285 m/s. The tests were then carried out in clear water with a ratio of U / U c =(0.79  0.94). Collar was positioned at the surface of the sandy bed level or below it at a depth of d =1 and d =2 cm, corresponding to a ratio d / b equal to 0.25 and 0.5 for the circular pier and 0.33 and 0.66 for the round-nosed rectangular piers. Figure 2a shows the maximum dimensionless scour y s / b versus the dimensionless collar location d / b for circular piers. Despite