**1. Introduction**

The local scour of bridge piers is one of the main natural causes to bridge structure damage [1,2]. A survey conducted by the US Federal Highway Administration (FHWA) in 1973 showed that more than 20% of pier damage and 70% of abutment damage in a total number of 383 bridge accidents were caused by scour [3]. Wardhana and Hadipriono [4] studied over 500 instances of bridge damage that occurred between the year 1989 and 2000 in the United States and found that more than 50% of the bridge failures were caused by scour due to flood. In a Transportation Research Board's (TRB's) National Cooperative Highway Research Program (NCHRP) report, scour was implicated in up to 60% of bridge failures [5]. Yi et al. [6] reviewed 160 bridge collapse accidents that occurred from the year 2000 to 2014 in China and found that more than 30% of the bridge accidents were caused by scour. Scour is a natural phenomenon caused by the erosive action of flowing streams on erodible beds. Scour can remove the sediment around the bridge foundation, change the structural natural frequencies, reduce the lateral bearing capacity of the foundation and even reduce the seismic capacity of the bridge [7–11].

Scour is usually classified as general scour, contraction scour or local scour. Compared with the first two types of scour, the local scour is generally taken as the main hazard responsible for bridge damage [2,12]. Local scour is the flow erosion due to the blockage of the bridge pier or foundation, where sediment is carried away from the bridge piers or foundations and the scour hole is formed near the piers or foundations [1]. Many researchers have carried out theoretical analysis and experimental research on the local scour of piers. Ataie-Ashtiani and Beheshti [13] studied the local scour around pile groups under steady clear-water scour conditions experimentally. Zhang et al. [14] explored the local scour mechanism of cylindrical piers. Khosronejad et al. [15] carried out experiments and numerical simulations to investigate the clear-water scour around three bridge piers with di fferent cross sections. Sumer et al. [16] summarized some recent research about scour around coastal structures, including flow and scour processes with the subheadings, and sediment behavior close to the structure with the subheadings and further examined the local scour around a pile subject to combined waves and current through testing [17]. Bouratsis et al. [18] carried out a quantitative spatio-temporal characterization study on scour at the base of a cylinder.

Local scour of a bridge foundation is almost inevitable in the actual river or ocean environment. Therefore, adopting some appropriate scour countermeasures to reduce the local scour hole around the bridge pier and protect the foundation of the bridge structure has become an important issue for the design and maintenance of bridges located in the erodible sediment beds. Traditional countermeasures used in bridge engineering can be divided into two categories: (a) passive countermeasures and (b) active countermeasures [2]. Passive countermeasures mainly include: the enlarged foundation [19], the concrete protection [20,21], the partial grouted riprap [5] and the riprap [22,23]. These measures aim to increase the resistance capacity of the bridge foundation to the inflow shear stress by laying a protective layer on the riverbed around the bridge foundation, so as to protect the erosion-prone sediment in the lower layer. Passive countermeasures have been widely used in current engineering practice, but they cannot eliminate the fundamental causes of scour. On the contrary, active countermeasures focus on the disturbing the flow field and thereby reducing the e ffect of inflow, and thus attract research interest. Some active countermeasures reported in the previous research consist of: the anti-scour collar [24,25], the ring-wing pier [26], the slot [27] and the sacrificial piles [28], etc. The active measures aim to improve the original hydraulic characteristics, such as changing the direction and reducing the flow speed, to weaken the e ffect of downflow and horseshoe vortexes and hence protect the bridge foundation from local scour.

Tafarojnoruz et al. [29] reviewed various flow-altering scour countermeasures and concluded that the di fficulties in the field applications of active countermeasures limited their practical use. Among the active methods, the anti-scour collar can be a useful and simple countermeasure for local scour around cylindrical bridge piers, which can reduce the horseshoe vortex strength by diverting and acting as an obstacle in the downflow passage. Previous literature shows that placing a collar around the bridge pier can reduce the depth of scour [30]. The anti-scour collar external diameter and installation height proved to be the most important factors influencing its protection e fficiency [2]. Chen et al. [31] proposed a new type of collar named hooked-collar, and found that that maximal downflow is highly reduced along with a corresponding decrease in horseshoe vortex strength for the experiments with the hooked-collar, compared to cases without the collar. Chiew [24] investigated the local scour of a pier using a collar, a slot, and their combination and found that placing either a collar or a slot around the bridge pier can reduce the scour depth, and their combination can further reduce the scour depth. Gaudio et al. [32] carried out clear-water tests on the bridge pier scour with various combinations of collars with some other countermeasures and concluded that an improper combination of two countermeasures may be less e ffective than each individual countermeasure. Similar studies about collars in reducing the scour depth can be found in several studies [25,27,29,33–35]. However, these studies seldom focused on the distribution of the scour depth around piers and how to design the anti-scour collar and optimize its protective e ffect on the local scour of piers.

In order to address the issues related to the optimal design of anti-scour collars and the local scour depth distribution around cylindrical bridge piers with anti-scour collars, this study is organized as follows: (1) A scour experimental program is firstly presented to study the protective e ffect of anti-scour collars at a local scour of a cylindrical bridge pier model; (2) local scour of the pier model

with and without anti-scour collar is studied experimentally; (3) the e ffect of collar installation height, collar external diameter and collar protection range on its scour protective e ffect were investigated; and (4) the design of anti-scour collars on the characteristics and development of local scour hole around the pier model are discussed parametrically.
