REDUCTION OF SCOUR AROUND BRIDGE PIERS USING PROTECTIVE DEVICES

Abstract

A cylindrical bridge pier constructed in a channel with its axis perpendicular to the approach flow creates an adverse pressure gradient in the longitudinal direction and a favorable pressure gradient in the downward direction in front of the pier resulting in a complex three-dimensional flow system comprising downflow, horseshoe vortex, cast-off vortices and wake vortices. This results in local scour around the bridge pier. During the last few decades, River Engineers have tried to predict the maximum depth of scour under different conditions in order to make a rational design of the pier foundation possible. Many of these equations are empirical in nature and a great many of them are based on the Regime Theory. Some effort has also been made to base the estimates by considering the stochastic nature of the turbulent flow and using boundary layer theory. Studies of CarstensC1966), Nakagawa and Suzuki(1975), Baker(1980b) and Kothyari(1992a) can be quoted in this regard. The results of these studies have been found to be widely differing by MelvilleC1975), Johnson(1995) and many others. This may be due to the absence of a clear understanding as to the main cause of scour around bridge piers. There are two schools of thought as to the main cause of scour in front of the pier. Shen et al.(1966), Baker(1981) and Kothyari(1989) consider the horseshoe vortex to be the main cause of local scour. On (i) the other hand, investigators like HJorth(1975), Raudkivl(1986) and Johnson and Ayyub(1992) consider downflow, which acts as a vertical jet, to be the main agent of local scour. However, both these hypotheses have not been subjected to quantitative evaluation as yet. Several investigators have studied control of local scour around bridge piers by streamlining the pier shape or by using flexible mattresses, rip-rap, piles in front of the pier etc. Schneible(1951) and Tanaka and Yano(1967) placed a collar on cylindrical piers to combat the horseshoe vortex. Kikkawa et al.(1973) found a number of horizontal plates called guards to be effective in reducing the scour depth while Gupta and Gangadharaiah(1992) proposed a delta-wing like structure on the pier. Considering the downflow as the main cause of scour Chiew(1992) used a rectangular slot through the pier for reducing the scour. However, the present state-of-art is not such that the actual amount of scour reduction can be computed for design purposes over a wide range of flow conditions and geometries of the pier and the appurtenances. Also information on the location of the deepest scour is lacking. The present study was taken up primarily with two objectives in mind. The first was to use a combination of theoretical analysis and experimental data to ascertain whether the downflow in front of the pier is indeed the main agent of scour as hypothesized by some investigators. A second objective was to study scour reduction achieved by using several appurtenances like a fin in front of the pier, a slot through the pier and a collar fitted around the pier. Detailed studies were (ii) planned to be carried out on piers fitted with collars so as to develop a predictive equation for the reduction of scour depth. An exlenHlve sot of experiments wan conducted In tho llydrn.nl Ion Laboratory of the CIVILENGINEERING department, University of Roorkee, Roorkee, India. A fixed bed masonry flume 30.00m long, 1.00m wide and 0.60m deep with a working section 2.00m long, 1.00m wide and 0.30m deep located at 12.00m from the inlet was used. The writer's experiments were restricted to the case of clear-water flow- a condition in which there is no sediment transport from the upstream- past circular bridge piers. Only uniform sediments were used. Three fins of height equal to that of the cylindrical pier and lengths equal to 0.5b, 1.0b and 1.5b were used on a cylindrical pier of 112.5mm diameter. Here b is the diameter of the pier. Sediment used in this part of the study was 1.542mm in size. A slotted pier 112.5mm diameter with a rectangular slot 0.25b wide was tested with and without extension into the scour hole as well as for angles of attack of approach flow ranging from 0 to 45 . The sediment used here was also 1.542mm in size. Studies related to the collar were carried out using a cylindrical pier 112.5mm in diameter, with five circular collars of diameters 1.5b, 2.0b, 2.5b, 3.0b and 4.0b and thickness 3mm placed at three elevations, viz. at bed level and at 0. 15Y and 0.25Y above the bed. Here Y is the depth of flow. The o o o sediment used was again 1.542mm in size. Additional data were collected along with one more pier of diameter 61mm with 2.5b collar at bed level for three flow conditions and also using two more sediments of sizes 0.775mm and 1.183mm. (iii) The data from previous investigations regarding equilibrium scour depth for cylindrical piers without appurtenances [ds ranging from 0.04m to 0.549m] and for cylindrical piers fitted with collars [ds c ranging from 0.016 to 0.0945m ] were also used in the analysis. Here ds is the equilibrium scour depth with pier only and ds is the equilibrium scour depth when a collar is placed around the pier. Ettema(1980) has given a quantitative estimate of the downflow in front of the pier. According to him, the maximum downflow velocity is equal to 0.4times the approach flow velocity and it occurs at 0.46b above bed level. The downflow at any elevation has a velocity equal to zero at the pier and also at a distance of 0.02b to 0.05b, being closer to the pier at lower levels. In the Jet-Flow model developed during this study, this downflow is made equivalent to a circular vertical jet, diffusing to a velocity V =0.4U at y=0.46b above bed level. Here U max o o is the approach flow velocity and y is the distance above bed level. The diffused jet diameter D at y=0.46b is taken equal to 0.10b (=2x0.05b). Assuming the vertical velocity distribution to be Gaussian, Albertson et als'(1950) relations for the diffusion of jet were used to determine the equivalent nozzle location and the characteristics of jet issuing out of it. Once the equivalent jet parameters are known, these were then used to calculate scour depths from the relations for scour in vertical jets given by Sarma and Sivasanker(1967), Rajaratnam(1982) and Stein et al.(1994). These results of scour were compared with the actual scour for the data of Chabert and Engeldinger( 1956), Shen et al.(1969), Verstappen(1978), Walker(1978), Jain and Fischer(1980), Ettema(1980), Kothyari(1989), Raudkivi and Ettema(1983), Dey et al.(1992) and the (iv) writer for clear-water scour In the case of cylindrical piers. It was found that the estimated scours due to downflow computed from the Jet-Flow model were much smaller than the observed scour. It was thus concluded that downflow is not the only cause of local scour around bridge piers. The pressure distribution at the junction of the pier and the fin for three different flow conditions were compared with those without the fin. It was found that there was considerable reduction in the magnitude of the pressure gradient due to the introduction of a fin on the pier. As such, the downflow is expected to reduce appreciably due to the introduction of the fin. However, the corresponding scour reduction was not appreciable. This further supports the finding mentioned above that downflow is not the only agent for scour. A slot of length equal to the depth of flow or a slot extending into the bed gave significant reduction in scour when the flow was parallel to the slot, thereby confirming the results of Chiew(1992). However, the scour depth was found to be very sensitive to the angle of attack of the approach flow. The reduction in scour decreased with increase in angle of attack and there was no scour reduction due to a slot if the slot was oriented at 35 to the flow. A further increase in the angle of attack resulted in more scour than in the absence of the slot. The collar unlike the fin and the slot, is not sensitive to the angle of attack and was thus chosen as a protective device for detailed (v) study. Detailed information was collected on the temporal variation of scour as well as on the migration of the scour hole with time. Measurements using a collar at bed level showed that, for smaller collars, the section of deepest scour occurred in the front and always remained there. However, for larger collars, the section of deepest scour moved both longitudinally and laterally with respect to the line of symmetry. Analysis of the temporal variation of scour depth around bridge piers fitted with collars has shown that the model of Islam et al.(1986) for piers without collars is applicable to the case of piers with collars too but with different scaling parameters. The variation of these parameters with the geometry of the collars and their elevations above the bed has also been studied. Dimensional analysis of the various parameters affecting the maximum scour depth in the case of cylindrical piers fitted with collars was carried out. The data collected in the present study as well as those available from earlier studies were analyzed in accordance with the forgoing relationships. The following equation was obtained as a result of such analysis: ds P Y o 0.6 --°n-2o°m3 +4. 0n .774/177 ljD~^pl -_ 0n . 1u0[I £* ] ,ds -ds -> As rY -h, ,Y -h.2 rY -h.3 55.26-127 rl -lt\ rl -ll-i rl -IK (vi) (1) Here B is the diameter of the collar and h is the elevation of the collar above the bed. One can use Eq. (1) in conjunction with a predictor for maximum scour around a pier without any protective device to estimate the scour reduction due to the placement of a collar on a pier. Equation (1) shows that, In general, scour decreases with Increase in collar size and decreases with its elevation above the bed. This equation also shows that a smaller collar at lower elevation may be more effective than a larger collar at a higher elevation. (