PINSA 64, A, No. 4, July 1998, pp. 569-580. (© Printed in india SCOUR AROUND BRIDGE PIERS RJGARDE* AND U CKoTHvarie* *INSA Senior Scientist, Central Water and Power Research Station, Khadakwasla, Pune-411 024 **Civil Engineering Department, University of Roorkee, Roorkee (U.P.) (Received 10 June 1997; Revised 05 December 1997; Accepted 04 February 1998) ‘The paper describes the phenomenon of scour around bridge piers and then enumerates the methods for its prediction. ‘The scour data from prototype bridges are analysed to comment on the relative accuracy of four methods of scour prediction. Brief comments are made on scour around bridge piers in clayey bed and gravel-bed rivers. Lastly, various techniques studied for scour control and protection are deseribed. Key Words: Scour in Alluvial Streams; Scour Estimation Equation; Laursen-Toch Equation; Melville & Sutherland's Equation; Chitale’s Lacey-Ingl Scour Prevention; Scour Protection Devices; ‘Method; Kothyari-Garde-Ranga Raju’s Method Introduction Scour is the local lowering of stream bed elevation which takes place in the vicinity or around a structure constructed in flowing water. Scour takes place around bridge piers, abutments, around spurs, jetties and breakwaters due to modification of flow pattern in such a way as to cause increase in local shear stress. This in tum dislodges the material on the stream bed resulting in local scour. In the case of bridges, the estimation of correct depth of scour below the stream bed is very important since that determines the depth of foundation. Huber' has stated that since 1950 over 500 bridges in USA have failed and that the majority of the failures were related to the scour of foundation material. Such data are not available for the Indian bridges; however, this has been the matter of concer to the Government of India and some detailed hydrologic and scour studies have been undertaken at selected bridges by the concerned organisations. This concern about safety of bridges is primarily due to three reasons which are: (1) inadequate knowledge about scour phenomenon when the bridges were constructed; (2) inadequate data on which the design flood was chosen; and (3) increase in the loading on the bridge due to increase in size of trucks, wagons, and their frequency of operation. * Author for Correspondence: Professor RJ Garde, Professor Emeritus, CWPRS, Pune-411024 ‘The stream bed lowering at the bridge can take place due to four primary reasons. If the bridge is located downstream of a large dam, there is a slow lowering of the bed and reduction of stream slope due to degradation. Degradation takes place when the stream transporting sediment becomes deficient in sediment supply due to sediment being stored upstream of the dam. In extreme cases this lowering can be as much as 4 to 6 meters. Secondly, if for reducing the cost of the bridge the stream is contracted by building guide bunds etc., such contraction can cause additional lowering of the stream bed. The depth in the contracted section is given by DyD,=(By/B,°""™, -a) where B, and D; are width and depth of flow in the uncontracted section and Bir and D; are the corresponding values in the contracted section. The third type of lowering that tzkes place around the bridge pier is due to modification of flow structure due to presence of the pier. Depending on the pier shape and free stream condition, an eddy structure comprising of one or more of the three eddy structures, namely horse-shoe vortex, wave vortex system and the trailing vortex system can form; this increases the local shear on the bed and causes scour. Typical formation of horse-shoe vortex is shown in Fig. 1. Lastly, additional scour can also570 a a wonse- Shoe voRTex RJ GARDE AND UC KOTHYARI Fig. 1 Definition sketch take place if the flow direction is inclined to the pier axis. Factors Affecting Scour Depth A number of papers have been published since 1940 on various aspects of scour around bridge piers. Based on the experimental work and some theoretical analysis it is found that the following factors affect the scour depth at the bridge pier. (a) Whether the Incoming Flow is Clear Water Flow or it Carries Sediment: Clear water flow occurs when uJ/u., is less than unity while for sediment transporting flow w/u., is greater than unity. Here U.=VgDs is the shear velocity of flow and wis the shear velocity at which bed material starts moving, D is the depth of flow in the river and 5 is the river slope. Average shear stress on the bed is 1,= G,U3.7,DS where y,is the unit weight of water and C; its mass density. All factors remain- ing the same clear water scour is found to be about 10% more than scour in sediment transporting flows. Further, whereas in clear water flow it takes several hours to reach the maximum scour below iver bed dj, in sediment transporting flow corresponding equilibrium scour depth dy is reached in relatively shorter time. () Effect of Change in Depth of Flow: Experiments by Melville and Sutherland? have shown that when (depth of flow/pier width) ratio i.e., Dé is greater than 2.6, scour depth does not depend on the depth of flow; for smaller depths, the scour depth depends on the depth of flow. (0) Effect of Shape of Pier Nose: The shape of the pier nose affects the strength of horse-shoe vortex as well as the separation of the flow around the bridge pier; hence it affects the maximum scour depth. The following Table I prepared on the basis of studies by Laursen and Toch’, Chabert and Engeldinger‘, Garde‘, and Paintal and Garde’, gives relative effect of pier-nose shape on the maximum scour if for cylindrical pier the ‘maximum scour is taken as unity. Table Average values of shape coefficients K, Shape K Cylindrical 10 Rectangular (/b=2106) 1.1 to 1.25 Lenticular (2:1, 3:1,4:1) 0.93, 0.79,0.70 Elliptical 21,31) 10,086 Joukowsky profile (4:1, 5:1) 1.0,0.80 Triangular: 15° apex angle 048 60° apexangle 0.75 90° apex angle 0.88 120° apex angle 0.98 150° apex angle _ 1.00 (@) Effect of Angle of Inclination of Pier on Scour Depth: When the axis of the pier makes an angle @ with the general direction of the flow, two ‘major changes take place in the flow field. Except in the case of cylindrical pier, the separation pattern is drastically changed resulting in change in vortices. Secondly, the open width between the piers, perpendicular to the flow direction reduces as the angle of inclination @ increases.SCOUR AROUND BRIDGE PIERS ‘The coefficient Ke= Scour at an angle of inclination @ ‘Scour when the flow is axial depends on pier shape and @ and increases as shown in Table II. For rectangular piers Ky will be function of 8 and I. Curves between these parameters are given by Laursen’. Table It Effect of 8on Ke for rectangular pier (Ub~6.0) o Opes oes loess 0s aaa Ke Lo. 137237 3.77 (©) Effect of Opening Ratio on Scour Depth: The opening ratio a. is defined as a=(B-b)/B where B is centre to centre spacing of the piers and b is the pier width. When b is very small compared to B, a is close to unity and flow around one pier does not affect that around the other. However, as a decreases, the interference effect becomes more pronounced and scour depth increases; in such a case D,/D or D,/d-a*. Here D, and Dy are scour depths below the water surface for sediment transporting and clear water flows respectively. The analysis of extensive data collected by Garde et al,” indicate that n=0.30. ©) Effect of Bed Material Characteristics: In the case of noncohesive materials, the characteristics of the bed material that affect the scour depth are sediment density, median size d of the bed ‘material, its standard deviation and stratification. For all practical purposes the density of natural sediments can be taken as 2.65, a constant value ‘As regards the sediment size, Lacey-Inglis approach (see below) suggests that D,~d"™*. Since the average shear stress on the bed (=y, DS) at which bed material moves-known as the critical shear stress-increases as the sediment size increases, it stands to reason that scour depth should be affected by the size of the bed material. Hence, for given flow condition, larger than the sediment size d, smaller should be the scour depth. ‘The clear water scour depth should decrease with increase in sediment size. Analysis of data over large range of sediment size by Kothyari* has 117 71 indicated that d,~d*' while for sediment transporting flow d.~d*”, Here d,, and d., are the scour depths below general bed level for sediment transporting and clear water flows. The effect of size distribution of the bed material on the scour depth is more significant, When the standard deviation o, of the bed material is large and the bed material contains some nonrhoving sizes for a given discharge, the coarser material would tend to accumulate in the scour hole and inhibit further development of scour depth. Hence for the same median size, scour depth will be smaller for material with larger geometric standard deviation o,. Here = V2 dyaldsotdso/die) and dy, diy and dy, are such sizes that 84%, 50% and 16% material is finer than dys, dip and dy sizes respectively. The percentages 84 and 16 are such that for normal or Gaussian distribution dy=(dyotstandard deviation) and (dy,= dyy-stanidard deviation). If the correction factor K, is defined as Equilibrium scour depth for non - uniform material . Equilibrium scour depth for uniform material of the same median size K, would depend on o,. On the basis of experimental. data of Raudkivi? and Kothyari* the following table is given. (g) Stratification: Ettema"” and Kothyari* have studied the effect of stratification of the bed material on scour depth in case of clear water scour. It is concluded that the stratification, in which a relatively thin coarse top layer covers a thick fine bottom layer, is the critical condition. Once the top coarse layer is scoured away, scour depth will rapidly increase. (A) Effect of Flow Parameters: Based on certain theoretical analysis, physical reasoning and analysis of experimental data, investigators have arrived at the basic flow parameters to which the dimensionless scour depth is related. Thus Breusers ef al", Laursen and Toch", Larras” and Ettema'® consider 6/D as the important parameter and hence they related d,/D to b/D. Thus according to Breusers ef al,'' 4,JD=1.4(b/D) =)$72 RJGARDE AND U C KOTHYARI Table Mt Variation of K with 0, 3 To TS 2 Ta 2B 33 7 78 k, 10 0.90 07s 050 038 025, 0.160 0.08 where b is the pier width or diameter. Lacey-Inglis Equation Lacey and Inglis" compute Lacey’s depth Dig for flood discharge Q as D040)" @) in metric units and relate scour depth below water surface D,, to Dig as Dj Dio=2.0, - @) Here fis known as Lacey's silt factor and is given by eq. 9 (see below). On the other hand, Garde’, Shen et al.'*, Venkatadri et al.'*, Coleman", and Jain show the importance of Froude number Fr=U/vgD, where U is the average velocity of flow. Then d,J/D is related as ,JD=KFr'(bIDy. ) where K is the constant. Typical equation of this category is U.S. Army Engineers’ equation 4JD=2.10 (Fr) (b/D)"*, © Shen et al.” have related the clear water scour to the pier Reynolds number Ub/v on the assumption that the strength of horse-shoe vortex is a function of Ub/v. Here v is the kinematic viscosity of water. They have proposed the following equation for enveloping curve between d,, and Ub/ 4,=0.000223(Ub/»)"*", +) Generally, the above types of relationships are valid for cylindrical piers and for piers of other shape and inclination to flow, the d,, or dj. value needs to be multiplied by K, and Ky. Aiso these equations are valid for nearly uniform bed material, where armouring is not pronounced. Recent Equations for Prediction of Scour Depth Below are briefly discussed Lacey-Inglis approach commonly used in India and a few recently developed equations. In the earlier part of this century Lacey analysed the data from stable irrigation canals flowing through loose noncohesive sandy material in Indo- Gangetic plain and obtained the following equations for depth (or hydraulic radius) Dg and perimeter (or width) P. Dyg=0.41( OH)" - @) and P=4.751Q, (8) where Q is the discharge in m’/s, Dig and P are in ‘mand fis Lacey’s silt factor related to median size of the bed material d by the equation frl.76vd, 0) d being in mm. On the basis of analysis of scour data on 17 bridges in alluvial rivers in North India, inglis'* found that the maximum scour depth below water level, D,. is related to computed value of Dig as D,=KDio. « (10) where K varied from 1.76 to 2.59 with an average of 2.09. Hence according to Inglis, D,, is given by the equation DH=2.0 Dio, (4) When bridge pier foundation is to be designed, this equation will be used for a flood discharge of return period 50 to 100 years, even though eq. (3) is at best valid for bankful discharge. In the light of the variables affecting scour depth mentioned above, it will be clear that K in eq. (10) should depend on pier shape, sediment size, obliquity of flow etc. Since these factors are not explicitly taken into account, Lacey-Inglis method should not be used outside the range of data on which it is based.‘SCOUR AROUND BRIDGE PIERS Laursen-Toch Equation The equation proposed by Laursen and Toch'? for prediction of d, is d,JD=1.35(6/D)™, Melville and Sutherland's Equation Melville and Sutherland? have proposed a method for estimating the scour at bridge piers. The method is completely based on the analysis of the laboratory data. Basically they assumed that the largest possible scour depth around the bridge pier is given by dn b, o (11) - (12) This scour depth below the general bed level is reduced by multiplying factors which depend on whether the scour is clear water scour, depth is shallow and sediment is graded. The multiplying factors are determined from the analysis of experimental data covering a wide range of pertinent variables. Chitale’s Method” The method proposed by Chitale estimates the probable maximum scour depth at the bridge pier. Maximum scour depth at the pier diy is computed as den =25 b. ++ (13) If the bridge is located at a constriction caused by guide bunds, the average depth D, in the contracted section is related to that in the uncontracted section Dyby DyD=(ByB,Y? °°, ey where B, and B, are the unobstructed and obstructed widths of the river channel. This average depth D, in the contracted section may not be uniform across the width because of nonuniform flow distribution and curved entry. ‘Analysis of eight bridges in Indo-Gangetic plain indicated that ratios of maximum depth to the average depth D, varied from 1.2 to 1.67. Hence, Chitale recommends a ratio of 1.7. :. Maximum local depth=1.70 D, w+ (14) Hence, Dya=2.5 b+1.70 D; 573 where Dien is the maximum anticipated scour depth below water surface. Kothyari-Garde-Ranga Raju's Method Based on the extensive laboratory data collected using uniform and nonuniform sediments, stratification and steady flows, Kothyari et al.**" have proposed equations for determining clear water scour depth d., and equilibrium scour depth d,, for steady flows. The analysis was done using the mathematical model based on the assumption of formation of horse-shoe vortex on the upstream side of the pier. Such a vortex increases the shear stress on the bed and causes scour. Their equations for scour depth are Clear Water Scour: te -006(8)"(2)" pe (U*-U? ) 030 6 a) \a * (Ay,d/ py) .. (13) where the average critical velocity U, is given by (C2 dyallp)=1.20(b/4) "(Did)"... (16) Scour Under Sediment Transporting Flow: d,/b=0.88(b/d)?"(D/dy“a?* cee Here Ay, is the difference in specific weight of sediment y, and water y,, and p,tis the mass density of water. It may be seen that in sediment transporting flows, the scour depth is not dependent on velocity. It may also be noted that the opening ratio a affects the scour depth. When the sediment is nonuniform, the scour depth is reduced as compared to that for uniform sediment. The reduction factor K, is the function of the geometric standard deviation o, of the bed material as shown in Table I. Altematively, when the sediment is nonuniform, effective sediment size, d., be used in eqs 15, 16 and 17 instead of d the median size, the former being given by d,Ja=0.925 6,° wes (18), for o,>1.124574 Field data To assess the relative accuracy of the above mentioned formulae, all available field data on scour around bridge piers were compiled and analysed. The Indian data on scour around pier for 17 bridges in Indo-Gangetic plain, collected by Inglis", were available. In addition, data collected by RDSO (Research Designs and Standards Organisation”"7"7*), Lucknow, on railway bridges, and some data on scour at bridge piers on Ganga canal were also used. Scour data for 55 bridges in USA published by Frochilch®, six bridges in New Zealand reported by Melville”, and for five bridges in Canada reported by Neil have also been used. Their summary is given in Table IV. In passing, it may be mentioned that not enough information is available on scour measuring equipment, even though the principles and broad circuitry used in imported equipments are known. ‘There is an urgent need to fabricate the equipment in India-and make it available to user agencies. Amalysis of Field Data for Scour Depth Using the data mentioned above, four relationships were tested for their accuracy of prediction of scour depth. These were Lacey-Inglis, Laursen- Toch, Melville-Sutherland and Kothyari et al. The comparison between observed and predicted scour depths were plotted for each of the four methods. Typical graph for comparison between the observed and computed d,, or d,. for Kothyari et al RJ GARDE AND UC KOTHYARI method is shown in Fig. 2. The results of the four methods are compared in Table V. It can thus be seen that among the four methods tested, namely Lacy-Inglis, Laursen-Toch, Melville-Sutherland and Kothyari er al, the methods by Kothyari et al. and Melville- Sutherland give results of almost the same accuracy. These methods are also superior in that these take into account the effect of flow depth, velocity, pier size and shape, and the size distribution of bed material on scour depth, Some comments can also be made about Lacey- Inglis method. The method is basically empirical and gives scour depth below the high flood level in the case of meandering rivers in flood plains in sandy materials with sediment size varying from 0.2 mm to 0.4 mm. The method, though based on a very limited data from prototype bridges, seems to give satisfactory results or oversafe values. It should not be used for rivers with cohesive or gravelly bed materials. Further, itis important that it should be used as was recommended by Inglis ie., computing D,g from eq. 3 and then finding Da. Also, since Lacey-Inglis did not independently account for scour due to nonuniform flow distribution, contraction and pier geometry and inclination, all these effects are inherent in Lacey’s scour depth, Scour in Clayey Soils When the river bed consists of clayey material different types of forces act between soil particles which resist the dislodgement of particles that Table IV ‘Summary of field data Source ‘Sediment Flow depth T ier diameter Scour depth sized mm ™ ms ‘or width m below bed level m RDSO 0.43"-1.6" 1.46-19.11 NA 2335.18 2.40-16.25 Inglis 02-039" 44-183 NA. 3113 7.60°-35.7* Upper Ganga Canal ——0.18-0.21 0.88-3.00 035-10 0.68-2.4 1.20-5.87 USA 0.25.90 058-195, 0.46-3.67 0.24-13.0 0.30-7.80 New Zealand 94-230 2738 087-421 09224 275-488 Canada 0.50 40-75 NA 150, 5.30°.9.8¢ N.A—Not Applicable, +Lacey’s Regime depth, *Measurement below water surface.‘SCOUR AROUND BRIDGE PIERS LUNE OF aoneewent 375 LEGEND UNG cama ata Ganga a1 woeanen Fig. 2. Observed vs computed scour depths by Kothyari etal. ‘method cause scour. These are Van der waal's forces, electric surface and other bonding mechanisms such as hydrogen bond, and chemical cementation between particles. Hence scour in clayey materials fs more complex and less understood than the scour in noncohesive sandy material. Unlike in the case of noncohesive sands, flow conditions at which clayey material will erode is very difficult to predict because it depends on the type and percentage of clay, quality of water and time Some investigators have tried to relate the critical shear or critical velocity to plasticity index, vane shear strength and such other properties; but these attempts are hot very successful. Some basic work ‘on scour in cohesive soils has been done by Mirtskhulava™”, One idea that he has introduced is to increase the specific weight of cohesive soils to account for increased resistance. The increase in the specific weight over the actual specific weight is proportional to its cohesion. He has also indicated that when cohesive soil is detached, aggregates of 3 to 5 mm in size come out. Hence, it may be necessary not to use the actual characteristic size of cohesive sediment but the size of aggregate soil. Because of such difficulties no rational method is available for estimation of scour depth around bridge piers in cohesive material. Hence further experimental work in the laboratory is needed under controlled conditions; in addition some field data on scour in clayey soils need to be collected. ‘Namjoshi"' and Kand™” have proposed methods for estimation of scour depth in cohesive soils, but ‘Table V Comparison of accuracy of prediction of cour depth by different methods Method % of Data points faling within given error band 30 550 BO Lacey-Inglis 39 85 100 Laursen-Toch 38 6s 98 Melville-Sutherland n 9s 100 Kothyari et a. 86 96 100 these are based on the data from one or two bridges. Hence these methods need further verification. According to Namjoshi scour depth below general bed level d, in cohesive soils does not exceed 1.5 b . Kand suggests use of Lacey- Inglis method with enhanced value of silt factor f. Seseine=F+ VC) 1) where F varies between 1.50 and 2.0 as friction angle decreases from 15° to 5° or less; here C is the cohesion in kg/em*. Scour in Gravel-Bed Rivers Gravel-bed river is that river the bed material of which is usually characterised by relatively large median size and large standard deviation. Hence, the bed material consists of material ranging from very fine to very coarse particles. It is only during relatively large flood that all the particles in the576 bed material move; as the discharge reduces the coarse particles, which cannot be moved by the flow, accumulate on the bed surface and form a layer of nonmovable particles known as armour layer or paving. For low discharge there is no sediment transport since the original material is overlain by armour layer. The standard deviation of the top layer is usually much smaller than that of underlying original material. The top layer thickness is one to two times the largest size in the bed material When the bridge pier is constructed in such a strata and the discharge is sufficiently large, the scour development would progress. During such development, the coarser particles would accumulate in the scour hole and partly inhibit further development of the scour. Ultimately the accumulated coarser material would stop further scour and the scour depth obtained would be much smaller than that in uniform material of the same dy The IRC-78-1979 code recommends that scour depth in gravel-bed rivers can be estimated using Lacey-Inglis approach involving discharge intensity q m’/sm namely DA33gip' (20) and silt factor of 24. Here q is the discharge per unit width of channel and Dj, is depth of flow calculated using g. In this connection, it may be stated that no field data have been published to support this contention in view of the fact that bed material size of the gravel bed rivers varies over a wide range. Published data of gravel bed rivers indicate the depth relationship. D=aQ'e', - Ql) see Hey and Heritage”. Here b varies between 0.33 and 0.49 and c, between ~ 0.03 and ~ 0.12. This is different from Lacey’s eq, (3). In addition such a method does not take into account the effect of pier width and its shape. Bridge foundation are normally designed for a flood of 50-year return period whereas the average annual flood has a return period of 2.33 yrs. Hence, at such a high discharge all the available sediment sizes in the bed would move and would thus destroy the armour layer or pavement formed RJ GARDE AND UC KOTHYARI earlier. Thus one can consider scour to occur with the original bed material in place without the presence of armour layer. The methods proposed by Kothyari ef al* and Melville and Sutherland? take into account the effect of sediment nonuniformity and hence armouring effect indirectly. Therefore, it is recommended that these methods be used in place of: Lacey-Inglis method using g and f=24. However, to study the relative accuracy of these methods there is need to collect scour data from gravel-bed river which are not available at present. To indicate the large variation in scour depth one can see the results obtained for a hypothetical problem solved by Garde and Kothyari™ with the following data, U=2.5 m/s, D=2.80 m, Diameter of circular pipe 5-25 m, a=almost equal to unity, 0-09, dyy=45 mm, 6,=2.125 Table VI Method ri Rank sism Kothyari er a.” 2.142 m IRC Code 1.520m Melville and Sutherland 4.800 m The large differences in scour depth predicted emphasize the need for further study of scour in gravel-bed rivers. However, it seems that since Kothyari et al. and Melville and Sutherland? methods take sediment nonuniformity into account, these methods be used in place of IRC- 78-1979 code till additional data on scour in gravel-bed rivers are available. Methods of Scour Control and Prevention Since taking the bridge pier sufficiently deep into the bed to take care of anticipated maximum scour depth and the grip length requirement is quite expensive, some attempts have been made to reduce the scour either by some modification of the pier, or some addition to it, and/or by increasing the ability of the bed to resist the scour. These methods are briefly discussed below. It may, however, be mentioned that, to the authors’ knowledge only a couple of methods discussed below have been used in prototype bridges, and hence their feasibility from the point of view of structural design, construction and economy need‘SCOUR AROUND BRIDGE PIERS to be evaluated further before these methods can be used in the field with confidence. Pier Modification Provision of the caisson or well having the diameter three times the diameter of the pier is recommended by Chabert and Engeldinger’, Shen and Schnieder’, and Jones et al.®*. While Shen and Schneider® have suggested the use of a lip, Charbert and Engeldinger suggested that the top of caisson be at 6/2 depth below the general bed level. The top surface of caisson protects the bed from scouring action of the horse-shoe vortex and thus reduces scour. The caisson top should be between 0 and 2.4 ¥/b, see Fig. 3 for definition of ¥ The efficiency of such a device can be quantified by the scour ratio Sr defined as scour with device Sp = SCOUT WH ceviee cour without device both under otherwise identical conditions. Scour ratio Sr for caisson varies from 0.30 to 0.50, see Chabert and Engeldinger* and Jones ef al. LIP (OPTIONAL) Fig. 3. Pier with cassion Stot in piers sm Tanaka and Yano”, Chiew* and others have tested slots in cylindrical piers (Fig. 4). With the optimum dimension and location of the slot in the pier in the direction of flow the scour ratio was 0.85 to 0.70 and reduction in width of scour hole from 0 to 25 per cent. The slot near the water surface reduces the effective depth of flow whereas the slot near the bed causes the jet issuing downstream. This jet deflects the downward flow in front of the pier and reduces the scour. It is felt that keeping such a slot in the pier may create structural problems and may endanger the safety of the bridge, Thomas”, Ettema”, Chiew™, and Haghighat‘' have experimented with circular collar of appropriate diameter placed around the circular pier at a certain elevation above or below the bed, (Fig. 5). The optimum diameter of collar is found to be 3b while location above the bed is 0.2 D. For this condition the scour ratio would be 0.80, while for a collar of 6 b diameter this ratio would be 0.45. Visual observations have shown that the collar of adequate diameter inhibits the growth of horse-shoe vortex and prevents it from reaching the bed; as a result the scour is reduced. Ettema®” studied the reduction in: scour when collar was placed on or below the bed; such a collar would provide nonerodible surface but will not the inhibit growth of horse-shoe vortex. Gupta and Gangadharaiah” experimented with the delta-wing-like triangular plate placed just in front of the pier as shown in Fig. 6. The two vortices released on the two sides of the triangular plate are in opposite direction to the horse-shoe vortex and hence the scour around the pier is reduced. The devices experimented with by Levi and Luna‘ are shown in Fig. 7. These included an obstacle, a plate of small height and a group of piles placed in front of the pier. Among the three a) aus eee sil O- Fig. 6 Delta-wing-like triangular plate ~ 7 cour “988 Fig. 5. Pier collar578 devices tested, the vertical plate with b/b=2, s/b=2 and 1/D 0.30 to 0.40 seems to be a better device. For such case the scour ratio was 0.30. Vittal etal. have replaced the solid cylindrical pier of diameter & by a group of three small piers of diameter 0.302 b each, and placed at an angular spacing of 120°. This was found to be effective in reducing the scour. The scour ratio obtained was 0.60 (Fig. 8). They also tested the scour reducing 0 <> -Gic -i Fig. 7. Devices used by Levi and Luna capacity by full pier group with that of cylindrical pier with slot or collar. Paice and Hey“ have tested four piles placed in front of the pier in a diamond shape formation at three prototype bridges in UK and found them to be effective; no severe scour was observed Odgaard and Wang have used IIHR (lowa Institute of Hydraulic Research) Vanes, and two plates in front of the pier (Fig. 9). Vanes or the plates develop secondary flow which persists some distance downstream and cause deposition of sediment in the scour hole. The effectiveness of vanes and plates was further increased by the use of collar. These vanes were tested on a small prototype bridge in USA and found to be effective in reducing the scour. Fig. 8 Pier group tested by He RJ GARDE AND U C KOTHYARI Fig. 9. Vanes tested by Odgaard and Wang. Riprap Protection Protecting the river bed and banks prone to erosion by large size nonmovable stones (called riprap) is an age old practice. Riprap blanket being flexible, is not weakened by slight movement or lowering of the bed. If r, is the average shear stress on the bed in N/m?, the size of nonmovable stone around the pier is given by t/120 m. If such stones are placed on finer bed material, the fine material underneath may get washed. For this reason proper gradation of armour layer is needed. Otherwise a filter needs to be provided underneath the riprap. Limited experience about riprap protection underlain by properly designed filter has indicated that it is rather difficult to place relatively thin layers of filter under deep water which is flowing. Hence, efforts have been made to provide riprap protection without filters. This has been done by Worman". According to Worman’” a geometric standard deviation of 2 can be assumed for riprap and dyis can be determined. The thickness of riprap Tat the scour hole is given by UfIgT-6(das/dys), where U,=twice the flow velocity in the river, ds is such a size of river bed material that 85 per cent of the material is finer than this size, and djs is such a size of armour layer that 15 per cent ‘material is finer than this size. Worman has stated dasdys should be less than or equal to 0.10. Twenty bridges in Sweden have been provided with riprap protection according to the above method and according to Worman no significant scour is reported. With this design criteria, no filter ig needed underneath the armour layer. Concluding Remarks The critical review of available literature and analysis of prototype scour data around bridge‘SCOUR AROUND BRIDGE PIERS piers have brought out certain major observations. During the past four or five decades a number of ‘equations have been developed for predicting the scour depth. Many of these are based on limited laboratory data and a few on the basis of limited field data. These studies have brought out the effect of flow conditions, pier diameter and its shape, sediment size and its nonuniformity and the nature of flow (clear water or sediment transporting) on scour. There are difficulties of getting proper instruments for measuring transient bed level in the scour hole and maximum scour depth in prototype bridges. Such equipment though available abroad is not available and used in India. ‘When available scour data in sandy beds are analysed using methods of Lacey-Inglis, Laursen- Toch, Melville-Sutherland, and Kothyari et al., itis found that the methods proposed by Melville- ‘Sutherland? and Kothyari et al. give more or less the same accuracy. Further, these two methods take: into account all the factors affecting scour ‘around bridge piers. Hence these are superior to the other methods. It is also concluded that Lacey- Inglis method should be used for sand bed rivers in precisely the same manner as recommended by Inglis. This should not be used for rivers with clayey or gravel bed. Not enough information is available on scour around bridge piers in clayey material. The phenomenon being very complex further laboratory studies under controlled conditions and field studies on measurement of scour are needed. In the case of gravel-bed rivers the provisions of IRC code sem arbitrary. The methods of Melville and Sutherland’, and Kothyari et al.™ seem more logical for prediction of scour in gravel-bed rivers and should be used. Yet there are no field data available to comment on the relative accuracy of prediction by these methods. Hence efforts need to be made to collect scour data in gravel-bed rivers. Several devices have been tested which would reduce scour at bridge piers or inhibit its development. These work on the following principles: (i) prevent formation or reduce ‘effectiveness of horse-shoe vortex; (ii) develop circulatory flow near the bed in the direction opposite to that of horse-shoe vortex to reduce or nullify its effect; (iif) provide device on the 579 upstream side of the pier which will scour material there and deposit it in the scour hole of the pier; and (jv) provide armour layer of adequate thickness and appropriate size distribution which ‘would inhibit scour. ‘Among the various devices, collar, vanes, and armour layer seem promising. Field studies need to be conducted in India to gain experiénce about their usage and cost effectiveness. Lastly, there is an urgent need to review codal provisions for estimation of scour, in view of available additional information. Notations Pier width or pier diameter ‘Channel width Sediment size Size of armour coat or riprap material Characteristic size of bed materials; also scour depth below bed level Clear water scour depth below bed level Scour depth below bed level in sediment transporting flow D_ Average depth of flow Lacey depth computed using the equation with Q Lacey depth computed using the equation with q Ded, Did : Lacey's still factor (=1.76 Vd) Froude number (=U/YgD) Coefficient of proportionality between D, and Diy Shape coefficient of pier Coefficient to take into account effect of sediment nonuniformity on scour Obliquity coefficient Pier length Discharge per unit width of channel Dicharge Channel slope ‘Thickness of riprap ‘Average shear stress onthe bed Average velocity of flow Local maximum average velocity Critical velocity for sediment Shear velocity (=¥2/p) Difference in elevation between river bed and top surface of caisson Opening ratio (=(B-6/B) Specifie weights of water and sediment 6 B a 4, 4, RARPOPS FeEece WbO8 =p %% Mass density of fuid @ Angle between axis of pier and the flow direction ¥ Kinematic viscosity of the fluid @ Geometric standard deviation of sediment! (dy/dyy+dsy +4) Subscript 16, 50, 84 Sediment size such that 16, 50 ‘or 84 per cent of material is finer than the corresponding size,580 References 10 n 13 “4 1s 7 18 19 20 21 B By 25 F Huber Civil Enginering ASCE 61(9) (1991) B/W Melville and A J Sutherland JHE ASCE 114(10) (1988) 1210 E M Laursen and A Toch Sth Congr IAHR Minneapolis USA (1953) 123 J Chabert and P Engeldinger Lab Nat d'Hydraulique Chatour France (1956) R J Garde Roorkee Univ Res J 8(1,2) (1965) 51 A'S Paintal and R J Garde Roorkee Univ Research J 8 (1, 2) (1965) 51 EM Laursen Jowa Highway Res Bd Bull 8 (1958) UC Kothyari Ph D Thesis Univ Roorkee (1990) A.J Raudkivi 4¢h Intl Conf Applied Numerical Modelling Taiwan (1984) R Ettema Univ Auckland New Zealand Rep 117 (1980) HUN C Breusers, G Nicollet and H W Shen J Hyd Res IAHR 18(3) (1977) 211 E™M Laursen and A Toch lowa Highway Res Bd USA Bult 4(1956) J Larras Ann Ponts Chausse’es 133(4) (1963) 411 CC Inglis Ann Rep (Tech) CWPRS Pune (1944) H W Shen, V R Schneider and S Karaki NBS US Dept Commerce Inst Appl Technol (1966) C Venkatadri, AM Rao, § T Hussain and K C Asthana J Trrig Power CBIP (1965) 35. NL Coleman /5th Cong /AHR Paris France 3 (1971) 307 SC Jain JHD ASCE 107(5) (1981) 611 HW Shen, VR Schneider and S Karaki JHD ASCE 95(6) (1969) 1919 'S V Chitale J Irrig Power CBIP 45(1) (1988) 57 R J Garde, K G Ranga Raju and U C Kothyari Res Rep Civil Engg Dept Univ Roorkee (1987) RDSO Bridges and Floods Rep No RBF-3 Prog Rep | Lucknow (1967) RDSO Bridges and Floods Rep No RBF-5 Prog Rep 2 Lucknow (1968) RDSO Bridges und Floods Rep No RBF-10 Prog Rep 3 Lucknow (1972) RDSO Bridges and Floods Rep No RBF-17 Lucknow (1991) 6 27 28 29 30 31 2 3 34 35 36 37 38 40 41 a 4a 45 47 RJGARDE AND U CKOTHYARI DC Froehlich Proc ASCE Nat Conf Hyd Engg (1973) $34 B W Melville Sch Engng Univ Auckland New Zealand Rep 117 (1975) CR Neil Proc Inst Civil Engrs Canada 30 (1965) 415 ‘Ts E Mirtskhulava CWPRS Golden Jubilee Symp Pune 1 (1966) 14 Ts E Mirtskhulava CWPRS Golden Jubilee Symp Pune 2 (1966) 333 AG Namjoshi Proc int! Sem Bridge Struct Foundation Bombay Document 3-V2 (1992) CV Kand Bridge Engineering India IX and X (1992- 1993) RD Hey and G L Heritage Bridge Engineering India IX and X (1992-1993) R J Garde and U C Kothyari Report Submitted Indian Inst of Bridge Engineering (1995) HW Shen and VR Schneider ASCE nai Mig Transportation Engineering Boston USA Paper No 1238 (1970) J'S Jones, R J Kilgore and M P Mistichelli JHE ASCE 118(2) (1992) 280 Tanaka and M Yano 12th Cong IAHR Fort Collins USA 3(1967) 193 Y M Chiew JHE ASCE 118(9) (1992) 1260-1269 Z Thomas 12th Cong IAHR Fort Collins USA 3 (1967) 125 R Ettema Civil Engg Dept Univ Auckland New Zealand Rep 216 (1980) M Haghighat ME Thesis Civil Engg Dept Univ Roorkee 1993) AK Gupta and T Gangadharaiah &1h Cong APD-IAHR Pune 2 (1992) 471 E Levi, and H Luna 9h Cong IAHR Dubrovnik (1961) 1061 N Vittal, UC Kothyari and M Haghighat JHE ASCE 120(11) (1994) Paice and R D Hey JHE USA (1993) 1061 J Odgaard and Y Wang JHE USA (1987) 523 ‘A Worman JHE ASCE 115(12) (1989) 1615