EFFECT OF WASH LOAD ON TRANSPORT OF UNIFORM AND NONUNIFORM SEDIMENTS

Abstract

Knowledge of the mechanism of entrainment of sediment and its subsequent transport by the flow is required for handling many problems like soil erosion and conservation, design of stable channels, aggradation and degradation studies, reservoir sedimentation etc. Depending upon the hydraulic conditions the total sediment load in a stream may consist of bed load and suspended load together known as bed-material load. Suspended load sometimes consist of a significant quantity of sediment that is not found in appreciable amounts in the bed. This part of the suspended load is conventionally termed as wash load. Wash load mainly comes from the catchment area into the stream during periods of high rainfall and flows as such in the stream along with water and is believed not to settle on the streambed. As per generally accepted concept, the wash load is not considered to be a function of the hydraulics of flow but is assumed to be dependent on the erodibility of the soil within the catchment and hence is difficult to predict. However, the presence of wash load is considered to affect the resistance to the flow. The effect of its presence on transport of bed-material load is not well known. The present study was therefore, taken up to study the effect of the presence of wash load (fine material load) on bed-material transport. Einstein et al. (1940) were among the first to introduce the concept of wash load and Einstein (1950) related the limiting size for wash load to the percentage of sediment in the bed. Shen (1970) and Woo et al. (1986) related the limiting size for wash load to the sediment carrying capacity of the channel. According to Partheniades (1977), the actual wash load may consist of two distinct types of sediment: one having a bed load function for a limited range of discharge and other having no function of bed load. Einstein (1968) further studied deposition of fines in a coarse bed stream. Diplas and Parker (1992) also studied the deposition and removal of fines from gravel bed streams. According to Vanoni (1946), Vanoni and Nomicos (1960), Cellino and Graf (1999) and many others the resistance to flow decreases in the presence of suspended load The studies carried out by Taggart et al. (1972), Lyn (1991) and many others have shown that friction factor in rigid-boundary channel increases with increase in suspended load concentration. Kikkawa and Fukuoka (1969), Pullaiah (1978) and Arora et al. (1986) have found both a decrease and an increase in the value of friction factor with change in suspended load concentration. Every channel has a limiting capacity to transport suspended sediment through it under given hydraulic conditions and for a given sediment size. Arora et al. (1984), Celik and Rodi (1991) and Nalluri and Spaliviero (1998) and many others proposed different relationships for determining the limiting capacity ofa channel for transport ofsuspended sediment load. Einstein (1950) proposed the first semi-theoretical model for fractionwise calculation of bed load transport rates. The characteristic phenomenon of hiding of the finer particles behind coarser particles was studied first by Einstein and Chein (1953 a) and later modeled theoretically by Egiazaroff (1965). Misri et al. (1984) pointed out the non-consideration of exposure effects of coarse fractions as a weakness of these methods and presented the concept of exposure-cum-sheltering parameter for the bed load transport rates- a concept also used by Proffit and Sutherland (1983), Samaga etal. (1986) and Mittal et al. (1990). Patel and Ranga Raju (1996) proposed an empirical relationship for the fractionwise transport of nonuniform sediment using a wide range of flow conditions and sediment nonuniformity. Wu et al. (2000) proposed a probabilistic model for predicting the hiding and exposure correction factor to be used for computing the bed load and suspended load in the case of nonuniform sediment. Samaga et al. (1986 b) and Gracia and Parker (1991) proposed methods for the suspended load transport of uniform and nonuniform sediment. Kikkawa and Fukuoka (1969) found that the transport rate of bed-material increases with increasing concentration of wash load. The critical review on the subject brings out the fact that transport of the bedmaterial load in alluvial rivers is affected by the presence of wash load in suspension. However, no definite relations for computation of transport rates are available under these scenarios. The study was therefore undertaken to study in detail the effect of presence of wash load on transport of uniform and nonuniform sediments and to propose a transport law for the computation of the same. An extensive set of experiments was conducted in a tilting, recirculating flume having 30.0 m length, 0.204 m width and 0.5 m depth. . Two uniform sediments and one sediment having non-uniform size distribution were used as the bed material. The arithmetic mean size, da , and geometric standard deviation, aK, of sediment were varied from 1.03 mm to 2.73 mm and 1.15 to 2.25 respectively. The wash material used was 0.064 mm in size. Ten series of experiments were performed. The first run in each series was carried out with inflow being free from wash load in order to provide a reference set of measurements. The subsequent runs in each series were taken with increasing concentration of wash load till the deposition conditions develop. The flow conditions were selected in such a way that bed material moved only as bed load. Bed load transport rates under uniform flow condition were ascertained. Aggradation was not found to occur so long as fines continued to settle into the pores of the bed material. Equilibrium was considered to be achieved when the rate of entrainment of fines from the bed became equal to the rate of deposition. After each equilibrium run the bed material was sampled for its composition at different locations and its characteristics determined. For a given discharge and slope, the runs were taken by increasing the concentration of wash load till deposition of wash material over the bed surface started taking place. At the end, clear-water flow was allowed into the flume so that entrainment of deposited material took place. Observations were also made for bed-material transport rate for all flow conditions. In addition, data were also collected from literature for use in the study. The data were used for the development of a mathematical model for predicting the change in bed composition in the presence of wash load, for finding out the relationships for the resistance to flow in the presence of wash load, limiting capacity of wash load transport of alluvial channels, bed load transport of nonuniform sediments in the presence of wash load and suspended load transport of nonuniform sediments. The process of infiltration of fines into the pores of the coarse bed material during the transportation of wash load is modeled. The differential equation governing this process is derived and given as in dPn+ dQ dQs + a, + a1 = 0 dt ' dt 2 dx where Qs is suspended load transport, b is the width of the stream, Az is the thickness of 1 the transport layer (active bed layer), pp is the porosity ofthe bed material, a, = UbAz and a2 =-- . The above equation is derived on the assumption that only the active b Az layer thickness of the channel bed participates in the infiltration/deposition process and simultaneous transport of bed-material has no effect on this process. The predictor-corrector based finite difference numerical scheme of MacCormack (1969) is utilized for the solution of the above equation through the use of appropriate boundary and initial conditions. Data collected in the present study were used for model validation. Good agreement is obtained between model computations and the observations. Verification of Arora et al. (1986) criterion for change in value of friction factor revealed that the limit specified by them in terms of the value of for increase or US decrease in the value of friction factor does not hold good for the data used herein. Even for rigid-boundary channels the criterion does not perform well for a wide range of data. Further analysis of data collected in this study along with the data of Vanoni (1946), Vanoni and Nomicos (1960) and Cellino and Graf(1999) revealed that in closely packed non-alluvial and alluvial channels, in the absence of any noticeable change in bed form size with change in wash load concentration, there is a decrease in the value of rease in me value ot [s -1; follows the relationship friction factor with increase in the value of (s-\) . The variation of friction factor 'us ^ =1-10~5(,-1)^ /„ v 'us In the above equation / is the friction factor for sediment-laden flow, f0 is the friction factor for clear water flow, s is the relative density of the sediment and co is the fall velocity of the sediment. IV After making a number of trials using all relevant dimensionless parameters, it was found that for rigid-bed channels ^-<1.0 when Cl/8 L and -£->1.0 when C,/8 (u* d^ fu*d\ K V J <0.65 >0.65 Here u, is the shear velocity, d is the size of sediment in suspension and v is the kinematic viscosity of the fluid. Predictors for friction factor are developed for each of the above cases as below: fu*d^ 1/8 a) For C V V j 0.65 (i.e. •£-> 1.0) f Jf_ = e8x10-6(j-1'u)—s fo 1/8 b) For C (ru£ /0 10 <0.65 (i.e. -*-< 1.0) / (5-1)Ceo Us + 10 (5-1) -[2 Ceo us -5 5x10 (5-1)Co) The data collected during the study were also used to verify the existing relationships of limiting capacity of wash load/suspended load transport through channels. The relationship of Arora et al. (1984) for the limiting concentration of wash load in rigidbed channels is also applicable for alluvial channels carrying wash load. In all the experiments of the present study the original bed material was much coarser than the fine sediment recirculated as wash load. The fine sediment settled within the pores of the coarse bed and thus did not affect the exposure of the coarse particles of the bed material significantly even though it itself is sheltered by the bed material. It was found that the method of Patel and Ranga Raju for computation of transport rate of the bed material can be used based on the properties of the original bed material. A check on the methods of Samaga et al. (1986 b) and Wu et al. (2000) for transport of suspended load revealed that they are not satisfactory when applied to a wide range of data. A new method is therefore proposed for fractionwise computation of suspended load transport of nonuniform sediment using laboratory and field data covering a wide range. The suspended load transport law for uniform sediments, proposed by Samaga et al. (1986 b), was taken as the basis for studying the effect of sediment nonuniformity. The suspended load transport rate of a particular size fraction, of,, can be estimated, if its T r sheltering-cum-exposure-cum-interference parameter E, is known. Here £ = -2-1 where reff is the bed shear stress which would cause the same transport rate of a uniform size fraction, dt, as caused by r0 in a nonuniform sediment bed of that particular size fraction, dr After studying the effect of various parameters on §, the following functional relationship is proposed: \roc daj 4,=f Here xoc is critical shear stress of the arithmetic mean size da. The detailed analysis of the data led to the development of the equation fd^ log «. vT<w = 0.703 + 0.54 x log \d*J ( + 0.03 log log - \daJJ aYY \daJJ + 0.0308 Using the above equation t)s for a particular size fraction dt can be computed and hence suspended load transport of that size fraction can be computed by substituting instead of A7s di &Ys d, in the suspended load transport law, <j>s = 28 r. vi significant to note that wash load too can be calculated from the above relationship if the surface layer composition (including fines) is used in the computation of the bed material