ROUTING OF SUSPENDED SEDIMENT THROUGH GRAVEL BED RIVERS

Abstract

Bed material extraction from bed and banks of channel (erosion) and its subsequent deposition at any other location by the flowing water is an essential part of flow in an alluvial channel. Deforestation and logging are responsible for worsening the disastrous effect of flood generated by extreme rainfall and subsequent sediment discharge (Bathurst 2010). Hence knowledge of sediment entrainment and deposition mechanism and its associated phenomena have become essential for a hydraulic engineer for various purpose as such design of stable channels, design and safety of hydraulic structures, handling soil erosion problem etc. Out of total load a river carries, there exist two major types of load depending upon the hydraulic condition. They are bed load and suspended load; together called bed material load. Suspended load may further consist of another type of load called wash load, not appreciably available in the channel bed and banks. These types of load join flow from the catchment area and supposed to have no relation with the hydraulics of the flow. It only depends upon the erodibility of catchment area. These types of load join river during periods of heavy rainfall and is believed not to settle on the stream bed. Several definitions are available for wash load based on different hypotheses. Einstein et al. (1940) introduced the concept of wash load and Einstein (1950) defined wash load as particles of those sizes which correspond to less than 10% of the size of bed material. A high flow can carry large particles whereas low flow can’t; i.e., no specific size of wash load can be attributed. This concept was forwarded by Shen (1970) and Woo et al. (1986). Presence of fine sediment in the flow has several hydraulic and environmental consequences. Vanoni (1946), Vanoni and Nomicos (1960), Cellino and Graf (1999), Samaiya (2009) and many others observed decrease in flow resistance due to presence of fines. Yano and Daido (1964), Taggart et al. (1972), Lyn (1997) and many others reported increase in friction factor i whereas Kikkawa and Fukuoka (1969), Arora et al. (1986) and many others reported conditional increase and decrease. Presence of suspended sediment in the flow also affects the velocity distribution by affecting the Von Karman constant. Many reporters have reported different opinions about the velocity distribution. Vanoni and Nomicos (1960), Elata and Ippen (1961), Holtrof (1985) and others observed a decrease in Karman constant with an increase in suspended sediment concentration and mentioned that effect of suspended sediment is limited to the near wall region. Lau (1983), Coleman (1986), Parker and Coleman (1986), Vetter (1986), Cioffi and Gallerano (1991) observed a change in wake function due to presence of suspended sediment. Bed load transport rate also gets affected by the flowing suspended sediment. Very few studies are carried out to study the effect of suspended sediment on the bed load transport. Simons et al. (1963) mentioned a conditional increase or decrease in the bed load transport rate with suspended load. Colby (1964) proposed a graphical solution to the estimation of bed load transport in the presence of suspended load. Wan (1985) observed a decrease in total load transport at low flow intensity and an increase at high flow intensity. From literature review, it was observed that, there is no definite relationship available in literature for bed load transport under the influence of suspended transport. On the environmental side, suspended sediment has several adverse impacts such as benthic smothering, irritation of fish gills and transport of sorbed contaminants (Davies-Colley and Smith 2007). It reduces or even kills the aquatic biota, macrophytes that form the base of riverine ecology (eg. Cline et al 1982; Lewis 1973a,b). Suspended sediment even harms or kills the animals ranging from microorganism to large fishes (eg. Lemly 1982; Barton 1977). 1987. They reduce the availability of oxygen in water, reduce the space for fish spawning, bury the eggs and fry (eg. Bash et al. 2001). In order to study the various effects of suspended sediment on hydraulics and sediment transport capacity of flow, an extensive set of experiments were conducted in a tilting flume using 5.2 mm, 2.7 mm and 1.9 mm size uniform gravels as bed material and 0.062mm size uniform sediment as suspended material. Seven series of experiments each having 10-12 individual experiments were conducted. Each series has its own hydraulic properties but each successive run of a series was at an increasing suspended concentration. The first run of each series was a clear water run (water free of suspended sediment). Clear water run facilitated a set of reference ii measurements. For all the runs, the flow condition was such that bed material moved purely as bed load only. The flow was steady and uniform. No aggradation or degradation was observed during or after each run. After each run, bed material was sampled from three locations along the flow and over the full depth of bed. The successive runs with increasing concentration in the flow were continued till the fines completely filled the bed pores. At the end of sediment laden run, clear water was allowed so that the fines deposited in the bed get entrained in the main flow. Point velocity distribution data that corresponds to the near wall region (y+ B 0:2) were picked up and plotted on logarithmic scale to check the variation of Karman constant. It was observed that logarithmic velocity distribution fitted well with the data but the Karman constant were different for runs of different concentrations. Compared to the clear water flow, the Karman constant gradually decreased with increasing suspended sediment concentration. A relationship that fits well the variation of Karman constant with sediment concentration has been proposed and is given as = 3 × 10−10C2 − 5 × 10−6C + 0:407 Where is Karman constant and C is suspended sediment concentration in ppm.Arora et al.’s (1986) conditional criteria of increase or decrease of friction factor with the limiting value of C! US are tested with the data of present study. It is found that Arora et al. (1986) criteria doesn’t hold good with the present data. It is observed that as the concentration of suspended sediment increases, the friction factor decreases. Apart from the present study data, data of Vanoni (1946), Vanoni and Nomicos (1960), Cellino and Graf (1999) and Khullar (2002) are taken to arrive at a new relationship for change in friction factor due to the presence of suspended load. A new relationship relating the friction factors and the parameter (s − 1)C! US ,is developed as; 0:985 − f f0 = 8 × 10−6(s − 1)C! US Where f and f0 are respectively friction factors for sediment laden flow and clear water flow. s is relative density, ! is the fall velocity of a sediment, U is the flow velocity and S is the channel slope. The fine sediments completely filled the pores of the gravel bed, specially the pores of the top gravel layer. But, since the gravel on the bed is large enough in size that fine sediments in pores don’t hinder the coarse gravel exposure though fines are sheltered by coarser, it is supposed that the bed material properties remained same as that of parent bed materials. iii Visual observation showed that fine sediment infiltrated down to the bottom of the bed composed of 5.2 mm and 2.7 mm gravel whereas only half way down infiltration was observed in bed of 1.9 mm gravel. The deposition gradually piled up to the surface. The proportion of fine sediments infiltrated in the bed at the upstream section of the channel was found to be more than that at the downstream section. During entrainment run, at lower discharge, only the top most layer of gravel was cleaned up. The flow entrained deeper and deeper as the discharge was increased. But from visual observation, it is seen that complete entrainment of fines is not possible until and unless the gravels hiding the fines also gets washed away. The process of infiltration of suspended sediment during its routing was modeled using flow and sediment continuity equations and flow momentum equation. The governing differential equation for the process is @PG @t + a1 @Qs @t + a2 @Qs @x = 0 Where Qs is suspended load transport, PG is porosity of bed material, a1 = −1~Ub z and a2 = −1~b z , b is the width of the channel and z is the thickness of active bed layer, x and t are flow direction and time ordinates respectively. Predictor and corrector based finite difference numerical scheme of MacCormack (1969) is used to solve the above equation along with initial and boundary condition. Validation of the above model gave a good agreement between computed and observed porosity. Some of the well known bed load transport equations (Meyer-Peter and M¨uller 1948; Misri et al. 1984 and Patel and Ranga Raju 1996) are tested with the present data. None of the relationship found to perform satisfactorily in predicting the bed load in the presence of suspended sediment. Probable reason for the poor performance may be the exclusion of suspended sediment effect on bed load transport. After attempting several other ways to incorporate the suspended sediment effect on the bed load transport, a more fruitful attempt came out as the inclusion of change in hydraulic parameter as a function of change in friction factor. Modified flow velocity and flow depth were obtained from the decreased friction factor. Incorporating these modified parameters in the original relationships gave much better results. The best estimation of bed load transport rate for the present data is by Misri et al. (1984) with incorporation of modified parameters. The result showed just 10% of observed data falling out of range of ±30% error band. iv The turbulence characteristics of flow is analyzed with the instantaneous three dimensional velocity components measured by Acoustic Doppler Velocimeter (ADV). The vertical distributions of the streamwise velocity are found to follow the standard velocity distribution pattern. It is observed that streamwise flow velocity slightly gets increased with an increase in concentration. This supports the hypothesis that modified velocity calculated (which comes bit higher) from the reduced friction factor. The streamwise turbulence intensity (TIu) is found to attain maximum value in the near wall region. Below and above that region, TIu decreases. It is also observed that TIu of the successive runs gradually decreases indicating the effect of suspended sediment and bed load transport on the turbulence intensity. It is observed that turbulence intensity decreased with increased concentration supporting the reduction/dampening of turbulence intensity. The rough elements present on the bed are supposed to be responsible for the dampening of intensity. The Reynold shear stress attains maximum value in the wall region and decreases towards bed. The turbulent kinetic energy (TKE) synchronizes well with the turbulence intensity and Reynold stress. Ejection and sweep events dominated the flow structure. They well synchronize each other. In the middle part of flow depth, the ratio of contribution of sweep to ejection is found to be about 0.8.