STUDY ON TURBULENT FLOWS IN OPEN CHANNELS

Abstract

In this study, a series of experiments including rough beds and smooth bed have been performed with the objectives of estimation of appropriate methodologies to estimate bed shear stress based on mean velocity as well as fluctuating components of velocity in three perpendicular directions, evaluation of resistance laws which use roughness as one of the parameters, ascertaining the presence of dip formation in velocity profiles, ascertaining the advantage of using a shift in bed location on simulation of velocity profile, variation of turbulent intensity across and along the flow, variation of turbulent energy in a vertical and assessment of performance of some of the widely used softwares in simulating experimental observations. The focus of the work has been primarily motivated by practical considerations. While the resistance laws are very much needed in analysis and design of several problems in engineering, the use of turbulence kinetic energy is usually important in many applications ranging from dissipation of energy, re-aeration process in open channels. Velocity profile is simulated using relationships of Swamee (1993). Swamee's resistance relationship is essentially developed from the resistance relationship used by ASCE Task Force on Friction Factors in Open Channels (1963). In addition, use of Manning,s resistance relationship is also evaluated. Shifting of bed to another reference level was found useful in simulation of velocity profiles. It has been observed that arrangement or packing of roughness also plays an important role. A relationship involving roughness height, von Karman constant, slope and intercept of best fit line between velocity and vertical distance ( measured with respect to reference level) is developed. It is observed that velocity equation of Swamee (1993) works well in the inner region. This agreement is relatively better when shifting of the bed is done to account for roughness effects. Also, near the free surface, the performance of this equation is found not that good as considerable deviation of it occurs from the observed velocity profile as one approaches the free surface. The deviation of velocity profile is observed and is in conformity with the earlier investigation of velocity profile in open channels, as observed by several investigators. 111 Discharge computations are found sensitive to the location of section across the flow. However, consideration of integrated effect of various sections favors the use of Swamee resistance law more strongly in case of uniform flow. A major part of the work has dealt with estimation of shear velocity using different approaches. If turbulent measurements are not available, log law based approach of shear stress estimation is found a better alternative. One of the major contributions of this work is towards estimation of bed shear velocity or shear stress using turbulence measurements. Chapter 5 clearly outlines the advantages in using the proposed approach. Variation of turbulent intensity has been investigated across and along the flow. Definitely, this variation also appears sensitive to the normalizing shear stress. The work has examined the use of six scaling parameters in Chapter 6 while studying the variation of turbulent intensity. The relative magnitude of different stresses follow the patterns reported in the literature. Use of shear velocity based on linear approaches of Grinvald and Nikora (1988) yields too low values of equivalent bed roughness in rough beds. Reynolds stress is observed to be maximum towards the bed and decrease towards the surface. However, near the bed, across flow variation are observed in different runs. Same is found true for the variation of normalized turbulence intensities. A relative comparison of turbulent stresses on different rough beds at a given Reynolds number does indicate that the turbulence is more pronounced where bed has non-uniform roughness. Numerical simulation of turbulent flow for different experimental series indicated that the nature of bed plays an important role in convergence of the solution to an acceptable level. The nature of velocity and TKE distribution in Chapter 7, suggests need for more accurate simulation using LES for a better understanding of the turbulent flow over rough bed channels. iv