NUMERICAL MODELLING OF AGGRADATION AND DEGRADATION IN ALLUVIAL STREAMS

Abstract

A long stretch of a stream with its bed and banks composed of unconsolidated sediments would normally be in equilibrium until it is disturbed. Most of the water resource development projects require certain changes to be introduced in the stream. As a result, the stream bed may be raised or lowered causing problems in the beneficial use of the river and hydraulic structures on it. Hence it is "important to predict stream bed variations when major changes are introduced in a stream. Prediction of bed level variation in aggradation and degradation caused by a dam is therefore of importance to hydraulic engineers. Prediction of bed level changes due to different human activities involves modelling the flow and sediment transport phenomena. Mathematical description of these phenomena can be made by a set of differential equations which are solved for depth of flow d and bed level z. Preliminary study indicated that even though a number of models exist there is no report available to indicate that they have been used to make prediction for very long period and hence about their stability for such simulations. It is also noticed that no satisfactory armour model has been developed for problems of degradation. Therefore, in the present study a numerical model for the prediction of transient longitudinal bed profiles, which can account for armouring in the case of degradation, has (ill) been developed. In this model standard step method is used to solve the backwater equation for prediction of water surface profile. The sediment transport is calculated by using method proposed by Samaga. The sediment continuity equation is solved for the bed level z, by an explicit finite difference difference scheme. For the scheme to be stable the kinks produced in the bed profile have been removed using a special smoothening procedure. The numerical scheme developed takes both rotational and parallel degradation. The parallel degradation is assumed to take place when the shear stress is less than that defined by Mittal. For prediction of the surface layer composition in parallel degradation, an armouring model has been developed based on an idea of mixing original bed material with the surface layer which is coarsening. This modelling involves computing bed-form height and defining a mixing volume. For the case of flat bed, a conceptual mixing volume of cross-sectional dimension (aD xbD ) is assumed 50 50 Using experimental observations on armour coats by Hasan the constants a and b are found to be 500 and 4 respectively. Every time sediments are removed -(transported) from this volume, the remaining content is mixed thoroughly with a compensation from the bed material, beneath which original size distribution prevails. This model has been verified using experimental data of Gessler, Little and Mayer and (iv) Jaswant Singh. The model is also capable of predicting surface layer composition as a function of time which has been tested with data collected by Garde et al. The model has been used to study degradation in a hypothetical run as well as in Middle Loup river downstream of Milburn dam. There is a good agreement between the observed and computed bed levels in the latter case. Aggradation profiles for the river Sutlej upstream of Bhakra dam are computed. The computed and observed profiles are in good agreement . In order to predict the future bed levels, stream flows. are synthesised using Thomas Fiering model. Using the observed flows for the period 1959 to 1981 and the synthesised flows bed profiles are computed for 40 years.