DISPERSION OF CONSERVATIVE POLLUTANTS IN CLEAR WATER AND SEDIMENT LADEN FLOWS IN OPEN CHANNELS

Abstract

Whenever rivers are used as sinks for intentional or accidental spillage of effluents resulting from the activities of the society, it becomes necessary to calculate the concentration of the effluent downstream of its point of entry into the river to ensure whether the required environmental standards regarding permissible concentration level of the effluent in the stream are being satisfied or not. If the pollutant is conservative, after a certain distance the decrease in its concentration due to its mixing with ever-increasing volume of water is a result of the longitudinal dispersion process. In the present investigation experiments were designed to check whether the one-dimensional Fickian-diffusion equation with a constant longitudinal dispersion coefficient, D adequately describes the dispersion of conservatice pollutants in open channels. The data collected in the present study in clear-water flow and sediment laden flow alongwith the data from previous investigators have been used for this purpose as well as to develop an alternative method of obtaining the c - t curve downstream of the section of injection of tracer. Experiments were conducted in a recirculating flume of 0.40 m width and 17 m length. Firstly, the experiments were carried out in clear-water flow and these were followed by experiments in sediment laden water using smooth and rough channel beds. Rhodamine B dye was used as a tracer whose passage was monitored in the flume using a 10-005 R Turner Fluorometer of continuous flow type. Two uniform sands of size 0.082 mm and 0.058 mm were used as sediment. The Reynolds number of the flow ranged from 6 x 10 to 1.4 x 105 , Froude number from 0.25 to 0.95 and concentration of sediment in suspension from 90 ppm to 5000 ppm. These data were analysed alongwith the data available from previous studies on dispersion. (ii) Predictors for D. given Elder, Sumer, Krenkel, Yotsukura and Fiering, Glover, Thackston and Krenkel, McQuivey and Keefer, Liu and Fischer were verified using the data from the present study and the available data from previous studies. It has been shown that these relations yield values of D. which are far from the observed values. The effect of variation in D. on the predicted concentrationdistribution downstream of tracer injection section for an actual river flow was also examined and it was found that predicted concentration distributions are quite sensitive to the variation in the value of D. used in computations. The similarity analysis approach of Day and Wood has been examined in detail and it has been shown that with proper choice of dependent and independent variables, the method can be sat sfactorily used for predicting the concentration-distributions. The main findings of the present study are as follows : (i) Taylor's model of longitudinal dispersion based on one-dimensional Fickian diffusion equation with a constant longitudinal dispersion coefficient, D. , is not suitable for describing the dispersion in natural streams as D. tends to increase in the downstream direction. (ii) Many of the well known predictors for D. do not yield reasonable values of D. . (iii) Similarity analysis of concentration distributions initially proposed by Day and Wood and substantiated by Beltaos is applicable over a very wide range of variables. Based on dimensional analysis empirical relations have been obtained for the various scaling parameters appearing in the foregoing similarity curve. These are superior to those proposed earlier by Day and Wood. (iv) The presence of suspended sediment does not have a significant effect on the longitudinal dispersion process other than its possible effect on the fric tion factor within the range of sediment concentrations used