Experimental verification of the partial mixing model

Abstract

Several tracer mathematical models such as the piston flow model, perfect mixing model, partial mixing model, dispersive models and discrete-state compartment models are currently used to calculate residence times of groundwater, typically between recharge and discharge areas. These models are based on theoretical considerations that are not subjected to experimental verification prior to their application in field studies. Therefore, the suitability of these tracer models to complex field situations or even simple laboratory controlled-conditions has not been experimentally documented. This study experimentally verifies the partial mixing model. It determines the extent to which the model is applicable to simple laboratory conditions and ultimately the more complex natural systems of groundwater. The objective is achieved by simpulating groundwater flow through porous media using a packed column, one meter in length and 15cm internal diameter. The experiment was performed using two flow rates, 25 and 50 ml/min. Sand and angular rock fragments were used separately as packing materials. Sodium chloride was used as a tracer. Verification of the model was conducted for three different types of groundwater flow: horizontal flow, vertical flow and flow with horizontal and vertical flow components. This was achieved by orienting the packed column in the horizontal, vertical and inclined positions, respectively. Simulation of mixing by the partial mixing model was compared to that of the axial dispersion model. Results of the simulation were obtained by fitting the two models to the experimental data obtained from the packed column. Results of the two models show that the partial mixing model and the axial dispersion model behave similarly. This further confirms the experimental verification of the partial mixing model.