Experimental and numerical investigation of DNAPL infiltration and spreading in a 2-D sandbox by means of light transmission method

Chlorinated solvents such as trichloroethene (TCE) and tetrachloroethene (PCE) are widespread groundwater contaminants often referred to as dense non-aqueous phase liquids (DNAPLs). Accuracy description of the spreading behavior and configuration for subsurface DNAPL migration is important, especially favourable for design effective remediation strategies. In this study, a 2-D experiment was conducted to investigate the infiltration behavior and spatial distribution of PCE in saturated porous media. Accusand 20/30 mesh sand (Unimin, Le Sueur, MN) was used as the background medium with two 70/80 and 60/70 mesh lenses embedded to simulate heterogeneous conditions. Dyed PCE of 100 ml was released into the flow cell at a constant rate of 2ml/min using a Harvard Apparatus syringe pump with a 50 ml glass syringe for two times, and 5 ml/min water was continuously injected through the inlet at the left side of the sandbox, while kept the same effluent rate at right side to create hydrodynamic condition. A light transmission (LT) system was used to record the migration of PCE and determine the saturation distribution of PCE in the sandbox experiment with a thermoelectrically air-cooled charged-coupled device (CCD) camera. All images were processed using MATLAB to calculate thickness-averaged PCE saturation for each pixel. Mass balance was checked through comparing injected known mounts of PCE with that calculated from LT analysis. Results showed that LT method is effective to delineate PCE migration pathways and quantify the saturation distribution. The relative errors of total PCE volumes calculated by LT analysis at different times were within 15% of the injected PCE volumes. The simulation are conducted using the multiphase modeling software T2VOC, which calibrated by the LT analysis results of three recorded time steps to fit with the complete spatial-temporal distribution of the PCE saturation. Model verification was then performed using the other eight recorded time steps. Simulated results showed that the model could successfully reproduce the migration pathways and distribution configuration observed from the laboratory experiment and LT analysis, excepted for a smaller pool height on the lenses, and a lower saturation values in PCE accumulation area due to local heterogeneities. Due to the influence of water flow, the PCE distribution was asymmetrical, and the PCE distribution area, pool height as well as PCE saturation in the accumulation region at right side was much greater. Acknowledge: This work is financially supported by the National Nature Science Foundation of China grants No. 41030746 and 41172206.