Problem 6 -Air Displacement of a NAPL from a Laboratory Column (rad)

This problem is an adaptation of the numerical simulation of a laboratory experiment presented by Ho (1995). It is also the problem no. 6 in TOUGH3/TMVOC user manual. The flow system is a sand column 18 cm high with a cross sectional area of $3.4\times 10^{-4}m^2$. The mesh has 18 evenly spaced elements with a spacing of 1 cm; two additional elements, placed on top and bottom of the column, are used to set different boundary conditions. The simulation consists of four parts that involve (1) obtaining a gas-static equilibrium as initial state, (2) injection of a NAPL consisting of a mixture of xylene and toluene, (3) redistribution of NAPL under gravity and capillary forces, and (4) removal of NAPL by airflooding. The first three simulations are run in isothermal mode and last simulation is run in non-isothermal mode.

The first part of the simulation obtains a gas-static equilibrium by specifying constant conditions of P = $1.01 \times 10^5$ Pa and T = 20 °C for inactive element AJ 1 at the bottom of the column. Initially, the column contains gas phase only with no VOC present. Gas composition is specified as mole fractions of 20.79 % for oxygen and 78.21 % for nitrogen; the remaining 1% mole fraction corresponds to water vapor. The system is initialized with three-component (water, oxygen, nitrogen) and a constant temperature for the simulation. The pressure-equilibrated conditions obtained are introduced as data block INCON in the input file for the second part of the simulation.

The second part of the simulation involves the injection of a liquid mixture of o-xylene (component #4) and toluene (component #5) into the top of the column (grid block A11 1). We use the initCType "FM1L123004' for reading the INCON data section (because INCON is from part 1 simulation which has less mass components). 'FM1L123004' represents that the initial condition data in the INCON section is in 1 line for each elements. Primary 1,2,3, and 6 are read from data number 1,2,3 and 4 respectively, and the primary variables 4 and 5 are 0.0. Constant boundary conditions are used at both ends of the column, by specifying numerically infinite volumes for the two boundary elements in block ELEME. The NAPL mixture is injected at a constant total rate of 1 ml/min for 216 s, to inject a total volume of 3.6 ml of NAPL. The various constants used to compute the thermophysical properties of oxylene and toluene are contained in data block CHEMP, where this time NHC = 2 is assigned. The Stone I three-phase relative permeability functions (IRP = 32 ) and Parker’s three-phase capillary pressure functions (ICP = 31) are selected for the simulation, respectively. Molecular diffusion is included for all components in all phases.

The third part of the simulation models redistribution of the injected NAPL under gravity and capillary forces over a two-day period, with both sides of the column closed. These boundary conditions were realized by simply removing the connections to the two boundary elements in block CONNE, so that the two connections across the boundary involve unknown elements and will be ignored by the code. The initial conditions for this part of the simulation are the conditions at the end of the second part of the simulation. Block GENER has been removed to stop the injection of NAPL. The SAVE file generated by the second part of the simulation was placed into the input file. It is seen that the NAPL has drained towards the bottom of the column. Composition of NAPL and gas phases is rather uniform throughout. Details can be seen in the output files.

The fourth and final part of the simulation consists of air injection into the bottom of the column for a time period of 7200 s (120 minutes). Again, the initial conditions for this part of the problem are the conditions at the end of the previous simulation. GENER data specify injection of oxygen and nitrogen at a total rate of $1.2 \times 10^{-5}$kg/s and in proportions that correspond to mole fractions of 21 % and 79 %, respectively. Both isothermal and non-isothermal simulations are run for this part of simulation.

Simulation results for all the 4 parts match the corresponding results from TOUGH3/TMVOC simulations, except the non-isothermal simulation in part 4 which had not been run in TOUGH3/TMVOC simulation.

Part 1:

Input Files: INFILE_rad1.zip

Output Files: output files

Part 2:

Input Files: INFILE_rad2.zip

Output Files: output files

Part 3:

Input Files: INFILE_rad3.zip

Output Files: output files

Part 4:

Input Files: INFILE_rad4_isothermal.zip, INFILE_rad4_non-isothermal.zip

Output Files: output files isothermal simulation, output files non-isothermal simulation