Published online 25 February 2008
Published in Vadose Zone J 7:316-324 (2008)
DOI: 10.2136/vzj2006.0119
© 2008 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
SPECIAL SECTION: TOUGH2
Modeling Multiphase Organic Spills in Coastal Sites with TMVOC V.2.0
Alfredo Battistellia,*
a RISAMB Dep., Snamprogetti SpA (a Company of Saipem), Via Toniolo 1, 61032 Fano (PU), Italy
* Corresponding author (alfredo.battistelli{at}snamprogetti.eni.it).
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.
Received 23 August 2006.
 |
ABSTRACT
|
|---|
Contaminant spills are frequently encountered in coastal sites where many industrial plants are located. Refineries and petrochemical plants are often built close to the sea for easy transport of crude oil and final by-products. The migration of organic compounds spilled in the subsurface of coastal sites can be influenced by the effects of seawater intrusion in the aquifers discharging to the sea. An improved version of TMVOC, belonging to the TOUGH2 family of numerical reservoir simulators, can model the migration of multicomponent organic mixtures under multiphase conditions accounting for the effects of sodium chloride dissolved in the aqueous phase. The thermophysical properties of groundwater as a function of temperature, pressure, and salinity are evaluated following the basic approach used for saline brines in EWASG, a specialized thermodynamic module of TOUGH2 originally developed to model geothermal reservoirs. Simulations of a multicomponent organic spill using a two-dimensional vertical numerical model show the effects of seawater intrusion on the distribution of contaminants within the aquifer. The effects of the construction of an impervious wall on the nonaqueous phase liquid (NAPL) plume migration, as a means to contain the spreading of plume toward the sea, are also investigated.
Abbreviations: asl, above sea level • BTEX, benzene, toluene, ethylbenzene, xylene • EOS, equation of state • NAPL, nonaqueous phase liquid • NCG, noncondensable gas • VOC, volatile organic compound
|
INTRODUCTION
|
|---|
Coastal sites contaminated by hydrocarbons and organic solvents are often encountered, as many industrial plants are located near the coast. In Italy, for instance, many refineries and petrochemical and chemical plants are located along 7400 km of Italian coastline (Rusi et al., 2005). The organic contaminants spilled in these sites can reach the sea as both free NAPL and dissolved in the groundwater. The design and implementation of containment and remediation activities in these sites should properly take into account both the processes linked to the three-phase flow of volatile organic compound (VOC) mixtures as well as the density-driven groundwater flow that controls the extension of seawater intrusion in coastal aquifers. Due to the higher density, the seawater penetrates inside the coastal aquifer at distances from the coast depending on the aquifer thickness and hydraulic conductivity and on the amount of flow discharged to the sea. The lighter groundwater flows above the seawater intrusion with a concentration of discharged flow in the upper aquifer section. The presence of seawater intrusion can then affect both the migration of VOCs dissolved in the groundwater and that of a NAPL lens migrating at the vadose zone–aquifer interface. Modeling tools able to simulate the three-phase flow in coastal aquifers are then necessary for the proper design of remedial actions and further interpretation of postoperation monitoring. The simulation of density-dependent flow in coastal aquifers is a classical field of application of numerical simulators; among those conventionally employed are codes like SEAWAT-2000 (Langevin et al., 2003), SUTRA (Voss and Provost, 2002), and HST3D (Kipp, 1997) with capabilities to model three-dimensional saturated–unsaturated flow, the transport of multiple solutes, and heat. Examples of codes available to model three-phase flow as required in environmental applications dealing with the migration of NAPL spill in the subsurface include NUFT (Nitao, 2000), STOMP (White and Oostrom, 2000), and TMVOC (Pruess and Battistelli, 2002). Only a few codes can afford the simulation of three-phase flow with the aqueous phase properties dependent on salt concentration, such as UTCHEM (Reservoir Engineering Research Program Center for Petroleum and Geosystems Engineering, 2000).
While three-phase compositional flow capabilities were available with TMVOC V.1.0 (Pruess and Battistelli, 2002), the modeling of flow of a salt solution with variable salinity was possible within the TOUGH2 environment using the EOS7 or EOS7R thermodynamic modules (Pruess, 1991; Oldenburg and Pruess, 1995a,b) and the EWASG module (Battistelli et al., 1997). A first step toward the coupling of three-phase flow and density-dependent groundwater flow was made by including the EOS7 brine treatment into the T2VOC code (Sanese et al., 2003). The necessary capabilities are now available within an improved version of TMVOC simulator, which can be used to study the migration of organic contaminant mixtures in coastal sites under natural as well as anthropically modified flow conditions.
|
Materials and Methods
|
|---|
TMVOC Reservoir Simulator
TMVOC V.1.0 (Pruess and Battistelli, 2002) is a numerical simulator developed for three-phase nonisothermal flow of water, a user-defined set of gaseous species, and a mixture of VOCs in three-dimensional heterogeneous porous media. TMVOC, based on the M2NOTS code developed by Adenekan (1992), is an extension of the TOUGH2 general-purpose simulation program (Pruess et al., 1999). It was originally designed for application to contamination problems that involve hydrocarbon fuel or organic solvent spills in the saturated and unsaturated zones.
The main enhancements in TMVOC V.2 relative to V.1 include the following. First, a new class of compounds, dissolved solids treated as tracers, is added (Battistelli, 2004); when modeling dissolved solids, the complete disappearance of the aqueous phase cannot be treated. Second, organic compounds can be degraded by reactions mediated by microbial populations following a generalized multiple Monod's kinetic law, accounting for the concentration of substrates, electron acceptors and nutrients (Battistelli, 2004). Third, all mass components can adsorb at equilibrium on the rock matrix following a generalized adsorption isotherm that can replicate the linear, Freundlich and Langmuir isotherms. Fourth, all mass components can optionally decay following a first-order kinetic law. Finally, NaCl can be specified among the dissolved solids; brine thermophysical properties are then computed as a function of NaCl content in the aqueous phase basically following the EWASG formulation (Battistelli et al., 1997; Aquater, 2003). The halite precipitation is not modeled in the three-phase version of TMVOC V.2, whereas it can be simulated in a specialized version called TMGAS, developed for the injection of gas mixtures in geological structures (Battistelli et al., 2003).
In TMVOC V.2 the thermodynamic calculations are still performed on the more convenient mole fraction basis, but the secondary parameters are internally stored on a mass fraction basis. Mass balance equations are then assembled on a mass basis as made by standard TOUGH2 EOS modules, for a better compatibility with them. The mass components tracked by TMVOC V.2 are assumed to be distributed under local thermodynamic equilibrium conditions in the three possible flowing phases: gas, aqueous, and NAPL. Any combination of the three phases and related possible phase transitions are modeled by TMVOC V.2 as shown in Fig. 1
, with the limitations specified above when dissolved solids are treated. Single-gas phase conditions can always be specified for boundary elements to simulate atmospheric conditions often needed for environmental applications. While molecular diffusion is modeled, the mechanical dispersion is not taken into account. Special TOUGH2 modules that include a conventional Fickian model for hydrodynamic dispersion have been developed (Oldenburg and Pruess, 1993, 1995a; Wu and Pruess, 2000) but have not yet been coupled to TMVOC. In heterogeneous media, dispersion is often caused by mass exchanges between pore regions with different fluid mobilities (Coats and Smith, 1964; Harvey and Gorelick, 2000); such effects can be modeled with TMVOC using the method of "multiple interacting continua" (MINC; Pruess and Narasimhan, 1985). A detailed description of the thermodynamic and numerical formulations of TMVOC V.1.0 is given by Pruess and Battistelli (2002); the treatment of brine properties as a function of NaCl content derived from the EWASG module is described by Battistelli et al. (1997).
View larger version (12K):
[in this window]
[in a new window]
|
FIG. 1. Phase combinations and phase transitions modeled by TMVOC V.2: w, g, and n stand for aqueous, gas, and nonaqueous phase liquid. Shaded phase combinations cannot be modeled when dissolved solids are treated.
|
|
Other new features of the EOS module are as follows. The effect of noncondensable gases (NCGs) dissolved in the brine is accounted for when evaluating brine density following the approach suggested by Garcia (2001). In addition, the NCGs' internal data bank has been expanded from the 8 initial gases to 23 by adding new inorganic gases and hydrocarbons. Finally, the primary variable tracking the NaCl content in the brine is the salt molality; the mole fraction of solutes in the aqueous phase is computed assuming the salt is completely dissociated, which is a reasonable assumption at the conditions commonly found in coastal site applications.
TMVOC V.2 has been successfully verified by reproducing both sample problems and verification of test results previously obtained with TMVOC V.1, TMVOCBio, and TOUGH2-EWASG codes.
Simulation of a NAPL Spill in a Coastal Site
Although the example modeled is highly simplified, the main hydrogeologic parameters and the features of the simulated organic spill are derived from those experienced in a coastal site where characterization studies and remediation activities have been performed during the last three decades. The spill takes place at low rates within the vadose zone; it produces a lens of NAPL migrating preferentially in the direction of the water table gradient. A two-dimensional vertical grid perpendicular to the coastline is used; in this way, it is assumed that the NAPL spill takes place along a line parallel to the coast, as could happen with distributed leakages along buried pipelines, sewers, or from multiple storage tanks, as actually happened in the mentioned coastal site. The simulated containment works are represented by the construction of a vertical impervious wall parallel to the coastline from the surface to a depth of 5 m below sea level. The wall can stop the floating NAPL lens, while the aquifer can flow beneath, transporting the VOCs dissolved in the groundwater. The modeled impervious wall represents only one of the necessary containment and remediation works performed at the site. They included the recovery of NAPL using extraction wells equipped with dual-pumps, the construction of a well barrier for the withdrawal of contaminated groundwater, and the stopping of the NAPL spills.
Conceptual Model and Simulated Scenarios
The conceptual model is shown in Fig. 2
. It is a two-dimensional section 500 m in length and 35 m thick, of which 5 m are above and 30 m are below the sea level. The section has a width of 1 m. The spill point is located at a distance of 199.5 m from the coastline and at an elevation of +3.25 m asl. The impervious wall is 1 m wide and is located at 99.5 m from the coastline, reaching a depth of –5 m asl. The aquifer thickness increases from the right boundary to the coast, where its bottom is 30 m below the sea level; the unconfined aquifer flows from right to left, with an average hydraulic gradient of 8 x 10–3 m m–1. The rock domains distribution is shown in Fig. 2, and their main petrophysical properties are listed in Table 1
. The relative permeability and capillary pressure curves for three-phase systems are described according to the Stone (1970) and Parker et al. (1987) models, respectively. The corresponding parameters are listed in Tables 2
and 3
for the relative permeabilities and the capillary pressure, respectively. The simulations are performed at a constant temperature of 20°C. Atmospheric boundary conditions are fixed at the grid top by specifying a constant absolute pressure of 1.013 x 105 Pa and a gas phase composition with water vapor and air mole fraction of 0.868E-02 and 0.99132, respectively, corresponding to a relative humidity of 60%, which can drive a diffusive flux of water vapor from the vadose zone, where the gas phase is saturated by water vapor, to the atmosphere. Lateral boundary conditions are specified at the left and right grid sides assuming a hydrostatic pressure distribution with water table elevation of 0 and +4 m asl, respectively. The groundwater has a NaCl concentration of 1000 mg kg–1, whereas seawater salinity is fixed at 36,000 mg kg–1.
View larger version (12K):
[in this window]
[in a new window]
|
FIG. 2. Rock domain distribution on the two-dimensional vertical section. Water table values of +4 and 0 m above sea level (asl) are assigned on the right and left boundaries, respectively.
|
|
The applied boundary conditions are listed in detail in Table 4
. For this application, the formation heterogeneities, the seasonal water table fluctuations, and the effects of sea tides have been neglected. Sea tides are usually of minor importance in Italian industrial sites located on the Mediterranean Sea coast, owing to the limited fluctuation of seawater level. A constant meteoric recharge made up of pure water was assumed equal to 31.6 and 12.6 mm yr–1 for the clean coastal sands and the silty sands, respectively. The vertical section is discretized with 29 layers and 65 columns for a total of 1885 elements. The vertical spacing is 0.5 m asl and is progressively increased in the lower section. The horizontal spacing is reduced to a minimum of 1 m at the sea boundary and near the wall and spill point locations, where compositional gradients are higher. The simulations employ the automatic time-stepping procedure available within the TOUGH2 codes. If convergence of mass balance equations cannot be achieved within a prescribed maximum number of Newton–Raphson iterations, the time-step size is reduced and a new iteration process is started; if convergence is achieved within a prescribed number of iterations, the time-step size is doubled for the next time step. Convergence is achieved when all mass balances equations are solved within a prescribed relative error 
1. An absolute convergence criterion is used when the accumulation terms are smaller than a prescribed value 2 (Pruess et al., 1999). For the simulations performed, 1 = 2 x 10–5 and 1 = 1.
To highlight the effects of seawater intrusion, the simulation of spill was performed considering both a salt and a freshwater boundary at the left grid side. Three different spill scenarios are discussed. Case A assumes freshwater conditions at the sea boundary; a constant NaCl content of 1000 mg kg–1 has been assigned to both aquifer lateral boundaries. Case B assumes seawater conditions at the sea boundary; a constant salinity of 36,000 mg kg–1 has been considered for the seawater. Case C assumes the same boundary conditions of Case B, but with the presence of the impervious wall constructed after 1.5 yr of spill.
The simulations require several consecutive steps: (i) set up of initial conditions at the lateral left and right boundaries; (ii) steady state of two-dimensional grid, controlled by gravity and capillary forces and subject to the boundary conditions at the lateral and top grid sides, before the construction of the impervious wall; (iii) modeling of NAPL spill for 5 yr in the absence of the wall (Cases A and B); and (iv) modeling the effects of wall construction performed after 1.5 yr of spill (Case C). The NAPL migration is modeled for a total spill time of 5 yr. It is assumed that the wall construction is instantaneous.
|
Results
|
|---|
Modeling of Natural State
The initial conditions were determined by running the system to a steady state governed by gravity and capillary equilibrium under the assigned boundary conditions. The only difference between Case A and Cases B and C is that freshwater and seawater boundary conditions are specified for the former and latter cases, respectively. The density and viscosity of the aqueous phase at ambient conditions are listed as a function of NaCl content in Table 5
. The seawater density and viscosity are 2.3% and 6.0% higher than those of groundwater. The water table elevation representative of steady state (initial) conditions for Cases A and B is shown in Fig. 3
. The reduction of thickness available for aquifer flow due to seawater intrusion in Case B is responsible for the higher water table elevation close to the sea side. The seawater intrusion at steady state conditions for Case B is shown in Fig. 4
where the contouring of NaCl mass fraction is plotted. Seawater can be found at the bottom of the aquifer at more than 100 m from the coastline. The confinement of groundwater flow to the upper aquifer section near the coast results in quite different flow profiles, as shown in Fig. 5
: the groundwater flow toward the sea is concentrated in the upper grid layers for Case B, while it is almost constant over the vertical for Case A. Due to the mixing of seawater into the outflowing groundwater, there is an inflow of seawater to balance the outflow, as shown in Fig. 5.
View larger version (17K):
[in this window]
[in a new window]
|
FIG. 5. Steady-state groundwater flow at sea boundary for Case A and B. Positive fluxes exit toward the sea; negative fluxes enter from the sea.
|
|
Modeling of Spill Scenarios
The NAPL spill is modeled for 5 yr assuming a constant rate of 50 kg/day. The NAPL composition by mass is listed in Table 6
together with the main properties of the 6 mixture components considered. Linear adsorption isotherms are assumed for all the VOCs. Main properties of unweathered NAPL at ambient conditions are listed in Table 7
. They are computed with the TMVOC correlations using the VOC's properties published by Reid et al. (1987). The molecular diffusion coefficients of all VOCs in the aqueous phase and the NAPL are assumed constant and equal to 6.0E-10 m2/s. Values ranging from 7.1E-6 to 8.8E-6 m2/s, at the reference conditions of 0°C and 1.013 E5 Pa, are assigned to the VOCs in the gaseous phase. The Millington and Quirk (1961) model is used to evaluate the tortuosity effects. Observing the VOC properties, it can be seen that n-pentane has a strong tendency to partition into the gas phase; BTEX (benzene, toluene, ethylbenzene, xylene), mainly benzene and toluene, have a significant solubility in the groundwater, and n-decane has a remarkable tendency to adsorb on organic carbon. The partitioning behavior of spilled NAPL components, which can be inferred from Table 6, is clearly shown in Fig. 6
and 7
. Figure 6 shows the molar composition of spilled NAPL and the composition of VOCs dissolved in the groundwater at equilibrium with the NAPL: benzene, toluene, and p-xylene are clearly the most-soluble mixture components. In Fig. 7, the relative amount of adsorbed VOCs and of the VOCs present in the gas phase is shown.
View larger version (16K):
[in this window]
[in a new window]
|
FIG. 6. Molar composition of spilled nonaqueous phase liquid (NAPL) and volatile organic compounds (VOCs) dissolved in pure water at equilibrium.
|
|
View larger version (19K):
[in this window]
[in a new window]
|
FIG. 7. Volatile organic compounds (VOCs) adsorbed on organic carbon and VOCs in the gas phase at equilibrium with the nonaqueous phase liquid.
|
|
The 5-yr spill corresponding to Case A produces the NAPL distribution shown in Fig. 8
: the lens reaches the sea, and free NAPL is discharged outside the model grid. Figure 9
shows the distribution of total VOCs dissolved in the aqueous phase. The maximum, of course, is found where free NAPL is present. Finally, the amount of VOCs in the different phases is plotted as a function of time in Fig. 10
, which clearly shows when the NAPL starts to flow to the sea.
View larger version (14K):
[in this window]
[in a new window]
|
FIG. 8. Contouring of nonaqueous phase liquid saturation after 5 yr of spill for Case A. (asl = above sea level.)
|
|
View larger version (19K):
[in this window]
[in a new window]
|
FIG. 9. Contouring of dissolved volatile organic compound mass fraction after 5 yr of spill for Case A. (asl = above sea level.)
|
|
View larger version (23K):
[in this window]
[in a new window]
|
FIG. 10. Mass balances of volatile organic compounds (VOCs) in the different phases as a function of time for Case A. (NAPL = nonaqueous phase liquid.)
|
|
The corresponding results obtained after the 5-yr spill for Case B are shown in Fig. 11
, 12
, and 13
. Figure 12 shows that the seawater intrusion concentrates the flow of dissolved VOCs into the upper aquifer section compared with Case A. Reducing the thickness available for the groundwater flow in the last 120 m close to the seaside, the seawater intrusion reduces the aquifer flow rate and the average aquifer hydraulic gradient upstream. These changes slightly impact the migration of the NAPL lens, which is affected by the water table gradient.
View larger version (14K):
[in this window]
[in a new window]
|
FIG. 11. Contouring of nonaqueous phase liquid saturation after 5 yr of spill for Case B. (asl = above sea level.)
|
|
View larger version (22K):
[in this window]
[in a new window]
|
FIG. 12. Contouring of dissolved volatile organic compound mass fraction after 5 yr of spill for Case B. Blue dot dimensions are proportional to the dissolved NaCl mass fraction. (asl = above sea level.)
|
|
View larger version (26K):
[in this window]
[in a new window]
|
FIG. 13. Mass balances of volatile organic compounds (VOCs) in the different phases as a function of time for Case B. (NAPL = nonaqueous phase liquid.)
|
|
Case C is modeled assuming that the impervious wall is constructed after 1.5 yr of spill. Thus, the simulation is performed taking the conditions computed after 1.5 yr for Case B. The results obtained after 5 yr of spill are shown in Fig. 14
and 15
. Figure 16
presents the simulation results for a simulation time exceeding 6 yr. This time, the NAPL lens migration is completely stopped by the impervious wall so that VOCs can leave the simulation grid only dissolved in the aqueous phase or due to advective and diffusive flux within the gas phase toward the atmosphere.
View larger version (14K):
[in this window]
[in a new window]
|
FIG. 14. Contouring of nonaqueous phase liquid saturation after 5 yr of spill for Case C. (asl = above sea level.)
|
|
View larger version (21K):
[in this window]
[in a new window]
|
FIG. 15. Contouring of dissolved volatile organic compound mass fraction after 5 yr of spill for Case C. Blue dot dimensions are proportional to the dissolved NaCl mass fraction. (asl = above sea level.)
|
|
View larger version (26K):
[in this window]
[in a new window]
|
FIG. 16. Mass balances of volatile organic compounds (VOCs) in the different phases as a function of time for Case C. (NAPL = nonaqueous phase liquid.)
|
|
VOC Migration
Figure 17
shows the flux of dissolved VOCs at the sea side after 5 yr of spill for the three simulated cases. The presence of seawater intrusion concentrates the VOC flux in the upper aquifer section, whereas the contaminants are present at lower depths for Case A, albeit at fairly low concentrations. The impervious wall (Case C) is able to reduce the overall contaminant flux to the sea with respect to Case B, due to the containment of free NAPL in smaller volumes, which limits the contact between the free NAPL and the groundwater. Thus, the wall actually reduces the advective flow of contaminants in the groundwater. This is shown clearly by comparing Fig. 18
and 19
, where the concentration of VOCs dissolved in the groundwater at an elevation of –0.25 m asl after 5 yr of spill is plotted as a function of the distance from the coast for Cases B and C. Of course, additional containment works would be needed to further reduce the outflow of contaminants. The flux of VOCs to the atmosphere is due to both advection and molecular diffusion; the total flux in the gas phase after 5 yr of spill is shown for Cases B and C in Fig. 20
. The advective flux is due to the pressure gradient generated by the evaporation of volatile mixture components, particularly n-pentane, close to the spill location but also over the entire vadose zone where the NAPL plume is present. An additional contribution to the advection of the gas phase comes from the density-driven flow, as the air contaminated by organic vapors is denser than uncontaminated air. This density-driven flow has already been the subject of specific investigations such as that of Falta et al. (1989). Molecular diffusion, whose contribution dominates over the advection one, is relevant around the spill point as well as at the edge of the NAPL lens, where more VOCs vaporize in the interstitial air of the vadose zone. The maximum flow of VOCs toward the atmosphere is actually reached for Case B at a distance of 135 m from the coastline, where the NAPL plume enters the clean sands. Here, the effective diffusivity of the gas phase is higher due to the lower irreducible saturation of the aqueous phase with respect to the silty sands. For Case C, both advection and diffusion are rather high upstream from the wall, where the NAPL accumulates close to the surface. The wall stops the VOCs flux to the atmosphere downstream from the wall, but a strong flow is concentrated just upstream of it. This needs to be taken into consideration for the risk associated to the possible inhalation of VOC vapors by the personnel on site.
View larger version (17K):
[in this window]
[in a new window]
|
FIG. 17. Total volatile organic compound (VOC) flow in the aqueous phase at the sea boundary after 5 yr of spill for Cases A, B, and C.
|
|
View larger version (31K):
[in this window]
[in a new window]
|
FIG. 18. Mass fraction profile of dissolved volatile organic compounds (VOCs) after 5 yr of spill for Case B at an elevation of –0.25 m above sea level.
|
|
View larger version (32K):
[in this window]
[in a new window]
|
FIG. 19. Mass fraction profile of dissolved volatile organic compounds (VOCs) after 5 yr of spill for Case C at an elevation of –0.25 m above sea level.
|
|
View larger version (21K):
[in this window]
[in a new window]
|
FIG. 20. Total volatile organic compounds (VOCs) flow in the gas phase toward the atmosphere after 5 yr of spill for Cases B and C.
|
|
The effect of the impervious wall on the migration of individual VOCs is shown for Cases B and C in Fig. 21
. In Case B, VOCs are mainly lost by the outflow of NAPL; this explains why an average of 20% of spilled mass is lost after 5 yr of spill. A higher percentage of n-pentane is lost because of the flux to the atmosphere. When the wall is present, no NAPL flow outside the grid is allowed and VOCs are lost both dissolved in the groundwater and due to the advective and diffusive flux to the atmosphere. Benzene and toluene are lost due to the groundwater flow beneath the impervious wall, whereas n-pentane is lost because of advective and diffusive fluxes in the gas phase within the vadose zone. Due to the partitioning of NAPL components into different phases, the NAPL composition changes, as shown in Fig. 22
and 23
, where the concentration profiles of VOCs in the NAPL at an elevation of –0.25 m asl are plotted as a function of distance from the coast for Cases B and C. The NAPL is depleted upstream of more soluble compounds (benzene, toluene and p-xylene), and consequently is enriched mainly in n-decane and to a lesser extent in n-pentane. Downstream, because of n-pentane evaporation, n-pentane is lost with a related increase in n-decane and n-propylbenzene concentrations. These two hydrocarbons are those less affected by partitioning; thus, the weathered NAPL is enriched in these two components.
View larger version (32K):
[in this window]
[in a new window]
|
FIG. 22. Mass fraction profile of volatile organic compound (VOCs) in the nonaqueous phase liquid after 5 yr of spill for Case B at –0.25 m above sea level.
|
|
View larger version (28K):
[in this window]
[in a new window]
|
FIG. 23. Mass fraction profile of volatile organic compounds (VOCs) in the nonaqueous phase liquid after 5 yr of spill for Case C at –0.25 m above sea level.
|
|
|
Conclusions
|
|---|
The improved version of TMVOC numerical reservoir simulator can be used to model the compositional, multiphase flow of hydrocarbon and organic solvent mixtures spilled in coastal sites, where the seawater intrusion affects the groundwater flow. The spill in the unsaturated zone of a multicomponent mixture of six hydrocarbons is simulated using a simple two-dimensional vertical grid orthogonal to the coast line. The spill is simulated both accounting for the seawater intrusion and for comparative purposes, by assuming a freshwater boundary. The seawater intrusion concentrates the flow of dissolved VOCs in the upper section of the unconfined aquifer. This needs to be taken into account if extraction wells are drilled to control the transport of dissolved VOC toward the sea, a measure often requested by the controlling authorities. Wells located close to the coast should not be drilled down to the aquifer bottom to avoid the unnecessary pumping of seawater, which would increase the overall volume of extracted water and cause the progressive advancement of seawater intrusion inside the aquifer.
The effects of the construction of an impervious wall are also simulated and presented. The wall can intercept the NAPL lens avoiding direct discharge of free NAPL to the sea, while it allows the groundwater to flow beneath it. This could also permit an easier recovery of NAPL using specially equipped extraction wells. The simulations also show how the different properties of mixture hydrocarbons and related compositional effects control the VOCs retention in the subsurface and their migration along different pathways. This can be of help for risk assessment studies in sites contaminated by the spill of organic mixtures.
|
ACKNOWLEDGMENTS
|
|---|
TMVOC V.1.0 is based on the doctoral research of Adeyinka Adenekan. TMVOCBio was developed under the Workpackage 3 "Site characterization and modeling" of the PURE research project (Protection of groundwater resources at industrially contaminated sites) financed by the European Commission through the 5th Framework Program (contract EVK1-CT-1999-00030 PURE). Further improvements of TMVOC were performed within the Line D "Numerical model" of the ENI R&D project "Clean-up of groundwater and soils contaminated by chlorinated organic solvents" sponsored by the ENI Research Fund.
|
REFERENCES
|
|---|
- Adenekan, A.E. 1992. Numerical modeling of multiphase transport of multicomponent organic contaminants and heat in the subsurface. Ph.D. thesis, Univ. of California, Berkeley.
- Aquater. 2003. Line D: Numerical model. ENI R&D project "Clean-up of groundwater and soil contaminated by chlorinated organic solvents." (In Italian.) Rel_6004. Unpublished.
- Battistelli, A. 2004. Modeling biodegradation of organic contaminants under multiphase conditions with TMVOCBio. Vadose Zone J. 3:875–883.[Abstract/Free Full Text]
- Battistelli, A., C. Calore, and K. Pruess. 1997. The simulator TOUGH2/EWASG for modelling geothermal reservoirs with brines and a non-condensible gas. Geothermics 26:437–464.[CrossRef][Web of Science]
- Battistelli, A., C.M. Oldenburg, G. Moridis, and K. Pruess. 2003. Modeling gas Reservoir processes with TMVOC V.2.0. In Proceedings TOUGH2 Symposium, Berkeley, CA. 12–14 May 2003. LBNL-52494. Lawrence Berkeley National Lab., Berkeley, CA.
- Coats, K.H., and B.D. Smith. 1964. Dead-end pore volume and dispersion in porous media. Soc. Petrol. Eng. J. 4:73–84.
- Falta, R.W., K. Pruess, I. Javandel, and P.A. Witherspoon. 1989. Density-driven flow of gas in the unsaturated zone due to the evaporation of volatile organic compounds. Water Resour. Res. 25:2159–2169.[CrossRef]
- Garcia, J.E. 2001. Density of aqueous solutions of CO2. LBNL-49023. Lawrence Berkeley National Lab., Berkeley, CA.
- Harvey, C., and S.M. Gorelick. 2000. Rate-limited mass transfer or macrodispersion: Which dominates plume evolution at the Macrodispersion Experiment (MADE) Site? Water Resour. Res. 36:637–650.[CrossRef]
- Kipp, K.L., Jr. 1997. Guide to the Revised Heat and Solute Transport Simulator: HST3D, version 2. Water-Resources Investigations Report 97.4157. USGS, Denver, CO.
- Langevin, C.D., W.B. Shoemaker, and G. Weixing. 2003. MODFLOW-2000, the U.S. Geological Survey modular ground-water model: Documentation of the SEAWAT-2000 version with the variable-density flow process (VDF) and the integrated MT3DMS transport process (IMT). USGS Open-File Report 03-426. USGS, Tallahassee, FL.
- Millington, R.J., and J.P. Quirk. 1961. Permeability of porous solids. Trans. Faraday Soc. 57:1200–1207.[CrossRef]
- Nitao, J. 2000. NUFT flow and transport code V3.0, software configuration management. Yucca Mountain Project STN: 10088-3.0S-00, Lawrence Livermore National Lab., Berkeley, CA.
- Oldenburg, C.M., and K. Pruess. 1993. A two-dimensional dispersion module for the TOUGH2 simulator. LBL-32505. Lawrence Berkeley National Lab., Berkeley, CA.
- Oldenburg, C.M., and K. Pruess. 1995a. Dispersive transport dynamics in a strongly coupled groundwater–brine flow system. Water Resour. Res. 31:289–302.[CrossRef]
- Oldenburg, C.M., and K. Pruess. 1995b. EOS7R: Radionuclide transport for TOUGH2. LBL-34868. Lawrence Berkeley National Lab., Berkeley, CA.
- Parker, J.C., R.J. Lenhard, and T. Kuppusamy. 1987. A parametric model for constitutive properties governing multiphase flow in porous media. Water Resour. Res. 23:618–624.[CrossRef]
- Pruess, K. 1991. EOS7: An equation-of-state module for the TOUGH2 simulator for two-phase flow of saline water and air. LBL-31114, Lawrence Berkeley Lab., Berkeley, CA.
- Pruess, K., C. Oldenburg, and G. Moridis. 1999. TOUGH2 user's guide, version 2.0. LBNL-43134. Lawrence Berkeley National Lab., Berkeley, CA.
- Pruess, K., and A. Battistelli. 2002. TMVOC, a numerical simulator for three-phase non-isothermal flows of multicomponent hydrocarbon mixtures in saturated–unsaturated heterogeneous media. LBNL-49375. Lawrence Berkeley National Lab., Berkeley, CA.
- Pruess, K., and T.N. Narasimhan. 1985. A practical method for modeling fluid and heat flow in fractured porous media. Soc. Petrol. Eng. J. 25:14–26.
- Reid, R.C., J.M. Prausnitz, and B.E. Poling. 1987. The properties of gases and liquids. McGraw-Hill, New York.
- Reservoir Engineering Research Program Center for Petroleum and Geosystems Engineering. 2000. User's guide for UTCHEM-9.0: A three-dimensional chemical flood simulator. Vol. 1. Reservoir Engineering Research Program Center for Petroleum and Geosystems Engineering, Univ. of Texas, Austin.
- Rusi, S., F. Tatangelo, A. Battistelli, and E. Crestaz. 2005. Density-dependent groundwater modelling for coastal industrial sites. p. 20–25. In Proc. ModelCare 2005: 5th Intl. Conf. on Calibration and Reliability in Groundwater Modelling: From Uncertainty to Decision Making, The Hague. 6–9 June 2005. IAHS Publ. 304. IAHS Press, Wallingford, UK.
- Sanese, L., A. Battistelli, and G. Gottardi. 2003. Numerical simulation of organic contaminants migration in coastal aquifers with T2VOC-BRINE. (In Italian, with English abstract.). IGEA 18:65–80.
- Stone, H.L. 1970. Probability model for estimating 3-phase relative permeability. Trans. Soc. Petrol. Eng. AIME 249:214–218.
- Voss, C.I., and A.M. Provost. 2002. A model for saturated–unsaturated, variable-density ground-water flow with solute or energy transport. Water-Resources Investigations Report 02-4231. USGS, Reston, VA.
- White, M.D., and M. Oostrom. 2000. STOMP Subsurface Transport Over Multiple Phases: Version 2.0, theory guide. PNNL-12030. Pacific Northwest National Lab., Richland, WA.
- Wu, Y.S., and K. Pruess. 2000. Numerical simulation of non-isothermal multiphase tracer transport in heterogeneous fractured porous media. Adv. Water Resour. 23:699–723.[CrossRef]
This article has been cited by other articles:
|
|
|
|
 
H.-H. Liu and T. H. Illangasekare
Preface: Recent Advances in Modeling Multiphase Flow and Transport with the TOUGH Family of Codes
Vadose Zone J.,
February 25, 2008;
7(1):
284 - 286.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
