United States
Environmental Protection
Agency
National Risk Management
Research Laboratory
Ada, OK 74820
Research and Development
EPA/600/SR-97/099
September 1997
A
to Soil
and of
in
Linda M. Abriola, John Lang, and Klaus Rathfelder
Soil vapor extraction (SVE) and
bioventing (BV) are prowen
for of
soils. Mathematical are powerful
can be to
quantify the interaction of physical,
chemical, and biological
occurring in scaleSVE/BVsystems.
This the
formulation, numerical
implementation, and application of a
multiphase, multicomponent, bio-
tosimulate
physical, chemical, biological
occurring in
and BV systems. The model, the
Michigan Soil Vapor Extraction
is two-
be run in a
or
are discretized with a standard Galerkin
finite element approach using linear
triangular A modular,
algorithm is
of the include: the ability
to simulate multiphase gas and aqueous
fluid flow; the simulation of multi-
component transport processes,
Incorporation of rate-limited
transfer for volatilization and
dissolution of an entrapped organic
liquid, volatilization and of
aqueous constituents,
biophase update; and the simulation of
muttlcomponent biodegradation kinetics
microbial population dynamics. A
complete description of the computer
code, implementation and
SVE BV simulations Is
included.
This
by 's
Research Laboratory's Subsurface
Ada, OK, to of
project that is fully
In a
title
at back),
Introduction
Soil vapor extraction (SVE) and
bioventing (BV) are proven and widely used
remediation strategies which target the
removal of volatile organic compounds
(VOCs) from the unsaturated zone. SVE
and BV are similar in that they both employ
vadose zone wells and pumps to generate
gas flow through the unsaturated zone.
They differ fundamentally, however, in the
mechanism of contaminant removal. SVE
systems emphasize removal by
contaminant volatilization and above ground
recovery. BV systems are designed to
maximize contaminant removal by in situ
biodegradation, which is stimulated by
oxygen enhancement in the air stream.
The efficiency and degree of success of
SVE/BV technologies is controlled by a
complex combination of physical, chemical,
and biological factors. SVE systems
-------
characteristically exhibit diminishing
removal efficiency and long term tailing
behavior indicative of diffusion controlled
mass transfer. Contributing factors may
include flow bypassing in heterogeneous
soils, rate limited volatilization and
desorption, and contaminant occlusion due
to water table upwelling. Biodegradation
rates in BV systems can potentially be
affected by a number of factors including:
the soil moisture content and distribution,
delivery of oxygen, availability of inorganic
nutrients, and substrate concentration,
which can be inhibitory at high levels or
limiting under conditions of diffusion
controlled desorption.
Because of the complexity of the
processes influencing the performance of
SVE/BVtechnologies, design and operation
guidelines are frequently qualitative in
nature, based on experience or simple
design rules. Mathematical models are
recognized as powerful tools that can be
used to integrate and quantify the interaction
of physical, chemical, and biological
processes occurring in field scale SVE/BV
systems. In addition to predicting potential
mass removal, mathematical models can
be used to explore alternative system
designs and to investigate factors limiting
successful remediation. Existing simulation
models of SVE/BV systems vary in their
focus and processes considered. Each
may exhibit limitations in its applicability to
field problems by focusing only on fluid flow
or single phase fluid flow conditions, by
neglecting biotransformation, or by failing
to fully account for nonequilibrium mass
transfer processes between all fluid phases,
the soil matrix, and the biophase. This
report describes the mathematical and
numerical formulation of a two-dimensional,
field scale SVE/BV simulator that
incorporates processes of multiphase fluid
flow, compositional constituent transport,
nonequilibrium interphase mass transfer,
and Monod kinetics for aerobic
biodegradation. The simulator is entitled
the Michigan Soil Vapor Extraction
Remediation Model, or MISER.
Conceptual Framework and
Formulation
Threefluid phases are modeled in MISER:
a mobile gas phase, a mobile aqueous
phase, and an immobile residual organic
liquid phase. The gas and aqueous phases
can flow simultaneously in response to
pressure and density gradients arising from
fluid extraction or injection at wells, rainfall
infiltration, orapplied surface irrigation. The
movement of the aqueous and gas phases
is described by standard macroscopically
averaged flow equations. Hysteretic
behavior in the retention and relative
permeability relations, including those
arising from air entrapment, are neglected.
Changes in the assumed immobile organic
liquid saturation occur solely by interphase
mass transfer. Organic liquid saturation is
updated from the organic phase mass
balance equation.
A compositional modeling approach is
employed. The organic liquid is considered
to be a mixture of a variable number of
organic components. The gas phase is
modeled with a composition of nitrogen
(the major constituent of air), oxygen (the
single elector acceptor), water vapor,
volatile constituents of the organic liquid,
and a single limiting nutrient (e.g.,
ammonia). The aqueous phase is assumed
to be comprised of water, oxygen, soluble
constituents of the organic liquid, and the
limiting nutrient. To account for possible
rate limited uptake by the microbes, the
biophase is envisioned as a subset of the
aqueous phase. Sorption to the soil particles
is restricted to components of the organic
liquid.
The transport and transformation of
individual phase components is described
by a general macroscopically averaged
transport equation. Rate limited interphase
masstransferis modeled with a lineardriving
force expression. These expressions are
used to model volatilization and dissolution
of the entrapped organic liquid, mass
exchange between the aqueous and gas
phases, rate limited sorption, and rate limited
transport to the biophase. Complete drying
of the aqueous phase is not considered
and, consequently, solid phase sorption
from the vapor phase is neglected.
Nonlinear sorption to the solid phase is
modeled with a Freundlich isotherm.
Biodegradation is modeled solely as an
aerobic process. Biodegradation is
assumed to occur only within the aqueous
phase by an indigenous, spatially
heterogeneous, mixed microbial population
that is present as attached microcolonies.
It is further assumed that biomass growth
does not affect soil permeability, that there
is no biomass transport, and that
detachment or sloughing of the attached
biofilm is negligible. Monod-type kinetic
expressions are employed to describe
biophase utilization of substrates, electron
acceptor, and a single limiting nutrient.
MISER has the capability to simulate
biophase uptake of substrate, oxygen, and
nutrient as an equilibrium or rate limited
process. MISER incorporates inhibitory
effects on biokinetics resulting from the
presence of excessively high or low
substrate concentrations, or due to the
presence of oxygen below a minimum
threshold limit. Growth and decay of the
attached microbial population is modeled
with a Monod-type kinetic expression,
extended to constrain biomass
concentration between a minimum value
reflecting the indigenous population present
in uncontaminated subsurface environ-
ments, and a maximum permitted biomass
concentration.
Numerical Development
The flow and transport equations are
discretized in two space dimensions using
a standard Galerkin finite element approach
withlineartriangularelements. Thecoupled
nonlinear equations are solved using a
modular, set-iterative solution algorithm.
In this approach, the sets of flow, transport,
and biodegradation equations are
decoupled and solved separately. The set-
iterative approach substantially reduces the
size of solution matrices and enhances
flexibility. With this scheme, different
numerical approaches, as well as alternative
grid and time discretization schemes, can
potentially be applied for each equation
set. The modular approach also facilitates
application of the model to simplified
scenarios, such as steady flow situations or
SVE in the absence of biotransformations.
The flow equations are solved with a
pressure based simultaneous solution
scheme, using a Picard iteration scheme to
account for nonlinearities. Subsequently,
the entire set of constituent transport
equations for the mobile aqueous and gas
phases, and immobile NAPL residual and
solid phases are solved sequentially using
Picard iteration schemes. Mass exchange
terms and molecular weights are updated
at each iteration. The entire process is then
repeated until convergence for all
constituent concentrations in all phases is
obtained. NAPL saturations and
biotransformations are updated at the end
of the time step.
Verification
Extensive verification of numerical
solutions was conducted by material
balance computations, comparison with
analytical solutionsforsimplified scenarios,
intermodel comparisons, and comparison
with published column data.
Numerical solutions of aqueous flow were
verified with analytical expressions for the
one-dimensional Richards equation, and
by intermodel comparison for conditions of
two-dimensional axisymmetric moisture
infiltration into a two layer system.
Predictions of unsteady radial flow of gas
were verified against one-dimensional
quasi-analytical solutions which take into
account nonlinearities stemming from
compressibility and the Klinkenberg effect.
Numerical solutions of contaminant
transport were verified against one- and
two-dimensional analytical solutions for
-------
advective-dispersive transport of a single
constituent subject to first order decay.
Verification of biokinetics was conducted
by 1) comparison with the two-dimensional
analytical solutions by adjusting Monod
parameters to represent first order
biodegradation rates; and 2) intermodel
comparison for conditions of microbial
growth and degradation in a one-
dimensional soil column.
Program Description
Example Simulations
MISER is developed in the FORTRAN 77
programming language and is comprised
of 25 program modules. The modularformat
enhances clarity, facilitates logic tracing
and code modification, and enables the
code to be easily programmed to run in
different modes. Due to the large number
of equations being solved and the
complexity of the solution algorithm, the
code is intended to be used on work station
platforms or main frame computers. A
description of major code variables and a
complete source code listing is provided.
Input data are organized into two main
input files, one required error message file,
and an optional restart file used for run
continuations. Considerable parametric
inputs are required to describe the physical,
chemical and biological characteristics, as
well as the numerical control features. A
complete description of all input data and
the organization of input data blocks is
provided in the report.
Output formats and files are flexible.
Optional output files can be generated for
runtime information, material balance
computations, contour plot data files, time
series information, and restart information.
Example Simulations
Three example simulations are presented
to illustrate application of the MISER model.
Two-dimensional axisymmetric simulations
are presented for hypothetical SVE and BV
scenarios. The physical system is an
unsaturated medium sand, intersected with
a layer of slightly less permeable fine sand.
Entrapped toluene liquid is present in each
layer. An extraction/injection well is
positioned in the center of the initial NAPL
distribution, and an impervious surface cap
is present on the ground surface. A listing
of all input data files is included as an
appendix.
Remediation of the contaminated soil by
SVE was simulated by the numerical
application of a constant extraction rate of
100 standard cubic feet per minute (scfm).
Mass transfer coefficients were selected to
represent relatively rapid organic liquid
volatilization and slower rates for desorption
and aqueous/gas partitioning. A temporal
progression of predicted organic liquid
saturation profiles illustrates the effects of
organic liquid persistence in low
permeability zones due to flow bypassing.
Complete removal of the organic liquid
occurs relatively quickly. However, due to
the magnitude of the selected mass transfer
coefficients, there is long term persistence
of contaminant mass on the solid phase
and in the pore water, producing a persistent
low level of contaminant discharge at the
well. In this simulation, the long term SVE
efficiency is controlled by the aqueous-
solid desorption rate after the period of
organic liquid removal.
Remediation of the contaminated soil by
BV was simulated by the numerical
application of a constant injection rate of
one cubic feet per minute (scfm) to supply
oxygen and enhance biotransformation. Air
injection produces outward radial movement
of oxygen (electron acceptor), as well as
toluene (substrate) due to volatilization from
the organic liquid. Organic liquid removal
also progresses radially outward. Similar
to the SVE scenario there is substantial
persistence of organic liquid within the low
permeability zone due to flow bypassing.
An examination ofthetemporal progression
of the predicted growth of biomass overthe
course of BV simulation reveals that, due to
substrate inhibition, biomass growth is
concentrated along the outside edge of the
organic liquid core, developing a so-called
"biofence." Once the "biofence" has
developed, the predicted toluene removal
from the gas phase occurs over a relatively
short distance, and there is little or no
growth radially outside of the "biofence"
due to an absence of substrate. Biomass
growth is also observed to fill in regions
close to the well after organic liquid has
been removed and aqueous concentrations
fall below the inhibitory threshold.
-------
Linda M. Abriola, John Lang, and Klaus Rathfelder are with Department of Civil and
Environmental Engineering, The University of Michigan, Ann Arbor, Ml 48109
Jong Soo Cho is the EPA Project Officer (see below).
The complete report, entitled "Michigan Soil Vapor Extraction Remediation (MISER)
Model: A Computer Program to Model Soil Vapor Extraction and Bioventing of Organic
Chemicals in Unsaturated Geological Material, " ( Order No. PB98-115355; Cost:
$47.00, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
U. S. Environmental Protection Agency
National Risk Management Research Laboratory
Subsurface Protection and Remediation Division
P.O. Box 1198
Ada, OK 74820
United States
Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, OH 45268
BULK RATE
POSTAGE & PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use $300
EPA/600/SR-97/099
------- |