Soil Vapor Extraction

1
Soil Vapor Extraction
2
Contamination in the Vadose Zone

• The vadose (unsaturated) zone acts as a buffer
  zone for protecting the quality of the
  underlying ground water.
• When contaminated, however, it acts as a source
  zone for ground water pollutants and gaseous
  emissions.

3
Soil Vapor Extraction (SVE)

• Targets the removals of VOCs from the vadose zone
  by volatilization
• Shown to be effective at removing NAPL, aqueous,
  and sorbed phases
• Encourages aerobic biodegradation
• Proven technology with some design guidance
  (rule-of-thumb).

4
Case Study JP-4 Spill at Hill AFB, Utah

• A JP-4 fuel spill site at Hill, AFB , Utah was
  selected to be modeled using VENT3D.
• This is probably the most comprehensive data set
  collected from a field application of SVE.
• Several pilot studies, and a full scale soil
  vapor extraction (SVE) operation under a variety
  of flow conditions were conducted at the site.
• Data for gas concentrations, contaminant
  concentrations, mass removals and vacuum pre- and
  during-operation were provided by the
  researchers.

5
VENT3D Description

• VENT3D solves the 3-D vapor phase
  advection-dispersion equation for a mixture of
  compounds. It computes the 4-phase distribution
  for each compound between vapor moving periods,
  assuming equilibrium partitioning between phases.
• The model domain is discretized into blocks.

6
VENT3D Flow Algorithm

• The model solves the 3-D steady-state gas flow
  equation using finite-difference.
• where k soil vapor permeability tensor (L2), P
  soil-gas pressure (M/LT2), ? soil-gas
  viscosity (M/LT),W vapor mass flux source/sink
  (M/T), R universal gas constant (ML2/T2 mole
  oK), T temperature (oK), MW molecular weight
  of soil-gas (M/mole), k intrinsic permeability
  (L2),
• ? air filled porosity (dimensionless), v
  interblock gas flow (L/T).

7
VENT3D Transport Algorithm

• Knowing the 3-dimensional interblock flows, the
  3-D advection-dispersion equation is solved by
  finite-difference for each chemical compound.
• where Mn total molar concentration of compound
  n in mixture (mole/L3), Cn molar concentration
  of compound n (mole/L3), q vapor discharge
  vector (L/T), Fnvolumetric molar loss/addition
  rate of compound n (mole/L3T), Dn the
  dispersion tensor (L2/T) , Dm molecular
  diffusion coefficient (L2/T), Do free air
  diffusion coefficient (L2/T), I identity
  vector, vapor dispersivity values (L),
  qx,qy vapor flow in x,y,z, direction (L/T),
  magnitude of the discharge vector (L/T).

8
Phase Partitioning

• Equilibrium partitioning is assumed between the
  different phases for each compound. Equilibrium
  is re-calculated at the end of each time step.
  The total molar concentration of each compound is
  expressed as a function of the vapor
  concentration and the sum of the molar
  concentrations in the 4 phases
• where MHCmolar concentration of NAPL phase in
  soil (mole/L3), Pv compound vapor pressure
  (atm), MH2O molar concentration of dissolved
  phase (mole/L3),???nactivity coefficient of
  compound in water (dimensionless), Kdn
  distribution coefficient of compound
  (dimensionless), ? soil density (M/L3), MWH2O
  molecular weight of water (18 gm/mole), ?H2O
  soil moisture flag1 if present, 0 if not present.

9
Site History

• 27,000 gallons (76,500 kg) of JP-4 jet fuel
  spilled in January 1985.
• Contaminated area west of spill area 160 ft
  120 ft 50 ft.
• A field study of soil venting was performed at
  the site in 1988-1989.

10
Pilot Full-Scale Tests

11
VENT3D Vertical Layers

• The contaminated soil volume is divided into 11
  layers, each having different initial contaminant
  concentrations in soil.

12
VENT3D Horizontal Grid

• Each layer is divided into a 16 12 orthogonal
  grid, each grid cell is 10 ft wide 10 ft. long.
  The grid shows the subareas to which the vertical
  vent area was divided. Each subarea is presented
  by a soil boring V1, V2........V15.

13
JP-4 Composition

• Reported and estimated standard weight fractions
  for JP-4 components.

14
Determining Site Permeability

• The horizontal and vertical permeabilities at the
  site were estimated using GASSOLVE, a computer
  program developed by Falta (1996).
• GASSOLVE uses analytical solutions to the
  steady-state gas flow equation for different
  boundary and initial conditions.
• The permeability is found by fitting the
  analytical solution to pressure data collected
  from air permeability tests.
• GASSOLVE adjusts the permeability until it
  reaches a minimum residual sum of squares between
  the calculated and observed pressures.

15
GASSOLVE Results

• Pressure data from four pilot tests were used by
  GASSOLVE to determine the air permeability of the
  formation.
• Results indicated
• horizontal permeability 40 darcys.
• vertical permeability 1 darcy.

16
Flow Calibration Validation

• Five pilot tests were simulated by VENT3D using
  GASSOLVE estimated permeabilities.
• Three of the full-scale operation flow tests were
  simulated by VENT3D.

17
Vacuum Adjustment

• VENT3D calculated pressure was adjusted at
  extraction wells using the equation adapted from
  Anderson and Woessner, 1992.
• where Pw is the well pressure, P i,j is the
  calculated pressure, Qw is the pumping/injection
  rate, k is the permeability, ? is the fluid
  viscosity, ? is the fluid density, g is the
  gravitational acceleration, re is the effective
  well block radius and rw is the well radius.

18
Vacuum Adjustment Results

19
Transport Calibration

• The total initial contaminant mass was adjusted
  until the measured mass removals from the pilot
  tests and calculated mass removals from VENT3D
  matched within less than 20 error.
• The initial total mass was estimated to be as
  high as 76,000 kg based on the spill volume.
• Five pilot tests were simulated and mass removals
  were compared.

20
Transport Calibration Using Pilot Tests

• Data from pilot tests were not successful in the
  calibration step due to some discrepancy in
  reported data.

21
Transport Calibration Using Full Scale Operation

• Mass removals recorded at the beginning of the
  full-scale operation were used to verify the
  calibration process.

22
Transport Validation

• Mass removals from simulations representing the
  whole operation were compared with the measured
  values.

23
Initial Final Soil and Gas Concentrations

24
Average Extraction Rates

• To evaluate whether the detail of the flow
  history was necessary,the
• 25 flow tests conducted during the full-scale
  study were represented
• by one simulation.

25
Average Extraction Rates

• VENT3D estimated a 2 difference in removals.

26
Predicting System Behavior

• VENT3D was used to predict the extent of
  contamination if SVE was not carried out at the
  site.

27
Predicting System Behavior

28
Conclusions

• The design and performance of remediation systems
  can be greatly improved through the use of
  mathematical models.
• VENT3D proved successful in representing site
  characteristics with respect to subsurface air
  flow and for simulating the performance of a
  vapor extraction study conducted at a JP-4 jet
  fuel spill site at Hill AFB.
• Air permeability was appropriately estimated
  using GASSOLVE.
• VENT3D helped in determining the initial total
  contaminant mass at the site.

29
Conclusions (cont.)

• 3-D modeling provided a closer match to field
  measurements than 2-D modeling.
• The loss of accuracy in 2-D modeling was small
  from a design standpoint and was accompanied by
  considerable savings in computer time.
• JP-4 could be represented with a mass-equivalent
  10-compound mixture, and even a single component
  representing the mixture.
• The gain in accuracy provided from modeling the
  multi-component mixture also came at the cost of
  extra computational effort.

30
Conclusions (cont.)

• The slight improvement in accuracy by using 3-D
  modeling and detailed multi-component
  representation of the jet fuel does not justify
  the increased computational effort.
• We therefore propose that for similar
  applications, one can represent the mixture by a
  smaller number of compounds and use a
  two-dimensional model without considerable loss
  of accuracy.
• VENT3D was useful in demonstrating the changes in
  JP-4 composition during SVE .

31
Conclusions (cont.)

• In this case, a surface seal was predicted to
  have minimal effect on cumulative mass removals.
• If a model was to be used for design purposes, it
  would be more convenient to be able to use
  average (constant) flow conditions instead of
  going through a lengthy, complicated process of
  running a large number of simulations.
• Data from long-term studies give a better
  description of site conditions and system
  behavior than data from pilot tests.

32
Limitations

• VENT3D calculated soil gas concentrations did not
  match well with observed concentrations due to
  the following
• mass distribution from soil cores was not
  accurate.
• VENT3D does not allow for specifying different
  mass fractions at different locations.
• VENT3D does not account for mass transfer
  limitations.
• The site domain was modeled as a homogeneous
  formation.