The International Information Center for Geotechnical Engineers

Soil Vapor Extrusion

Design and Operation of SVE Systems

The design of soil vapor extraction (SVE) systems is largely empirical due to the great number of parameters and mechanisms involved. Accordingly, the commissioning of each SVE system should be preceded by an involved screening process, which transitions into a pilot-scale test at the site of interest. The following sections provide a summary of accepted methods for determining if SVE is an appropriate remediation technology for a given site and key components of pilot-scale testing. This is followed by a brief overview of the design of the full SVE system as well as operation and monitoring considerations. Finally, some of the potential modifications of SVE are discussed.

Determination of Applicability of SVE

The first step in designing a SVE system is determining if in fact it is an appropriate technology and is capable of removing the desired amount of contaminant from the site. Following the initial investigation of the site and determination of the extent and composition of contamination, the decision tree presented in Figure 3 can be used as an initial indicator of whether or not SVE may be applicable. The second to last box of the flow chart, labeled “Evaluate Relative to a Variety of Site Specific Factors, Considering Experience at Other Sites” entails an in-depth screening process to further ensure that SVE is viable, and is discussed below.

Figure 3. Technology Screening Decision Tree (USACE, 2002)

(SVE: Soil Vapor Extraction

BV: Bioventing)

Johnson et al (1988) present a comprehensive, and widely referenced, set of mathematical models that can be used as screening tools for SVE. The authors identify three primary factors that strongly influence the applicability of SVE and provide equations to model them. The three factors are: 1) vapor flow rate, 2) contaminant concentration, and 3) the location/setting of the contamination relative to the desired flow path. For steady state conditions, the radial pressure distribution around the well, the radial velocity distribution, and the vapor flow rate can be described by equations 1, 2, and 3, respectively. Johnson et al also present the transient forms of these equations as well as the derivations.

(1)

Where Pw: pressure at the vacuum well

RI: radius of influence

Patm: atmospheric pressure

r: radius from the well

(2)

Where k: soil permeability

μ: vapor viscosity

(3)

Where: H: screened length of the well (within the vadose zone)

Using these equations, for a homogenous soil system, the vapor flow rate, velocity distribution around the well, and vapor travel time from a given point to the well, can be estimated. For SVE systems through layered soils, the total flow rate is simply equal to the sum of the individual flow rates (according to equation 3) for each layer. It is also critical to understand how the concentration of the contaminant will change with time during the vapor extraction process. Johnson et al (1988) present a solution to the governing vapor phase concentration equation which can be incorporated into the mole balance equation (equation 4). This must be solved numerically to obtain the change in concentration with time. Figures 4 and 5 present examples of results from this type of analysis. They are the mass loss rates with time and concentrations with time, respectively.

(4)

Where Mi: total number of moles of component i

Ci: molar concentration of i

Bi: rate of degradation (biological or chemical) of i

t: time

Figure 4. Predicted Mass Loss Rates for a Hypothetical Venting Operation (Johnson et al 1988)

Figure 5. Predicted Soil Concentrations of Hydrocarbons for a Hypothetical Venting Operation (Johnson et al 1988)

Finally, it is important to consider the flow path, relative to the contamination, that will be induced by the proposed SVE system. The mathematical models presented above for factors (1) and (2) are based upon assumptions of vapor flow through homogenous soils with homogenous contamination. This however is often not the case. The contaminant may be spread through various soil layers and/or may exist below the ground water table or in a relatively impermeable layer. Under these conditions, the vapor extraction must take place through diffusion, instead of simply advection through unsaturated soils. Figure 6 illustrates some of the different scenarios under which SVE may take place. This must be understood for the site of interest and the boundary conditions of the mathematical models used to predict vapor flow and concentrations must be adjusted accordingly.

Figure 6. Scenarios of SVE flow paths (Johnson, 1990)

Following determination of the three factors presented above, an informed decision can be made to determine if SVE is appropriate for the given site. The models created can be used to estimate how an SVE system would perform. Johnson et al (1990) present a series of six basic questions that need to be answered alongside these factors to verify that SVE is feasible. They are as follows:

1. What are the expected contaminant vapor concentrations (at the beginning of extraction)?
2. Is the concentration great enough to achieve acceptable contaminant removal rates?
3. What vapor flow rates can be achieved?
4. Will the concentration and vapor flow rates produce acceptable removal?
5. What residual will be left in the soil following remediation?
6. Is there potential for any negative impact to the site or surrounding sites due to the SVE process?

Of these questions, Johnson et al identify 2) and 4) as being critical to the advancement of the project beyond the screening phase. If the concentration of contaminant within the soil is too low for the SVE process to remove a significant portion of the contaminant, a different remediation option must be considered. When examining this, the biodegrability of the contaminant must be considered, since it may result in changes to the achievable removal concentrations as the contaminant degrades. Also, with the available equipment and supplies, it may not be possible to achieve high enough flow rates to result in adequate contaminant removal. This also is grounds for abandoning SVE as a remediation option. There is also potential for the SVE operation to negatively affect the site or neighboring sites. One example of this is contaminants from neighboring sites being drawn in to the site of interest by the vapor flow or as a result in change in ground water table in conjunction with SVE. These types of issues however can be remedied with passive wells and by diligent monitoring during the operation of the remediation. A passive well is a well which has no pump but enhances airflow from the ground surface, allowing the designer to better control the subsurface airflow.

Pilot-Scale Testing

Once it has been determined that, according to the predictive models, SVE is an appropriate remediation technology, it is necessary to perform pilot-scale testing at the site. Due to the complexity of the process, predictive models may not capture all of the important aspects of the system. For this reason, pilot-scale tests are always conducted prior to the full design and implementation of a SVE system. In addition, bench-scale column tests, air permeability tests and ground water pumping tests are often conducted.

Pilot-scale tests typically consist of a single extraction well with additional wells for sampling and monitoring. Figure 7 presents a schematic of a typical pilot test setup. USACE (2002) identifies the following measurements that should be taken during a pilot-scale test: airflow rates, pressure levels, soil/air temperature, soil moisture levels, effluent contaminant concentrations, and quantity and composition of liquids retrieved from the air/water separator. The results of the pilot scale test are used to refine the site model that was developed during the screening process. Of particular interest from a pilot test is the time that it takes for the system to reach steady-state removal concentrations. Since this is dependent on a wide range of factors, it is very difficult to accurately model. A pilot test is run at least until the baseline concentration is reached. Typical pilot tests are completed within hours, but may take days, or even longer. Performing Modified Active Gas Sampling in conjunction with the pilot-scale testing can reduce the time needed for the test. In addition to helping refine the predictive models, the pilot test may also be used to refine the understanding of where the contaminant is located, since the initial site investigation can only provide estimates. This information can be gathered from the removal concentration-time history taken from the pilot test. If the concentration increases with time, it indicates that the pilot extraction well is located at a distance from the contaminant center, while a decrease in concentration with time indicates that the well is near the center of the contaminant. Finally, the pilot test can be used to monitor potential upwelling of the ground water table due to the vapor extraction process.

Figure 7. Typical pilot test setup schematic (USACE, 2002)

System Design

Once the pilot test has been conducted the complete SVE system can be designed. Using the predictive models from the screening process and the information gathered from the pilot test, the various aspects of the system, including well depth, screening depth, number of wells and spacing, well arrangement, pump/blower capacity, and above ground treatment, can be arranged to optimize the removal of contaminants from the soil. This is a complicated process which is largely empirical and requires vast experience and expertise. In order to illustrate this complexity, the list of professionals that USACE (2002) recommends be a part of the design team is reproduced below:

• Environmental/chemical engineer
• Health and safety specialist
• Mechanical engineer
• Regulatory specialist
• Chemist
• Cost engineer
• Geologist/geotechnical engineer/hydrogeologist
• Civil/structural engineer
• Soil scientist/soil physicist
• Electrical engineer

A critical component of the design of the SVE system is the number of extraction wells. There are multiple methods for calculating the number of wells, all with varying assumptions. Johnson et al (1990) suggest computing the number of wells using three methods, described here, and selecting the largest of the three. The first method is to simply divide the acceptable removal rate by the estimated removal rate for one extraction well. This neglects the change vapor concentration with time. The second method is to use the mathematical models presented in the screening process to calculate the required number of wells. Depending on the extent to which the models have been developed, and what assumptions were made, this approach may allow for a better representation of the transient nature of the process. The third approach is to divide the area of the contamination by the area of influence of a single extraction well.

Once the number of wells has been determined, their configuration and spacing must be designed. This will be heavily dependent upon how the desired vapor path will travel relative to the contaminant, as well as soil stratigraphy. Well spacings typically range between 15 and 100 feet (Hutzler et al, 1991). There are multiple types of configurations that may be used, including vertical wells and horizontal trenches, capped systems, systems which cycle between wells at a given site, and single well systems. Figure 8 illustrates some of the variations in possible well configurations. It should be noted that while designing the configuration, care must be taken to avoid stagnant zones between wells.

Figure 8. Examples of venting well configurations (USACE, 2002)

Upon extraction from the subsurface, the contaminant vapors must be treated prior to release. Several methods exist for the above ground treatment of offgas. Table 4 presents a table of VOC treatment technologies, reproduced from USACE (2002). Of these technologies, Johnson (1990) indicates that the ones most commonly used for SVE systems are vapor combustion, catalytic oxidation, carbon beds, and diffuser stacks. As indicated in the USACE table, each option has its advantages and disadvantages. Diffuser stacks are not listed in the table, because they do not actually treat the vapors, but rather release them in a controlled manner. This option is often not permitted and the health and safety risks associated with it must be carefully considered.

Table 4. VOC Treatment Technologies (USACE, 2002)

The design considerations presented here are only a brief overview of the design process that takes place for SVE systems. In addition to the relatively simple mathematical models that have been presented, several sophisticated tools are available to aid in the design process. The USEPA has made available a software program, HyperVentilate, for guiding engineers through the design process. This software is based on the approach presented by Johnson et al. Three dimensional finite difference programs, such as VENT3D, have also been used for modeling SVE applications (e.g. El-Beshry et al, 2000). However, many publications on the topic of SVE design stress the fact that experience and good engineering judgment are paramount in designing a successful SVE system.

System Operation and Monitoring

Following completion of the design, the SVE system is ready to be constructed and extraction can commence. (Figure 9 presents an image of an operating SVE system) However, the operation is still a dynamic process and parameters are still subject to change. Extensive monitoring must be conducted throughout the operation of the system for multiple reasons. The operators must verify that the equipment is working properly and must adjust the operation (well pressures, etc.) to further optimize the removal of the contaminants. Finally, the monitoring results are used to determine when the remediation is complete and the system can be shut down. Table 5 presents a list of items that USACE suggests be monitored throughout the operation. Of these measurements, Johnson et al (1990) suggest that, at the very minimum, vapor flow rates, pressure readings, vapor concentrations and compositions, temperature, and water table levels, must be taken.

Figure 9. Image of SVE System During Operation (serc.carleton.edu)

Table 5. Suggested SVE System Monitoring (USACE, 2002)

The duration and cost of an SVE operation vary widely from site to site, depending on numerous factors. The operation of the systems is typically on the order of years. A summary of 12 remediation cases presented by the Federal Remediation Technologies Roundtable (USEPA,1998) indicated project durations ranging from 8 months to 8 years. A table from the same report summarizing the cost of the 12 projects is presented in Table 6. As the table shows, the total cost of the projects ranged from $76,000 to$43 million.  The cost per unit of treated soil also varied widely, with the lowest being $0.35/lb and the highest being$220/lb.

Part of the reason for the wide range in operation duration and costs is the fact that there are several modifications and extensions of conventional SVE that may be used. SVE is very frequently used in conjunction with some other remediation technology such as air sparging or bioventing (Hutzler et al, 1991). Air sparging is a similar method to SVE, but air is injected below the ground water table to force volatile contaminates to rise to the top of the water surface, where they can then be extracted by SVE.  SVE may be performed in above-ground soil piles in cases where in-situ remediation is not feasible. The contaminated soil is excavated and stockpiled on the surface with a series of horizontal extraction trenches, as opposed to the conventional vertical wells. Also, despite the fact that the information presented in the screening process suggested that SVE may only be used for formations with high permeability, the USEPA has conducted SVE at low permeability sites by using hydraulic fracturing enhancement, in order to allow vapor pathways (Frank and Barkley, 1995). Overall, SVE is a very versatile and effective remediation option.

Table 6. Summary of Cost Data for Several SVE Projects (USEPA, 1998)