A description of the design, construction, and performance of a soil vapor extraction system in response to the discovery of the subsurface migration of methane-bearing LFG away from a California landfill.

By Stephen Ferry

The generation of methane during the decomposition of waste buried in landfills is a well-understood phenomenon for which no practical means of prevention exists.

The lateral migration through the subsurface of landfill gas (LFG) containing as much as 50% methane by volume has been documented at numerous landfill sites, including the subject site. Often the preferential pathways of migration are anthropogenic (such as gravel-filled utility easements) and tend to conduct the gas directly to areas of human activity (such as basement and ground-floor rooms of nearby buildings). Because methane is combustible at concentrations of more than 5% by volume, the uncontrolled subsurface release of LFG has been recognized as an immediate danger to life and limb in areas surrounding landfills. Local enforcement agencies have formally defined the conditions under which control measures must be taken and the standards of performance that constitute regulatory compliance. The regulations affect virtually every landfill operating today.

This article describes the design, construction, and performance of a soil vapor extraction system in response to the discovery of the subsurface migration of methane-bearing LFG away from a landfill in California. A barrier to gas migration approximately 6,000 ft. long, consisting of 28 vertical gas extraction wells, was constructed in two phases. The first phase was an eight-well pilot system that was operated for approximately one year to verify the theory that a unique design for the gas control system would be appropriate for this site. The second phase was a 20-well system incorporating the design guidelines developed during the pilot phase.

Site Description and History

The surface topography in the vicinity of the landfill is flat, and the native ground surface elevation is approximately 50 ft. above mean sea level (msl). The uppermost geologic formation in the area, within which all project activities were conducted, is composed of interbedded fluvial sands, silts, and clays with inclusions of locally extensive gravel lenses.

In the vicinity of the project, the uppermost lithologic layer is a deposit of silty clay approximately 25 ft. thick. Beneath the silty clay lay an extensive 25-ft.-thick layer of unconsolidated, poorly sorted stream-rounded cobbles and minor amounts of coarse sand. Beneath the cobble layer lay 70-80 ft. of interbedded medium- to fine-grain sands, silts, and clays. This highly permeable cobble layer, sandwiched between two lower-permeability zones, formed an unusually extensive fluvial bedload deposit. The extent of the cobble layer beneath the site was definitively confirmed prior to the opening of the landfill by gravel mining activities. The entire 100-ac. footprint of the present landfill was mined by first removing and stockpiling the upper 25 ft. of silty clay, then removing and selling the underlying 25 ft. of cobbles. The gravel mining operation at its conclusion exposed a 25-ft.-high vertical wall of cobbles surrounding the entire site, in the interval between approximately 25 ft. and 50 ft. msl.

After gravel mining ceased, the stockpiled overburden was replaced in the pit, filling it to approximately 28 ft. msl. Figure 1 shows the landfill site in cross-section as it existed immediately before the present landfilling operation began. Prior to the deposit of the first buried waste, the old gravel mining pit was reexcavated. All of the redeposited overburden and an additional 15 vertical ft. of the underlying dense, sandy silt were removed, establishing the bottom of the landfill at an approximate elevation of 15 ft. msl. Most of the waste accepted since that time has been relatively inert construction debris. When the LFG migration project was initiated, approximately 1.25 million tons of waste were in place, and the depth of buried waste was between 50 and 100 ft. over most of the 100-ac. extent of the landfill.

Pilot Gas Control System—Module 1

In response to the first offsite detection of subsurface LFG, a pilot gas control system (Module 1), consisting of five monitoring wells and eight vapor extraction wells, was constructed. Because it was known that the buried trash had an unlined interface with the relatively permeable cobble layer between approximate depths of 25 and 50 ft. below ground surface (bgs), it was suspected that the cobble layer might be the primary pathway for offsite gas migration. Module 1 was designed specifically to test the theory that the primary pathway of gas migration was through the cobble layer surrounding the landfill.

The first step was to construct five triple-completion gas monitoring wells. Each well consisted of three 0.5-in.-diameter PVC pipes installed in a common 8-in.-diameter casing and slotted at different intervals. In each monitoring well, the deepest probe was slotted within the upper part of the cobble layer. The shallowest probe was slotted in the interval between 5 and 10 ft. bgs. The third probe was installed with its slotted section midway between the other two, approximately in the interval between 15 and 20 ft. bgs. Each slotted interval was isolated from the others by sand-cement bentonite grout.

Click here for larger view.

The motivation for installing triple-completion wells was to explore the effect of the unique site geology on the pattern of gas movement. The use of multiple monitoring zones made it possible to observe the relative magnitude of response, and the temporal sequence of responses, of methane concentrations in the two subsurface facies (cobbles and overlying silty clay) while gas extraction activities were conducted nearby.

In response to documented methane monitoring indicating that methane was consistently present in three of the five monitoring wells at concentrations greater than 5%, a row of eight gas extraction wells on 225-ft. spacing was constructed about 10 ft. from the perimeter of the landfill, which was as close to the landfill property line as the drill rig could work. The intent was to construct the wells where they would intercept the undisturbed cobble layer immediately outside the mass of buried waste, and in fact all eight borings did encounter undisturbed cobbles. The top of the cobble layer was encountered in each boring at depths ranging from 22 to 31 ft. bgs, and all extraction wells intercepted about 25 vertical ft. of cobbles. All borings were terminated in the hard sandy-silt native material below the cobbles at approximately 60 ft. bgs (the depth of the deepest buried waste) and completed as vapor extraction wells to their full depth. Although the potential exists for minor concentrations of LFG to move downward within a soil profile as a result of concentration gradient-driven diffusion, negligible downward flux was expected because of the buoyancy of methane (specific weight equal to 0.55 times that of air) and the existence of the highly porous cobble layer. Accordingly, no vapor extraction infrastructure was installed deeper than the deepest buried waste.

The extraction wells were constructed in 10-in.-diameter boreholes drilled with a cable tool drilling rig. Well casings of 4-in.-diameter Schedule (Sch.) 40 PVC were installed in each borehole. Blank casing was installed from ground surface to 5 ft. bgs, and factory-slotted (0.020-in. slots spaced at eight slots per inch) casing was installed from 5 ft. bgs to the borehole terminus. The borehole annulus was filled with washed pea gravel opposite the slotted casing intervals and with hydrated bentonite pellets opposite the blank-casing intervals. The unusually coarse filter pack was appropriate, given the large average particle size of the target zone of remediation, the cobble layer.

The placement and construction of the extraction wells were different from typical active gas-control system designs described in the product literature. The placement of perimeter extraction wells is usually in the waste itself, rather than in native material immediately outside the waste. This is because, at most landfills, the waste is more permeable than the surrounding native subsurface. The subject landfill, with its surrounding cobble stratum, did not fit this characterization.

The design of the extraction wells also differed from typical designs described in the literature. The surface seals of gas extraction wells are typically one-third to one-half the depth of the well. For the pilot system, there was a rationale behind installing wells with much shorter surface seals and long, perforated intervals. For the purposes of the pilot program, it was desirable to create a sphere of extraction-well influence at the elevations of all monitoring-well completions, not only at the elevation of the deepest monitor probes (the probes in the cobble layer). It was reasoned that if the monitor probes in the cobbles responded faster and more strongly than shallower monitor probes when gas was being withdrawn from all elevations through the subsurface, the hypothesis that the cobble layer was the primary pathway of gas migration would be strongly supported.

Siting the wells outside the waste offered other advantages besides providing an opportunity to test the theory that extracting gas from the cobble layer would be the most effective way to mitigate offsite gas movement. Telescoping casings, required where subsidence is a significant problem, were avoided. Interference with, and gas system damage resulting from, daily landfill operations was minimized.

Each wellhead was provided with a sample port controlled by a 0.5-in.-diameter ball valve; a 10-ft. section of straight, 2-in.-diameter Sch. 40 PVC pipe with a 0.5-in. hole to accommodate flow readings with a Pitot tube and differential pressure gauge; and a 2-in. ball valve for flow control. Connections from each well to the gas transfer header were made with wire-reinforced flexible hose to accommodate movement of the aboveground gas header in response to daily temperature fluctuations.

The 2-in.-diameter pipe in the extraction-well discharge piping was needed to accommodate flow rate measurement with Pitot tubes and differential pressure gauges because relatively low gas flow rates were anticipated. Pitot tubes measure two pressures inside the pipe: static pressure (which acts equally in all directions) and total pressure (the sum of static pressure and velocity pressure, which acts in the direction of gas flow). Velocity pressure is calculated as the difference between total and static pressure, gas velocity is calculated as a function of velocity pressure, and gas flow rate is calculated as a function of velocity. In a 4-in. pipe, the flow rate must be about 30 standard cubic feet per minute (scfm) before velocity is great enough to create a total pressure measurably different from the static pressure. By constricting the gas flow path, velocity is increased and lower flows can be measured. In a 2-in.-diameter pipe, flow rates as small as about 7 scfm can be measured. Pitot tubes cannot be used in pipes smaller than 2 in., the measuring of flow rates inside 2-in. pipe employed the most accurate method available within economic constraints.

The gas transfer header to which each well was connected was made of 4-in.-diameter PVC pipe with a wall thickness of 0.075 in. and a standard dimension ratio (SDR) of 56. The gas header was hung on the perimeter fence, and its grade was established to trend toward two low spots, where condensate was accumulated and released from the gas transfer path via simple manometers (U-tubes with a vertical height greater than the gauge vacuum inside the pipe at the condensate release point). The gas transfer header was connected to a small regenerative blower with a capacity of 110 scfm and an applied vacuum of 20 in. of water column (WC) at the blower intake. The relatively thin-walled thermoplastic pipe had adequate collapse resistance to withstand this magnitude of vacuum, even at summer temperatures of 110°F.

For the first six months of operation (June to November 1992), the gas flow rate from each extraction well was approximately 14 scfm and applied vacuum at the wellheads ranged from 0.1 to 0.7 in. WC. By November 1992, it was apparent that offsite methane concentrations had more or less stabilized at levels that were lower than preextraction levels but still higher than 5% at most locations. It was reasoned that increasing the flow rate per well would result in more effective control of gas migration. A focused extraction plan was developed to quantify the extraction rate required to bring all monitoring wells into compliance with the maximum 5% regulatory mandate.

The focused extraction procedure was directed at gas monitoring well B (MW-B) and initiated on January 2, 1993. The extraction rate at wells near MW-B were doubled to 27 scfm by closing the control valves of four distant extraction wells and rebalancing the remaining nearby wells to equal (increased) flows. On February 1, the extraction rates near MW-B were further increased to 37 scfm by closing one additional extraction well.

The results of the focused extraction trial are illustrated in Figure 2. The figure shows how the methane concentrations observed at MW-B decreased as the gas extraction rate employed at the extraction wells closest to MW-B was increased. The application of a 27-scfm-per-well pump rate for one month did not bring the top zone of MW-B into compliance, although a reduction from 11% to 7% methane was achieved. When the pump rate was increased to 37 scfm per well, methane concentration in the top zone of MW-B dropped from 7% to 3% in less than two weeks. Methane in the top zone of MW-B remained below 5% for the rest of this focused extraction trial, which was concluded in mid-April. It was concluded that a gas extraction rate of 37 scfm, at an applied vacuum of 1-2 in. WC, was appropriate for achieving compliance with extraction wells spaced at 225 ft. at this site. Furthermore, the sequence of responses of the three completions at MW-B was clearly consistent with the hypothesis that the primary pathway of subsurface gas circulation in the area of the landfill was through the cobble layer. The lower monitor zone responded first and most strongly to gas extraction, and the top zone demonstrated the slowest and weakest response to gas extraction, with the middle zone intermediate.

Main Gas Control System—Module 2

The design of Module 2, situated along a different section of the landfill perimeter where the same subsurface lithology existed, incorporated the successful features of Module 1. A row of 20 gas extraction wells spaced at 200 ft. was built in undisturbed native material, 10 ft. outside the limit of the buried waste. The slotted interval of each well was restricted to the cobble zone. A positive-displacement blower capable of delivering 37 scfm per well was installed. The design vacuum was a minimum of 10 in. WC at all points in the gas collection header. The pipe schedules used were Class 63 PVC (SDR 64) for all points in the collection system where applied vacuum was expected to be 50 in. WC or less; Sch. 40 PVC (SDR 24) where applied vacuum was expected to be a maximum of 80 in. WC; and Sch. 80 PVC (SDR 15) at the blower intake manifold, where continuous vibration and an applied vacuum of 100 in. WC were anticipated.

Before construction of Module 2, offsite gas monitoring wells were showing methane concentrations between 35% and 45% by volume. After 10 days of gas extraction, all wells had either nondetectable methane levels or concentrations of less than 5%.

Condensate Handling

To collect and treat condensate from the extracted gas, seven condensate traps were situated at low points in the gas transfer path, where the accumulation of condensate was anticipated. Two sumps were placed at centralized locations to accumulate the condensate collected from these traps and transfer the condensate tinder pressure into the air-stripper feed header. Wherever possible, condensate transfer piping was configured to transmit collected condensate by gravity flow. In some cases, collected condensate was pumped back into the gas collection header in order to transmit it to the closest sump without constructing a separate line dedicated exclusively to condensate transmission.

A key part of the condensate collection system was a 6-in. electrically actuated motor valve installed at the vacuum blower assembly, on the suction side of the blower. The motor valve was sized to accommodate the entire intake capacity of the blower (8-in. inlet and discharge plumbing). The motor valve periodically opens and thereby reduces the vacuum applied to the gas collection system to zero. The relief of applied vacuum allows condensate to be released at various points throughout the gas collection system so that collected condensate can flow by gravity out of the gas header, through check valves, and into traps. The timing of the operation of the motor valve is controlled by a liquid level float switch in the moisture knockout pot (KO pot), located at the blower intake, where the rate of condensate accumulation was expected to require the most frequent discharge.

The need for periodic vacuum relief to facilitate the release of collected condensate from the gas transfer path was the result of the relatively high vacuum in the gas header. The applied vacuum in the gas collection header ranges from approximately 10 in. WC to 80 in. WC and can be as high as 100 in. WC in the KO pot. For a condensate trap to allow the continuous release (without vacuum relief) of condensate from the gas header, while still preserving the vacuum inside the gas header, a U-tube arrangement with a height equal to the applied vacuum would be required. The U-tube would have to be situated entirely below the elevation of the invert of the gas header, which in this case would have required numerous facilities to be constructed between 6 and 8 ft. bgs. The installation of numerous deep U-tubes was rejected because of the required excavation.

Engineered check valves, with cracking pressures precisely matched to the line vacuum at the trap location, were considered as an option to reduce required excavation depths. They were rejected because the line vacuum changes daily as a result of changes in barometric pressure, changes in well-field configuration or well flows, and changes in the condition of the ground-surface seal around the extraction wells. The dependable operation of engineered check valves could not be guaranteed in an application for which the required cracking pressure was not constant from day to day.

The cost-effective solution to the problem of releasing condensate from high-vacuum sections of the gas header was to periodically reduce the applied vacuum to zero, permitting the drainage of condensate through shallow U-tubes incorporating standard zero-cracking-pressure check valves. The check valves were installed in an orientation in which they would be opened by the weight of the collected water upon relief of vacuum and would be closed by atmospheric pressure upon reestablishment of vacuum.

Guest author Stephen Ferry is a consulting engineer based in Napa, CA.

MSW - July/August 2002