Frakturering

Bilag 6

Assessment of Hydraulic Fracturing Potential at Vestergade 10, Haslev, Denmark

1. Introduction

Continuing dry cleaning operations at the site have contaminated the underlying soil and groundwater with tetrachloroethylene (PCE) and daughter compounds. The sources and extent of contamination, as well as geology and hydrology, are thoroughly described and mapped by NIRAS (1999). In general, the chlorinated solvents are trapped in the upper clay till, which is wet. An underlying primary aquifer has suffered minimal though measurable impact. The contaminants have not spread laterally beyond the limits of the premises, so the remediation target remains compact. However, the principal structure, several outbuildings, and neighboring properties preclude excavation and thus constrain remedial options to recovery or in situ processes. Dual phase extraction has been shown to be marginally effective for recovery of the contaminants.

The site is proposed for demonstration of hydraulic fracturing in Denmark because of the characteristics of the soil, the type and distribution of contaminant, the known performance of conventional remedial technologies such as dual phase extraction, the typical urban setting, and the proximity to the Haslev gasworks site where soil has been extensively characterized. Presumably horizontal or sub-horizontal fractures can be created to enhance dual phase extraction at the site.

On June 21 and 22, 1999, Bill Slack, vice president of FRx Inc., toured the site and was briefed by the staff of GEUS and representatives of NIRAS, NCC, and Broker Drilling. FRx provides hydraulic fracturing and related services in the United States. The tour and briefing provided insight that coupled with FRx experience permits an assessment of the viability of hydraulic fracturing at the site. This report, which transmits our evaluation, was requested to support subsequent decisions and actions at the site.

2. Background – Induced Hydraulic Fractures

Hydraulic fracturing involves injecting fluid into an open well at a pressure such that a crack, or fracture, forms within the surrounding soil, and continuing injection to dilate the fracture and fill it with beneficial material. Most fractures are created with slurries of granular solids that have desirable chemical or physical properties. The first, and still most widely used application, couples sand-filled fractures to soil vapor extraction (SVE) or pump-and-treat recovery. This technique results in highly permeable layers that increase the flow rate of a well, typically by one to two orders of magnitude. A variety of other applications have been demonstrated since the initial use of hydraulic fracturing twelve years ago in United States Environmental Protection Agency (USEPA) research projects. Also, refinements in the techniques have broadened the spectrum of sites and problems that can be addressed.

A sequence of USEPA research and demonstration projects developed hydraulic fracturing as a tool for restoration of contaminated soil and ground water. Initially, the project objectives envisioned methods to improve the flow rate of recovery wells, such as those used in soil vapor extraction (SVE) or pump and treat. The development borrowed heavily from the technology used to increase petroleum production from low permeability reservoirs. In general, the effort defined the smaller scale methods that are required by shallow surface soils and identified or proved several useful applications for hydraulic fracturing at contaminated sites. The USEPA Superfund Innovative Technology Evaluation Program (SITE Program) sponsored an independent review of the technology and published the findings in a widely distributed report (USEPA 1993).

Since 1994 hydraulic fracturing has been commercially available. The first commercial applications were installed as part of SVE projects. These projects typically were of small scope, as might be expected for an innovative technology, and not heavily publicized. Murdoch et al. (1995) presents some case histories. Bilag 3 lists projects undertaken by FRx.

One project has several parallels to the Haslev site. In 1995 and 1996 hydraulic fractures were created at the Linemaster Superfund site in Connecticut to enable dual phase extraction of chlorinated solvents, which were trapped in surficial glacial tills. In a pilot test a, water recovered through fractures induced a fifteen-foot deep cone of depression in the water table in three weeks, whereas a conventional well failed to have any measurable effect. Air movement through the exposed soil column yielded satisfactory recovery of TCE and daughter compounds. Pilot results justified more widespread application. Use of fractures in six contiguous wells has resulted in a 30-foot deep and 200 ft wide cone of depression in the water table in six months.

More recently, hydraulic fracturing techniques have been used to create permeable reactive barriers that intercept and destroy contaminants. Iron-filled vertical fractures have been created in aquifers to remediate chlorinated solvent plumes, and horizontal fractures in low permeability media have been filled with a variety of materials to immobilize or destroy contaminants released from overlying sources. Thus fracturing should be considered an enabling technology, but is not a remedial process by itself

Certainly any of a variety of methods can be used to create horizontal hydraulic fractures. A 1994 USEPA publication (Murdoch et al. 1995) outlines and compares several. Furthermore, any method can be fine-tuned to most effectively satisfy project and site constraints. The site-specific factors that weigh in selection of methods include target depths for the fractures, remedial processes to be used, surface access, subsurface obstructions such as utilities or existing wells, schedule, budget, etc.

Nonetheless most horizontal fractures have been created in accordance with methods developed by the USEPA research projects conducted in the late 1980’s and early 1990’s. In summary, these steps include (1) installing a dedicated well by hammering to desired depth a piece of 2-inch pipe fitted with a drive point, (2) dislodging the drive point downward to expose a short section of open hole, (3) cutting a thin kerf in the wall of the borehole by means of a horizontal hydraulic jet, (4) pressurizing the kerf with liquid so as to nucleate a horizontal fracture from the hoop that constitutes its outer edge, (5) delivering sand-laden slurry to the open hole section of the well so as to propagate the fracture, and (6) monitoring the injection pressure and surface deformation, which permits deduction of the fracture form. This sequence of steps constitutes one of the simplest, reliable, and effective methods for creating fractures. More details can be found in Appendix "B". The method results in the maximum degree of control over the fracture, both during creation and during ultimate use.

Multiple fractures can be made from a single well, but the creation process and effective utilization become more complicated. For each fracture, an open section of borehole must be exposed by cutting away a section of casing. Straddle packers are used to isolate each open section of borehole during creation of the fractures. Individual access of each fracture, which is certainly the preferred method for operation of most remedial processes, requires packers with multiple passages. Such well completion components may need to be fabricated especially for the project, incurring expense as well as requiring long lead-time. The size of equipment used in the well, both during creation and subsequent operations, requires use of larger wells that are six or eight inches in diameter.

Rarely can existing wells, which typically have long screen sections, be used for creation of horizontal hydraulic fractures. First, the screen and surrounding gravel pack, which have been carefully engineered to exclude soil particles, needs to be cut away to permit passage of fracturing slurry. More importantly, pressurization essential for nucleation of a fracture is applied to the entire screened or gravel packed interval, whichever is longer. Pressurization of such a long section of borehole will tend to result in a vertical fracture through the axis of the borehole. Even if a horizontal fracture could be nucleated and propagated, the resultant combination of a linear well and a planar fracture does not effectively enhance remedial processes, especially those that rely upon fluid flow to or from the well.

3. Site Characteristics Relevant to Fracturing

Geological factors influence the form of induced fractures a strongly as variations in the methods of creation. State of stress and toughness of the formation exert principal control. Several other geological or geotechnical characteristics also correlate with fracture form, but these relationships result, at least in part, through effects on the predominant factors.

The state of stress in a formation affects orientation of an induced fracture once it has propagated away from the borehole. Fractures are usually flat lying where horizontal formation stresses are greater than vertical stresses, whereas they tend to be steeply dipping where vertical stresses are greatest. The state of stress of soils and unlithified sediments depends on several factors, including consolidation history, and wetting and drying history. Soils that were consolidated under a load greater than the present load are overconsolidated, and many such soils contain horizontal stresses that exceed vertical stresses. For example, glacial loads can result in overconsolidation, so soils deposited subglacially are good candidates for high lateral stress. Soils containing clay minerals that undergo large volume changes in response to changes in moisture content can become overconsolidated with repeated cycles of wetting and drying. For instance, vertisols (soils rich in swelling clays) are particularly susceptible to large lateral stresses. Soils of poorly sorted particles can be dense, and stress-inducing mechanisms, such as specific volume increases due to oxidation, can readily effect high lateral stress. Thus soils derived from weathered bedrock can sustain horizontal fractures. In some cases lateral stresses are greater in surficial soils, which have been heavily weathered, than in similar underlying units.

The toughness of the soil at the fracture tip, coupled with the elasticity of soil surrounding the fracture, determines whether a horizontal fracture is thick and confined close to the injection well or thin and of large extent. Tougher soil limits propagation of a fracture and favors thicker aperture. Anisotropic toughness can occur along contacts between different strata. Induced fractures may follow contacts in interbedded sediments. The effect of bedding can be capricious, with fractures following beds in some cases and crosscutting beds in others.

Permeability has little direct effect on fracture form but deserves careful attention because of its impact on the application of fractures and also on the creation process. Permeability critically effects the relative performance of wells installed to recover fluid. In order for wells with fractures to discharge at rates substantially greater than conventional wells, the fracture sand should have permeability more than 1000 greater than the surrounding media. If the target soils have sufficient permeability, fractures can not aid the greedy operator who desires faster recovery of contaminants. In cases where fractures are being created rationally in permeable media, such as for the construction of a permeable treatment barrier, the permeability of the media provides a mechanism for separation of fluid and granules in the fracturing slurry. Fracturing techniques are available for such circumstances.

The role natural fractures play in determining the form of induced fractures has not been extensively explored. Few sites have undergone thorough evaluation of natural fractures, and fractures have been created at yet fewer. Flat lying fractures have been created in the upper 5 meters of glacial till near Sarnia, Ontario, Canada, for the purpose of characterizing in situ flow around fractures. Below 5m, vertical fractures were favored. In separate projects, naturally occurring fractures have been described in similar soil nearby. Several natural steeply dipping fractures were traced from the surface to depths of 5m and appeared to have resulting from weathering processes. By inference, a predominance of near vertical natural fractures at a site does not preclude creation of useful horizontal or sub-horizontal fractures.

Water content of a formation appears to have negligible effect on creating fractures by injecting fluid.

The distribution of contaminant within the impacted soils generally effects only the locations selected for fractures and has essentially no impact on the resultant form of the fracture.

Ideally, the fractures will be placed to optimally remediate the site. Exact placement depends, of course, upon the radius of influence of a fracture. Fractures should be concentrated around hot spots or source zones. For remedial designs involving fluid recovery, at least one fracture should be placed at the down gradient limit of contamination.

Fractures can have an optimal orientation for intercepting or recovering contaminants. In homogeneous media, selecting the fracture orientation follows from evaluation of streamlines in the system. In multiple porosity media, such as naturally fractured clay, contaminants may be preferentially distributed in flow channels. Fractures should be oriented to intercept these natural flow paths so that the remedial processes can address the greater portion of contaminant.

A common concern is that fracturing may dislodge or otherwise move and disrupt structures or utilities at a site. The concern arises from the very real fact that creation of a hydraulic fracture displaces surrounding soil by a few millimeters. The magnitude of displacement depends upon the form and size of the fracture as well as the distance between the fracture and the point of interest. In the case of shallow horizontal fractures, the overlying soil and ground surface will be displaced upwards a distance that correlates closely to the aperture of the fracture. Displacement will taper to zero within a short distance outside of the extent of the fracture. At greater depths, the amplitude of displacement is diminished but a larger area is affected. (A pea under a mattress is a simple analogy for this effect.) Elastic soils attenuate the displacements caused by fractures.

Displacement is greatest at the conclusion of fracturing. Afterwards, the fracture closes and the dome of overlying soil subsides. The injected sand prevents the fracture walls from closing completely. The amount of contraction depends on the concentration of sand in the slurry. Here the ratio of maximum aperture when the fracture is pressurized to thickness of resultant sand pack after the liquid separates is similar to the ratio of total slurry volume to bulk volume of sand in the slurry.

Across a site, the displacements experienced by structures and utilities are gradual because the upward displacement follows the aperture of the fracture, both during and after fracture creation. Small fractures created at depths of two to five meters in glacial till, as probably will be the case at Vestergade 10, may have a radius three to four meters and maximum uplift of 1 cm. This 1:300 gradient can be tolerated by many structures as well as utility lines, which are often constructed to accommodate strain induced by temperature changes or subsidence. Rarely will steep gradients be created at the surface, although slabs of concrete paving, which are stiff and non-bending, may shift to reveal throw of several cm at their edges. Nonetheless, counsel of structural experts should be sought if structures are considered delicate and valuable.

Surface structures can impact the propagation of fractures. The trenches and excavations used to install subsurface utilities represent a path of weak soil. If the tip of a propagating fracture intersects such a feature, it will either quickly penetrate to the ground surface, or (less frequently) propagate along the bottom of it. Fortunately, most utilities are installed close to the surface, above the target depth of many fractures. Note, shallow, sub-horizontal fractures may easily climb to the depth of utilities.

The foundations of buildings and other heavy structures can have a dual impact upon the propagation of fractures. Like subsurface utilities, the bottom of a foundation may intercept a fracture and channel slurry to the ground surface. In such case, the fracture clearly does not propagate beyond the line of the foundation. Deeper fractures that do not intersect the foundation can be influenced by the weight of the overlying structure. Multi-story masonry structures or large tanks filled with liquids impose sufficient load to alter the in situ state of stress that governs fracture form. At the least, these structures deflect propagating fractures. In the extreme, the structures may induce conditions that favor formation of vertical fractures. The phenomena are not sufficiently reliable to be used to advantage during the selection of fracturing locations. A building that appears stout may have little influence on fractures while a light frame building may seem to repel fractures.

The presence of structures and utilities at a site usually complicates the activities of creating fractures. Positioning of a drill rig to create fracture wells requires consideration of overhead electric and communication lines, the course of crushable sewers etc., and the existence of buildings, alcoves, porches, etc. Walls and fences, even with doors or gates, represent obstructions to hoses that connect the fracturing equipment to the fracture well. Likewise, active roadways or railroads disrupt site activities. These factors do not prevent application of fracturing, but must be taken into account during planning for the work.

The creation and use of a fracture among existing monitoring wells and borings frequently arises at heavily characterized or confined site. Many effects can occur, and a definitive answer may not be readily forthcoming nor common among sites.

An attempt to create a fracture near an existing well may be frustrated by the tendency of the well casing to suppress opening of the fracture aperture. In a simple sense, the well can act as a reinforcing pin that holds the earth together. This effect is especially pronounced if the well is within a meter of the fracture nucleation point and if the fracture intercepts solid casing as opposed to a screen and gravel pack. Exceptions have been observed: we have created satisfactory horizontal fractures within 40 cm of an existing well.

A more distantly offset existing well may or may not interact with a propagating fracture. In some cases fractures have propagated to and around the well casing, as indicated by sand-filled aperture exposed during subsequent excavation. In other cases the aperture pinches to zero some distance before the fracture encounters the well, only to open at some distance beyond the well, i.e. the well is centered on an island of unfractured soil within the plane of the fracture. The reasons for either phenomenon have not been delineated. The consequences limit the options for planning to connect fractures into existing wells.

If a propagating fracture encounters a borehole or screen section of an existing well, the zone can be pressurized. Unsecured borings, such as those back-filled with cuttings or poorly constructed wells with insufficient annular seal, can provide a pathway for fracture slurry to reach the ground surface and frustrate fracturing operations. Likewise fracturing pressure has popped off the caps of monitoring wells. In an extreme case, well screen has collapsed when fracture intercepted an existing well. The screen is very effective at excluding the sand in fracturing slurry while passing the liquid. The resultant filter pack provides the area over which the fracturing pressure achieves sufficient force to collapse the screen. In addition to these mechanical effects, the pressurized borehole or well provides a pathway for fracture fluid to propagate in unplanned directions, especially if a secondary fracture can nucleate elsewhere along the length of the bore. We have no direct evidence of this phenomenon, but the opportunity exists.

The intersection of fractures with long sections of well screen complicates the flow patterns of almost all remediation schemes. Flow to a fracture can be characterized as linear and generally perpendicular to the fracture face. Flow to a well screen is characterized as radial. Even in a homogeneous infinite case, the flow resulting from merger of the two patterns can not be characterized by simple calculations, and thus becomes more difficult to utilize. In reality, the well often serves as a short circuit that floods the fracture with fluid from an undesired source.

As mentioned, practically no site considered for fracturing is devoid of wells and borings. Thus creative thinking and strategic use of existing wells has become the standard practice. Rarely have wells been abandoned and plugged. Often existing wells have been used as monitoring points during creation of fractures and continue in useful function during remedial activities.

4. Suggested Fracture Design

The native soil between 2 and 5 mbgl at Vestergade 10 is a moist, less compacted clay that contains an order of magnitude more natural fractures than the underlying more compacted clay. The soil appears to have similar origin and properties to the Sarnia, Ontario, site where several flat lying fractures have been created to depths as great as 5 mbgl. Both sites are composed of glacial tills that are demarcated into upper and lower units. We fully expect that horizontal or sub-horizontal sand-filled fractures can be created between 2 and 5 mbgl at Vestergade 10.

Contaminant distribution at Vestergade 10 may necessitate creation of a fracture at 8 mbgl. Achievement of the desired sub-horizontal can not be assured. At analogous North American locations, such as Sarnia, Ontario, or near Cincinnati, Ohio, the less weathered clay units deeper than 5 mbgl do not allow propagation of sub-horizontal fractures. However, the deeper unit at Haslev may be different. Indeed, it is more compacted and presumably denser than the overlying unit and thus should be as amenable to propagation of sub-horizontal fractures. The only definitive method of determining the form is to create a fracture.

The contaminant hot spots at Vestergade 10 are fairly small areas. The largest, under the northeast corner of the main building is about four meters in width and six meters in length. The smaller hot spots, in the back yard, are less than three meters across and within five meters of the largest hot spot. At some sites, one modest fracture with a radius of five meters would be installed to address such a distribution of contaminants. However, the building and the presence of several existing borings and monitoring wells suggests that fairly small fractures at multiple locations be used at this site. Small fractures can be fit in and around the various obstructions, should still prove effective enhancements to dual phase extraction, and allow tighter control of subsurface flow.

Fractures that fit with in the confines of the site may the smallest ever created for purposes of environmental remediation. Our proprietary fracture propagation model, which is based on the mechanisms outlined in Murdoch (1993), indicates that a fracture composed of 200 to 250 kg of sand and 200 l of gel may be confined to radius of 2.5 m. The inputs to the model are listed in Table 1 and an example output is shown as Appendix D. The use of alternative but equivalently reasonable parameters in the model suggests that an equal quantity of material may propagate a fracture as far as 6 meters from the injection well. A far-reaching fracture may adversely interact with structures and utilities on the site and on adjoining properties. A field test in similar soil is the only certain method of determining exact behavior.

Table 1 – Fracture Size Calculation Parameters

Fractures created in the backyard can be larger than those created under the building because of the additional space. In Figure 1 the backyard fracture is drawn with a 5 m diameter while the fractures under the build are drawn with 4 m diameter.

The fractures should be created to directly address the most serious contamination. NIRAS (1999) has documented that PCE occurs in two hot spots, one in the backyard and the other under the northeast corner of the principal building at Vestergade 10. These hot spots appear to be confined between 2.5 and 4.5 mbgl. The moisture in the soil necessitates dual phase extraction as the recovery process. Thus two fractures should be created at each hot spot. The lowermost fracture should be placed at the bottom of the hot spot but still within the less compacted clay. At such location, it can intercept any contaminant migrating from above. Also, water should be recovered through the lowermost fracture so that pore water can be removed from the hot spot to open more pore space to airflow. The uppermost fracture should be placed in the middle of the hot spot, so that the contaminant recovery can be maximized.

Two secondary locations for fractures should be considered. Soil gas and ground water analyses show contamination under the basement stairs and below the southeast crawl space at Vestergade 10. Two fractures should be created at each of these locations for similar purposes as the hot spots.

Figure 1 indicates the recommended locations for the fractures. Placement of fractures in the backyard will required consideration of the several existing borings and monitoring wells. Fractures should be created at least one meter away from existing wells; least the wells serve as reinforcing pins and prevent opening of a fracture aperture or the wells collapse upon nucleation of the fracture. The building probably will impact fracture propagation. These facets are discussed in Section 5 of this report.

The well placement methods that can be most effectively employed within the confines of the site should be used. In most other sites an individual well is dedicated to each fracture at depths encountered at this site. The individual wells can be of made from small diameter pipe. Two-inch pipe has been widely used in the United States, but fractures have also been created through 1-inch pipe. The well should be installed as described in Appendix "B". Regardless of pipe size, the well can be installed in segments that are screwed together, permitting well installation inside the basement.

Dual phase extraction can be performed most efficiently when draw down can be effected to the elevation of the fracture. If applied vacuum approaches 250 cm of water, then water can be slurped out of the wells. If less strong vacuum is to be applied, then a pump will need to be installed at the bottom of the well. Most pumps require modest submersion to operate, so the pump should be placed below the level of the fracture. Accordingly, a hollow drive point needs to be fitted into the bottom of the casing. Furthermore, the casing diameter needs to accept the pump.

If multiple fractures are created from a single well, the well necessarily will be of larger diameter, and a drill rig will be required. Directional or slant drilling can accomplish access under the building. The borehole should be cased to bottom and a swelling grout used along the entire length of casing. We have had poor experience with use of bentonite. Cuts in the casing can be made with mechanical, hydraulic, or explosive means. Presumably the cut will be perpendicular to the axis of the well. If the axis of the well is steeply inclined, nucleation of a horizontal fracture may be disrupted.

Effective application of dual phase extraction through multiple fractures in a single well requires individual access to each fracture. A system of packers and risers must be designed for the well before the final well size is selected. As with small diameter single fracture wells, consideration must be given to placement of a pump below the fractures for efficient dual phase extraction.

Fracturing should be demonstrated in uncontaminated are of the site so that planned size of the fractures to be created at Vestergade 10 can be assured. In particular, a fracture should be created at 8 mbgl if the design incorporates such depth. Eight-meter is well within the lower till unit and is below the depth where sub horizontal fractures have been successfully created at analogous North American sites. If this preliminary work can not be conducted in the confines of the site, it should be performed in the vicinity of Haslev

The fractures should be created with the most viscous slurry and with the greatest concentration of sand that can be prepared by the equipment. This consistency should help confine the fractures to the vicinity of the injection well.

Otherwise, fracturing should follow methods developed during the course of USEPA research projects and outlined in Appendix "B".

Fractures have the potential to adversely effect all structures on the site. As discussed in Section 3.3, above, creation of a fracture displaces surrounding soil. The fracturing locations and fracture sizes have been chosen to minimize the amplitude of uplift at the walls of the building and along water, gas, electric, and telephone lines. The sewer line, from which contamination leaked, overlies some of the hot spots, so its possible disruption can not be avoided.

Nonetheless, monitoring while creating the fractures can provide warning if fracture threatens the integrity of the buildings or utilities. Monitoring also provides documentation of the impact of fracturing.

Too many walls and blind corners exist at the site for surveying techniques to follow the course of fracture propagation. Besides, surveying is too slow to provide timely warning of detrimental reactions to the fracture. Rather, electronic sensors should be deployed. Strain gauges can be placed at critical points along the building foundation and above utility lines. Nearby monitoring wells should be loaded with water and transducers installed to alert operators upon fracture penetration into the wells. Tiltmeters can be used to measure deformation of unrestrained ground surface. The small fractures recommended herein should not cause significant soil displacements more than five meters away from an injection wells.

Success of the monitoring program depends upon the quality of the underlying plan. The key elements include selection of critical points for placement of monitoring sensors, identification for each sensor of the threshold signal that will invoke intervention or other action, and establishment of a set of actions, i.e. protocols, to be followed.

5. Probable Results of Creating Fractures

Fractures create at the target intervals listed above, i.e. between 3 and 4.5 mbgl, most likely will be sub-horizontal. That is, the fractures will nucleate in plane established by the techniques employed (presumably a horizontal notch will be cut), initially propagate horizontally, but gradually assume increasing dip. Dip at the extreme limits may approach 20°. The configuration after creating two fractures will resemble a pair of poorly nested saucers. The fractures will be more elliptical than circular, with major and minor axes of 6m and 4m, respectively. Maximum fracture aperture when pressurized with slurry will be on the order of 1 cm, and the average closed aperture after separation of the fracturing fluid may be 3 mm.

The fractures will be similar to fractures created in southwest Ohio (near Cincinnati) and Sarnia, Ontario, Canada. These sites have similar geological setting and fractures created there have been extensively characterized. Murdoch (1995) presents more details.

The majority of the utility services do not appear to lie within the target areas for fracturing. Gas, electric, telephone, and water services are laid under the street of Vestergade or the adjoining sidewalk. These utilities appear to be installed in the upper 80 to 100 cm of soil. Fractures are not expected to interact strongly with these utilities because of the distance from the fracturing locations and because the building foundation shields them from the injection wells. The service lines into the building are suspended from or integrated into the frame of the building and need not be considered separate from the building

The sewer system overlies the hot spots – as might be expected since it was the source of contamination. The sewer has several branches and laterals throughout the backyard. Fracturing activities can not avoid the sewer system. Hopefully the sewer can tolerate some displacement. Sewers, which are either short lengths of clay pipe or longer lengths of plastic pipe, are usually fairly flexible. Any sections damaged by fracturing should be easily replaced.

The main building on the site, and the building over much of the contamination, is of multi-story masonry construction. A basement has been dug under the northern half, while only a crawl-space provides access under the southern portion. The foundations extend under the exterior walls and under a central east-west load-bearing wall. Thus the soil underneath the building is divided into a north and south half. An examination of the exterior and interior walls revealed no indication of settlement, i.e. no pointing traces could be seen in the exterior walls and the interior walls were smooth and without cracks. In summary, the building appears to be heavy and well constructed. Any disruption of the building will be readily evident as unique cracks.

The building probably will block propagation of hydraulic fractures. The weight alone should be sufficient to deflect the fractures. The foundations may extend deep enough to intercept a sub-horizontal fracture nucleated at 3 mbgl. If so, fracture slurry will vent to the surface along the foundation. Nonetheless, the chance that the fracture may deform such a pristine building warrants the monitoring program defined in Section 4.5.

Secondary buildings at the site are of light construction and do not overlie any fracturing targets. There should not be any interaction between them and the fractures.

Fractures will be created at the site to enhance performance of dual phase extraction for recovery of PCE from moist clay soil. Previous tests at the site have yielded unacceptable vapor discharge and radius of influence.

Projection of system performance can be obtained by analogy to two sites where fractures were employed as remedial tools.

At the Linemaster Superfund Site in Connecticut, hydraulic fractures were created to enable dual phase extraction of TCE and daughter products from dense, silty glacial till. Fractures allowed draw down in the water table in a matter of days where none could be effected by conventional wells. Air movement through the exposed soil column recovered contaminants at a satisfactory rate. Linemaster differs from this site in that its contamination is spread through a 15m thick column. Nonetheless, the similarities in process and targets encourage a similar favorable projection for Vestergade 10.

Quantitative projections of well discharges and extent of influence draw upon flow characterization work conducted during the USEPA research projects of the early 1990’s. Fracture response at the Center Hill Site is presented in the EPA SITE report (1993) and discussed at length in Murdoch (1995). About 4% of the vacuum applied to a 6m diameter fracture located 1.7 mbgl could be detected over 6m away from the well. In contrast, similar influence of a conventional well was less than 2m. Although this comparison is encouraging, it needs to be considered in the context of the Vestergade 10 pilot test. A dual phase extraction pilot at Vestergade 10 exerted influence over 4m. By extrapolation, creation of a 5m diameter fracture may extend influence to only 5m. The contrast between 4m and 5m is not nearly as spectacular as between 2m and 6m. Thus the effect of fracturing may be less evident than presumed.

6. Conclusions

Our experience suggests that useful flat lying fractures can be created in the soil at the Vestergade 10, Haslev, Denmark. The soil has origin and character similar to sites in North America where several flat lying fractures have been created. At the particular North American sites, flat lying or shallow dipping fractures have been created at various depths that include the intervals expected to be fractured at Vestergade. The fractures were created for various applications, including dual phase extraction as is contemplated at Vestergade.

Contamination at Vestergade will need to be addressed by creating fractures at multiple locations and depths. In addition to hydrologic design issues, which yet have not been resolved and may well require multiple fractures, structures at the site preclude creation of fractures from a single location; uniform propagation under the various building foundations can not be assured. Fortunately, smaller fractures minimize the risk of adverse impact on the building and utilities. Smaller fractures also coincide well with the size of contaminant hot spots.

Performance of dual phase extraction operated through fractures should substantially exceed what was accomplished during a dual phase extraction pilot test that utilized conventional wells. The advantages should be realized as three characteristics, all of which contribute to accelerated contaminant removal. First, fractures effectively extend the radius of influence of wells in low permeability material. At all sites where flow to or from induced fractures has been measured, the radius of influence exceeds the size of the fractures, often by a factor of three or more. At Vestergade, short term SVE tests have shown influence to extend less than 4m (Well B12 to B10), which is certainly smaller than the anticipated fractures. Secondly, fracturing can improve the discharge from wells. Specifically, conventional wells at analogous sites in the United States discharged air and water at rates similar to those reported in the long-term SVE tests at Vestergade, while wells completed with fractures discharged at rates an order of magnitude greater. Finally, proper operation of multiple fractures can develop a wide cone of depression in the phreatic surface, thereby exposing large volumes of soil for vapor flow into overlying fractures. At the Linemaster Superfund site, this approach has proved successful when a similar configuration could not be effected by conventional wells.

In summary, fracturing should be undertaken at 10 Vestergade, Haslev for the purpose of recovering the contaminants from the subsurface soils and groundwater.

7. References

NIRAS. "Summary of Soil and Ground Water Investigations at Vestergade 10". Internal memorandum. June 1999.

Murdoch, L.C., D. Wilson, K. Savage, W. Slack, and J. Uber. "Alternative Methods for Fluid Delivery and Recovery". USEPA/625/R-94/003. 1995. www.epa.gov/ordntrnt/ORD/WebPubs/fluid.html (for entire document download in PDF format) or www.epa.gov/clariton/clhtml/pubtitle.html (search for 625R94003 for page-by-page on-line viewing)

USEPA Risk Reduction Laboratory and The University of Cincinnati. "Hydraulic Fracturing Technology - Applications Analysis and Technology Evaluation Report" USEPA/540/R-93/505. 1993. www.epa.gov/ORD/SITE/reports/051.htm (for entire document download in PDF format)or www.epa.gov/clariton/clhtml/pubtitle.html (search for 540R93505 for page-by-page on-line viewing)

Murdoch, L.C. "Hydraulic Fracturing of Soil During Laboratory Experiments, Part III: Theoretical Analysis". Geotechnique, 43 No 2, 277-287. 1993.

Figure 1 – Proposed Fracture Locations – Vestergade 10, Haslev

Two fractures should be created at the center of each shaded circle. In each set, the uppermost fracture should be nucleated at 3 to 3.5 mbgl and the lowermost at 4.5 mbgl. Fractures within the building have smaller (4m) diameter to fit within the confines of foundations and utilities. The backyard fractures can be larger and are shown with a 5m diameter.