Miljøprojekt, 1108 – Frakturer fra lodrette testboringer på Vestergade 10 Haslev

Frakturer fra lodrette testboringer på Vestergade 10 Haslev

Summary and conclusions

The objective of this project has been to establish whether vertical remediation wells extended with hydraulic fracturing will be a cost-effective remediation technique in low permeable glacial deposits. At the test site, fracturing has been applied to improve the hydraulic effect of vertical test wells into the till. Remedial pumping from the vertical wells was carried out by Dual Phase Extraction (DPE).

The report gives an account of the design of the remedial system, experience during the physical establishment of the system, and the results of a series of hydraulic tests of the performance of the installed wells, before and after the fracturing.

The site geology comprises fill in the upper 1-2 meters with till beneath this to a depth of approx. 17 m, where the “Grønsandskalk” (Paleocæn) limestone layer is found. In the till layer, a few thin layers of silt and sand schlieren were found.

In the fill layer and in the upper till layer, ground water was found near the surface, in an area where the water level was approx. 1-1.5 m below surface. The “Grønsandskalk” makes up the primary reservoir. In this confined primary reservoir, the hydraulic pressure level was found at a depth of approx. 6 m.

Slug tests carried out in short filters show a hydraulic conductivity in the till of approx. 1.7E-06 m/s in the upper 2 m below ground level, decreasing to approx. 6.7E-09 m/s in a depth of 4.5 m below ground level, and hereafter the value increases to approx. 2.0E-09 m/s in a depth of 8 m.

On basis of the feasibility project, it was recommended that the most suitable remediation technique for this locality would be to extend ordinary vertical remedial wells with hydraulic fracturing, followed by remedial pumping using DPE, and in this way, extract water and air from the polluted soil.

Based on experiences in other countries, it was expected that the fractures induced by hydraulic fracturing would spread to a distance of 3-5 m from the injection point. However, the effect of hydraulic fracturing has often been found to be approx. 3 times the physical extent of the induced fracture.

To assess the risk of unacceptable impacts on existing buildings and underground installations in connection with the fracturing work, geotechnical as well as constructional investigations were carried out. Based on the geotechnical parameters, the elevation in the terrain was calculated. On this basis, it has been assessed that it will be possible to apply the fracturing principle to the test site, also in cases where the remedial wells are placed close to existing buildings.

Based on a series of functional demands made by both builder and consultant, the contractor with assistance from his American consultant has designed the fracturing process, and carried out the fracturing work. The fracturing was established from 2 vertical test wells and carried out in depths of 4.5 and 8.0 m. The work was carried out in 2 days and is reported in a separate field report, enclosed as appendix D.

The elevation in the terrain was measured during the fracturing process by levelling of a number of field points. The upper fracture was found to cause a maximum rise in the terrain of approx. 10 mm, and the fracture radius approx. 5 m. The similar rise was in a depth of 8 m approx. 3 mm, and the radius approx. 3 m.

The mapping of the actual spreading of the fractures was examined by visual inspection of core samples taken out in distances up to 2 m from the test wells. In 3 out of 6 core wells, it was possible to identify the upper fracture. Based on the identification of the fractures, it can be concluded that the upper fracture is sub-horizontal and tends to seek in a downward direction. The aperture of the fracture was measured to be 3-10 mm, and the fracture radius approx. 2 m in the eastern direction. In some directions, it was not possible to identify the fracture. The lower induced fracture was identified in a sand layer, not earlier identified, approx. 0.5 m below the injection point. It is assessed that the induced material has spread in this sand layer, and thus the fractures established in the till did not spread very far.

It can be concluded that it is possible to establish 1-10 mm sand-filled fractures with a range of at least 1-2 m from the injection point. It is, however, extremely difficult to predict the orientation of the fractures.

To be able to assess the hydraulic effect obtained by hydraulic fracturing, a test of the double-phased extraction method DPE was carried out from each of the test wells before and after the fracturing. During the DPE tests, the changes in the groundwater level were measured in 3 depths from 0.5 to 5.0 m in the established monitoring wells, corresponding to a total of 20 monitoring filters.

The DPE tests performed prior to the hydraulic fracturing show that neither was it possible to extract air nor water from the borings, and nor was it possible to establish any effect in the adjacent monitoring filters (distances of 0.5 to 5.0 m). Consequently, each test was stopped after approx. 24 hours.

The DPE tests performed after the fracturing were carried out over a course of approx. 7 days, and the results showed a substantial improvement of the hydraulic conditions. Using the DPE, an air flow of approx. 3.5 l/s and a water flow of approx. 0.006 l/s (0.5 m³/day) were measured in a depth of 4.5 m, and the groundwater level in the monitoring filters nearby were found heavily affected at all 3 levels. Furthermore, it was possible to establish a clear effect at a distance of approx. 8 m from the pump well. On this basis, it is assessed that the influence radius in this test is up to 10 m, which is approx. 2 times the radius of the elevation in terrain. Further, the results show the establishment of a direct hydraulic connection between the upper pump well, 4.5 m below surface, and several of the monitoring filters in a depth of 8 m. This is probably due to the fact that the established fracture is downward with a sub-horizontal orientation. The induced fracture is thus in hydraulic contact with the sand layer 8.5 m below surface level via natural fissures in the till.

A DPE test in the deep well (8 m below ground level) showed a water flow of approx. 0.003 l/s (0.25 m³/day), however, it was not possible to measure the air flow. This is presumably due to the location of the filter (8 m below ground level), approx. 6 m below groundwater level. Furthermore, it was possible to establish a significant effect on the water level at all 3 levels. In a depth of 8 m, several of the filters are emptied and the groundwater level thus lowered with approx. 6 m. It is assessed that the sand layer found in 8.5 m’s depth has a great impact on the pumping results, especially results from the depth of 8 m.

Finally, a modelling of the fracture spreading was carried out applying the non-commercial fracture model, CIRFRX. The American consultant carried out the work, and the documentation is enclosed in appendix F. Based on a comparison of the data for the fracture spreading, the model is assessed to be a useful tool to predict and assess the initial fracture spreading. However, at the end of the fracturing, the model showed a difference of more than 100% between the observed and calculated values for upplift. Thus the design for a remedial project with hydraulic fracturing cannot be based on modelling, and it is recommended that a test fracture be carried out.