Remediation by steam injection

Research output: Book/ReportPh.D. thesis – Annual report year: 2003Research

Documents

View graph of relations

Chemical spills have created a large number of contaminated sites where toxic
compounds are present as nonaqueous phase liquids (NAPL). These sites have proved difficult to remediate and aggressive technologies are needed. Remediation by steam injection is such a technology and it may be the optimal technology at heavily contaminated sites. Steam injection involves the injection of steam into the subsurface and simultaneous recovery of fluids from extraction wells. The injected steam heats the soil and creates a steam zone that expands from the injection wells as more steam is injected. The physical-chemical properties of typical contaminants change with temperature and this makes them easier to extract. In particular the increase in vapor pressure is important. Additionally, two immiscible liquid phases will boil when the sum of their vapor pressures equals the surrounding pressure and this phenomenon proves very important for the contaminant mass transfer. The objectives of this study have been to further our understanding of the dominant processes and mechanisms involved in steam injection through the use of numerical modeling, laboratory experiments and field-scale data. The focus has been on the following four issues that have been reported separately:

1. Removal of NAPLs from the unsaturated zone using steam: prevention of
downward migration by injecting mixtures of steam and air

Steam injection for remediation of porous media contaminated by nonaqueous phase liquids has been shown to be a potentially efficient technology. There is, however, concern that the technique may lead to downward migration of separate phase contaminant. In this work a modification of the steam injection technology is presented where a mixture of steam and air was injected. In two-dimensional experiments with unsaturated porous medium contaminated with non-aqueous phase liquids it was demonstrated how injection of pure steam lead to severe downward migration. Similar experiments where steam and air were injected simultaneously resulted in practically no downward migration and still rapid cleanup was achieved. The processes responsible for the prevention of downward migration when injecting steam-air mixtures were analyzed using a non-isothermal multiphase flow and transport model. Hereby, three mechanisms were identified and it was demonstrated how the effectiveness of these mechanisms depended on the air to steam mixing ratio.

2. Remediation of NAPL below the water table by steam induced heat conduction
Previous experimental studies have shown that NAPL will be removed when it is contacted by steam. However, in full-scale operations steam may not contact the NAPL directly and this is the situation addressed in this study. A two dimensional
intermediate scale sand box experiment was performed where an organic contaminant was emplaced below the water table at the interface between a coarse and a fine sand layer. Steam was injected above the water table and after an initial heating period the contaminant was recovered at the outlet. The experiment was successfully modeled using the numerical code T2VOC and the dominant removal mechanism was identified to be heat conduction induced boiling of the separate phase contaminant. Subsequent numerical modeling showed that this mechanism was insensitive to the porous medium properties and that it could be evaluated by considering only one-dimensional heat conduction.

3. On spurious water flow during numerical simulation of steam injection into
water saturated soil

Numerical simulation of steam injection into a water saturated porous medium may be hindered by unphysical behavior causing the model to slow down. We show how spurious water flow may arise on the boundary between a steam zone and a saturated zone, giving rise to dramatic pressure drops. This is caused by the discretization of the temperature gradient coupled with the direct relation between pressure and temperature in the steam zone. The problem may be a severe limitation to numerical modeling. A solution is presented where the spurious water flow is blocked and this widely enhances the performance of the model. This new method is applied to a previously reported example exhibiting numerical problems. Furthermore, it is applied to the simulation of 2-D sand box experiments where LNAPL is remediated from a smearing zone by steam injection. These experiments would have been difficult to analyze numerically without the adjustment to prevent spurious flow. The simulation proves to be very sensitive to the type of relative permeability model. The LNAPL is removed by a combination of vaporization and flow. Based on the numerical results it is argued that it is necessary for the steam to contact the NAPL directly to achieve clean-up at field-scale.

4. Three-dimensional numerical modeling of steam override observed at a fullscale remediation of an unconfined aquifer
Steam injected below the water table tends to move upwards because of buoyancy. This limits the horizontal steam zone development, which determines the optimal spacing between injection wells. In this study, observations indicating steam override at a full-scale remediation of an unconfined aquifer are analyzed by numerical modeling using the code T2VOC. A simplified 3-D numerical model is set up, which qualitatively shows the same mechanisms as observed at the site. By means of the model it is found that it will be possible to achieve a larger horizontal extent of the steam zone in a layered geology compared to the homogeneous case. In the homogeneous case the steam injection rate increases dramatically when the injection pressure is increased, which is necessary to achieve a larger horizontal development. The development of the steam zone under unconfined conditions is found to be a complex function of the geologic layering, the ground water table at steady-state extraction and the injection/extraction system. Because of this complexity it will be difficult to predict steam behavior without 3-D numerical modeling.
Original languageEnglish
Place of PublicationKgs. Lyngby
PublisherEnvironment & Resources DTU. Technical University of Denmark
Number of pages44
ISBN (Print)87-89220-74-9
Publication statusPublished - 2003

Download statistics

No data available

ID: 5774913