Ronald W. Falta

Experimental method for characterizing CVOC removal from fractured clays during boiling

Conventional remediation methods that rely on contact with contaminants can be ineffective in fractured media, but thermal methods of remediation involving CVOC stripping at boiling temperature show promise. However, limited experimental data are available to characterize thermal remediation because of challenges associated with high temperature. This research reports an experimental method using uniformly contaminated clay packed into two types of experimental cells, a rigid-wall stainless steel tube and a flexible-wall Teflon tube in a pressurized chamber. Both tubes are 5 cm in diameter and approximately 25 cm long. This laboratory apparatus was developed as a 1D physical model for contaminant transport in a cylindrical matrix towards a fracture, which is represented by one end of the cylinder and serves as the outlet of vapor and contaminant. The clay was contaminated with dissolved 1,2-dichloroethane (DCA) and bromide, and the columns were heated to more than 100 °C and then the top end was depressurized to atmospheric pressure to induce boiling. The outflow was condensed and analyzed for contaminant mass. The flexible-wall cell was confined to 100 kPa (gage), allowing equilibrium boiling temperatures of approximately 120 °C to be maintained. The clay was sampled before and after heating and extracted to determine the DCA distribution along the length of the column. During a typical test in the rigid-wall cell, internal temperatures and pressures along the column during heating reached the saturated vapor pressure curve. DCA concentrations in the recovered condensate were up to 12 times of the initial pore concentration in the clay. Less than 5% of non-volatile bromide was recovered. Significant removal of DCA and water occurred along the entire length of the clay column. This suggests that boiling was occurring in the clay matrix.

Evaluation of liquid aerosol transport through porous media.

Application of remediation methods in contaminated vadose zones has been hindered by an inability to effectively distribute liquid- or solid-phase amendments. Injection as aerosols in a carrier gas could be a viable method for achieving useful distributions of amendments in unsaturated materials. The objectives of this work were to characterize radial transport of aerosols in unsaturated porous media, and to develop capabilities for predicting results of aerosol injection scenarios at the field-scale. Transport processes were investigated by conducting lab-scale injection experiments with radial flow geometry, and predictive capabilities were obtained by developing and validating a numerical model for simulating coupled aerosol transport, deposition, and multi-phase flow in porous media. Soybean oil was transported more than 2m through sand by injecting it as micron-scale aerosol droplets. Oil saturation in the sand increased with time to a maximum of 0.25, and decreased with radial distance in the experiments. The numerical analysis predicted the distribution of oil saturation with only minor calibration. The results indicated that evolution of oil saturation was controlled by aerosol deposition and subsequent flow of the liquid oil, and simulation requires including these two coupled processes. The calibrated model was used to evaluate field applications. The results suggest that amendments can be delivered to the vadose zone as aerosols, and that gas injection rate and aerosol particle size will be important controls on the process.

A semi-analytical method for simulating matrix diffusion in numerical transport models

A semi-analytical approximation for transient matrix diffusion is developed for use in numerical contaminant transport simulators. This method is an adaptation and extension of the heat conduction method of Vinsome and Westerveld (1980) used to simulate heat losses during thermally enhanced oil recovery. The semi-analytical method is used in place of discretization of the low permeability materials, and it represents the concentration profile in the low permeability materials with a fitting function that is adjusted in each element at each time-step. The resulting matrix diffusion fluxes are added to the numerical model as linear concentration-dependent source/sink terms. Since only the high permeability zones need to be discretized, the numerical formulation is extremely efficient compared to traditional approaches that require discretization of both the high and low permeability zones. The semi-analytical method compares favorably with the analytical solution for transient one-dimensional diffusion with first order decay, with a two-layer aquifer/aquitard solution, with the solution for transport in a fracture with matrix diffusion and decay, and with a fully numerical solution for transport in a thin sand zone bounded by clay with variable decay rates.

Numerical Analysis of Thermal Remediation in 3D Field-Scale Fractured Geologic Media.

Thermal methods are promising for remediating fractured geologic media contaminated with volatile organic compounds, and the success of this process depends on the coupled heat transfer, multiphase flow, and thermodynamics. This study analyzed field-scale removal of trichloroethylene (TCE) and heat transfer behavior in boiling fractured geologic media using the multiple interacting continua method. This method can resolve local gradients in the matrix and is less computationally demanding than alternative methods like discrete fracture-matrix models. A 2D axisymmetric model was used to simulate a single element of symmetry in a repeated pattern of extraction wells inside a large heated zone and evaluate effects of parameter sensitivity on contaminant recovery. The results showed that the removal of TCE increased with matrix permeability, and the removal rate was more sensitive to matrix permeability than any other parameter. Increasing fracture density promoted TCE removal, especially when the matrix permeability was low (e.g., <10(-17) m(2)). A 3D model was used to simulate an entire treatment zone and the surrounding groundwater in fractured material, with the interaction between them being considered. Boiling was initiated in the center of the upper part of the heated region and expanded toward the boundaries. This boiling process resulted in a large increase in the TCE removal rate and spread of TCE to the vadose zone and the peripheries of the heated zone. The incorporation of extraction wells helped control the contaminant from migrating to far regions. After 22u2009d, more than 99.3% of TCE mass was recovered in the simulation.

Modeling Plume Responses To Source Treatment

Source zone remediation at a site may be undertaken for a number of reasons. The plume response to source remediation is a complex function of many variables, including the fraction of mass remaining in the source compared to the plume, source concentration compared to regulatory limits, ratio of plume decay rates to groundwater velocity, relationship between source mass removal and source discharge, and local diffusive effects in the plume. As discussed in this chapter, models are tools that allow practitioners to quantify plume response to source remediation. This insight, gained from modeling, can then allow the practitioner to establish reasonable goals and expectations for what can be achieved through source remediation.

MODELING OF FLOW AND TRANSPORT INDUCED BY PRODUCTION OF HYDROFRACTURE-STIMULATED GAS WELLS NEAR THE RULISON NUCLEAR TEST

The Project Rulison test in Western Colorado was conducted in 1969 to determine if a nuclear device could be used to fracture low permeability, gas-bearing rock to enhance natural gas production. The presence of radionuclides (specifically tritium) in the produced gas persisted above acceptable levels, and the test was abandoned. Advances in hydraulic fracturing technology have made it feasible to extract natural gas from low-permeability reservoirs, and drilling activity near the Rulison site has raised concerns that remnant radioactivity in the detonation zone could migrate to nearby producing wells and enter the natural gas distribution system. The site was modeled using the massively parallel version of TOUGH2, TOUGH2_MP-EOS7R, with over 1,000,000 elements. The majority of the tritium is present as tritiated water, and since the gas permeability of the native formation is several orders of magnitude higher than the liquid permeability, any significant migration occurs with the gas phase. In the model, tritium partitions between the aqueous and vapor phases in relation to the Henry’s Law constant. The Henry’s constant for tritiated water vapor is simply the water vapor pressure. The model domain includes the entire targeted gas-bearing section and gas production wells installed in 2010 1.2xa0km (0.75xa0mile) from the Rulison site. The model was calibrated to historical production and pressure data from the Rulison reentry well and to data from the recently installed gas wells. The model was used to simulate the effects of current wells and of future wells that could potentially be installed nearer the Rulison site.