Modeling Trichloro-Benzene Desorption at the Eastern Woolen Mills Superfund Site

Myron I. Kuhlman, MK Tech Solutions, Inc., 12843 Covey Lane, Houston, TX 77099, Tel: 281-564-8851, Fax: 281-564-8821

Ian T. Osgerby, US Army Corp of Engineers, New England District, 696 Virginia Road, Concord, MA 01742-2751, Tel: 978-318-8631, Fax: 978-318-8663

Hot-air vapor extraction has been used to desorb both volatile and semi-volatile chemicals from subsurface or excavated soil. At the Corinna, Maine Superfund Site, 1,2,4 trichlorobenzene (b.p. 213°C) and other chlorinated solvents in river sediment were to be desorbed.  However, dry, hot-air injection had not achieved the remediation goals after two weeks of heating because the vapor pressures of the chemicals were reduced by adsorption on the dried soil.  This phenomenon is the result of the less polar chemical being adsorbed to replace water that had been hydrogen-bonded to the surface.  For instance, if the soil contains only enough moisture to cover 5% of the surface, the chemical's vapor pressure is less than 0.5% of that expected for a moisture content corresponding to 95% of a monolayer.  Moisture equivalent to 95% of a monolayer is about 5% soil moisture.  Therefore, the process had to either be operated with a controlled humidity or at a very high temperature.  

STARS, a well proven oil-field thermal simulator developed by CMG, Ltd of Calgary, Alberta was used to model and optimize the Corinna desorption process.  A three-component model of the adsorbed SVOC was used.  In this model, the SVOC could have a vapor pressure as high as its textbook value, or it could be reduced by as much as a factor of several million.  The vapor pressure reduction is dependent on the water saturation in the soil pile. 

The model was used to help select day-to-day operating conditions in the early stages of the project and showed the soil could reach the desired cleanup criteria if it was held above 140°F without drying out for one week.    If the soil was not kept moist, it would have had to be heated to over 320°F to be cleaned in seven days.  The thermal insulating properties of dry soil and condensation of vaporized materials in cooler portions of the pile make decontamination with hot dry air problematical.  Simulations like this example are invaluable in designing and optimizing remediation projects.

Application of Flow and Transport Optimization Codes to Groundwater Pump and Treat Systems 

Kathleen Yager, U.S. EPA, Technology Innovation Office, 11 Technology Drive (ECA/OEME), North Chelmsford, MA 01863, Tel: 617-918-8362, Fax: 617-918-8427
Dave Becker, U.S. Army Corps of Engineers, Hazardous, Toxic, and Radioactive Waste Center of Expertise, 12565 W. Center Road, Omaha, NE 68144-3869, Tel: 402-697-2655, Fax: 402-697-2613
Karla Harre, U.S. Navy, Naval Facilities Engineering Service Center, 1100 23rd Avenue, Port Hueneme, CA  93043-4370, Tel: 805-982-2636, Fax: 805-982-4304

With the recent focus on lowering the operating costs of environmental remediation systems, including groundwater pump and treat systems, there has been increased interest in automated approaches to optimization.  This project evaluates the benefits and utility of contaminant transport modeling optimization algorithms, operable on desktop computers, against traditional  (“trial and error”) modeling approaches.  The transport optimization approach requires a well-calibrated groundwater flow and contaminant transport model.  Three pump-and-treat facilities with varying site characteristics were selected for inclusion in the project.  Three separate optimization formulations (each formulation considers a different cleanup/pumping strategy) were developed for each facility, and solved by three modeling teams (two using optimization algorithms and one applying trial and error).  The results clearly indicate that mathematical optimization methods are able to identify solutions better than those obtained using trial-and-error approaches. The solutions found were 5% to 50% better (measured using optimal objective function values), with an average improvement of about 20%.  As the complexity of the site increased, more cost savings were obtained.  Potential savings are substantially greater than the incremental cost of applying mathematical transport optimization over a more traditional trial-and-error approach, estimated to range from zero to $40,000.  The challenges in applying optimization algorithms increased with the complexity of the site, with the greatest challenge being the computational requirements of the optimization algorithms.  If a single optimization run were set up to solve the entire problem as formulated, many thousands of simulation runs would be required and the computational times would be prohibitive. Instead, the teams employed sequential solution approaches, in which some parts of the problem were fixed while others were optimized. The teams also considered fixing one parameter, such as flow, and optimizing the locations, then using these locations, optimizing the flow rates.  These approaches require substantial expertise and effort to use effectively.

Top