Study of Parameters Influencing LNAPL Extraction Using Cofra System

No Thumbnail Available

Date

2008-11-17

Journal Title

Series/Report No.

Journal ISSN

Volume Title

Publisher

Abstract

This research consisted of a large scale laboratory experimentation program to evaluate the performance of a prototype, vacuum extraction, PVW system and numerical modeling of the experimentation. A large-scale (1.52 m x 1.52 m x 1.37 m) fine sand specimen was prepared and instrumented with a prototype PVW system and piezometers in the laboratory to investigate the fluid extraction phenomenon for a single PVW subject to a transient water table. Two key properties of the PVW system were varied during the study: the active length and the placement depth of the PVW within the subsurface. A transient water table was the result of the PVW fluid extraction with a no flow boundary around the specimen, which would be typical of field conditions for a single PVW in the middle of a network of PVWs. Experimental testing was divided into two phases. The first phase investigated the active lengths and placement depths of two PVWs for water extraction. A light Nonaqueous Phase Liquid (LNAPL) was placed in the sand sample during the second phase of testing, and the PVW active length and placement depth was again investigated for the extraction of B100 biodiesel, used as a model LNAPL. Two scenarios were conducted in each phase of testing and were described as Setups 1 through 4. After each set of tests were completed for both setups in Phase II of the experimentation, the distribution of the residual LNAPL within the testing domain was analyzed for mass balance of the LNAPL within the system. The fluid extraction rates and hydraulic gradients were determined in each scenario and ranged from 1.2x10-3 m/s to 4.8x10-3 m/s, with the higher fluid extraction rates and higher gradients occurring in the scenarios with the greater depth of placement that remained below the water table elevation for longer periods (Setups 2 and 3). The greater LNAPL extraction was observed by the scenario with the shallowest placement depth and the shortest active length (Setup 4). The distribution analysis of the LNAPL within the system after testing was completed for Setups 3 and 4 indicated the shorter active length and shallow placement depth utilized in Setup 4 resulted in less smearing of the LNAPL within the sand, hence less volume of residual LNAPL. Mass balance of the LNAPL in Phase II also suggests that Setup 3, with greater water and air flow rates through the system, provided improved conditions for biodegradation and dissolution of the B100 biodiesel. Numerical modeling was performed to simulate the results obtained in the laboratory testing in order to evaluate system performance, to better understand the extraction process of LNAPL using a single PVW, and to develop a strategy for field implementation. Two modeling programs were used to evaluate PVW system performance: i) SEEP/w was used to model the groundwater flow and water table depths, and ii) Bioslurp was used to model the LNAPL extraction and residual distribution. SEEP/w accurately modeled the water table drop with the rate of drop varying from 0.127 m (0.4 ft) per hour to 0.365 m (1.2 ft) per hour. The LNAPL transport modeling provided valuable prediction of LNAPL distribution and was compared with laboratory residual distribution results. The Bioslurp model yielded a more uniform distribution of residual LNAPL with specific volume ranges from 0.0184 m (0.72†) to 0.0349 m (1.37†), while the experimental results yielded specific volumes ranging from 0.0043 m (0.17†) to 0.0263 m (1.04†).

Description

Keywords

biodiesel, LNAPL, prefabricated vertical drain, vacuum extraction

Citation

Degree

MS

Discipline

Civil Engineering

Collections