Modelling gas-phase recovery of volatile organic compounds during in situ thermal treatment

Highlights

• Mass recovery during thermal treatment evaluated using mathematical modelling.

• Simulated cumulative mass recovery relationships consistent with field results

• Moment analysis shows importance of contaminant distribution on heating time.

• Location of dense non-aqueous phase liquid pools can extend heating time.

• Total mass not a good indicator of total required heating time.

Abstract

In situ thermal treatment (ISTT) technologies can be used to remove mass from non-aqueous phase liquid (NAPL) source zones. Ensuring the vaporization of NAPL and the capture of vapors are crucial, and numerical models are useful for understanding the processes that affect performance to help improve design and operation. In this paper, a two-dimensional model that combines a continuum approach based on finite difference for heat transfer with a macroscopic invasion percolation (macro-IP) approach for gas migration was developed to simulate thermal conductive heating (TCH) applications at the field-scale. This approach simulates heat transport and gas migration, but is different than a traditional continuum multiphase approach. Mass recovery for 60 randomly generated realizations under three degrees of heterogeneity of the permeability field were simulated. The mass recovery curves had an overall similar shape for the various permeability fields. However, a wider range of completion times was observed for domains with a higher permeability variance. Results also showed that NAPL pools that were highly saturated, deep, and away from the heaters needed more heating time to be depleted, and that total NAPL mass was not a good indicator of completion time. The completion time was positively correlated with the maximum value of the mixed spatial moment of NAPL saturation about the heaters in the lateral and vertical direction, and the NAPL pool with the highest moment could increase the heating time by as much as 35%. This effect was most notable in simulations with a high permeability variance and suggests the potential to reduce heating time by locating the largest NAPL pools and placing TCH heaters accordingly.

Introduction

In situ thermal treatment (ISTT) technologies have the potential to remove volatile and semi-volatile organic compounds (VOCs and SVOCs) from the subsurface. These technologies, including electrical resistance heating (ERH) and thermal conductive heating (TCH), rely on increasing the temperature of the subsurface to near the water boiling point or higher to achieve mass removal. It has been reported that ISTT technologies can be more effective than fluid injection-based remediation technologies to treat heterogeneous geology (Kingston et al., 2014; McGuire et al., 2006; Stroo et al., 2012) because their performance is mainly governed by electrical and thermal conductivity, which are typically less spatially variable than hydraulic conductivity.

Predicting the duration required for ISTT applications to remove non-aqueous phase liquid (NAPL) at field sites is challenging. Not only is the rate and extent of NAPL removal affected by the delivery and transport of energy (heat) in the subsurface, it is also affected by the rate at which gas (steam and vaporized VOCs) is produced and captured. For the application of ISTT at sites impacted by volatile NAPLs, over 90% of VOC mass from a treatment zone is removed in the gas phase (Heron et al., 2005, Heron et al., 2013; Vermeulen and McGee, 2000), and the majority of that mass is recovered during the co-boiling stage (Burghardt and Kueper, 2008; Zhao et al., 2014). Co-boiling (or steam distillation) plays an important role in removing VOCs and SVOCs as it allows the production of gas at a temperature lower than the boiling point of either water or the NAPL (Nilsson et al., 2011; Tzovolou et al., 2011). Therefore, although a common criterion for terminating heating is reaching a specified temperature and maintaining it for a defined period (Parker et al., 2017), uncertainty associated with NAPL distribution as well as variations in temperature due to heat transport, water boiling and co-boiling, complicate the relationship between measured temperatures and the achievement of treatment goals.

Numerical modelling can be a useful tool for investigating ISTT applications, and accurate ISTT models must account for heat transport, phase change and gas migration to simulate NAPL removal by co-boiling. Although several numerical studies have been conducted to reproduce aspects of ISTT experiments (Chen et al., 2012; Krol et al., 2011b; Krol et al., 2011a; Molnar et al., 2019; O'Carroll and Sleep, 2007; Xie et al., 2019), few studies have focused on the numerical simulation of ISTT in NAPL source zones at the field scale. Numerical models that make use of macroscopic invasion percolation (macro-IP) (Glass and Yarrington, 2003; Ioannidis et al., 1996; Kueper and McWhorter, 1992; Mumford et al., 2015; Trevisan et al., 2017) including those used to study ISTT (Krol et al., 2011a; Molnar et al., 2019) are particularly useful for studies that make use of multiple realizations of NAPL and permeability distributions to account for unknown site conditions because of their computational efficiency.

In this study, a two-dimensional model that combines a continuum approach based on finite difference for heat transfer with a macro-IP approach for gas migration was developed to simulate TCH applications at the field scale. The two-dimensional (2D) model incorporates conductive heat transfer, latent heat of NAPL and water, gas production and gas migration. Taking advantage of the computational efficiency of macro-IP, the objective of this study is to investigate the relationships between permeability heterogeneity, NAPL architecture and mass recovery by simulating key gas recovery mechanisms.