Formatted Title
The Challenges with Amendment Delivery in Fractured Porous Rocks: Using DFM Simulations to Forecast Remediation Efficacy
Background/Objectives
A serious gap exists in our current practice of realistic evaluation of corrective measures studies with respect to the challenges of the delivery of amendments to contaminant mass in fractured porous rock source zones: the variable length and aperture of fractures cause tortuous and channelized flow paths, and thus incomplete delivery of amendments; it is uncertain whether the injection well will intercept fractures that are well-connected to the entire zone of contamination; and finally, the contaminant has had decades of additional time to diffuse to regions deep within the rock matrix blocks indicating contrasting time scales necessary to quantify and judge “success.” Diffusion within the rock matrix is a key, limiting process, as the time scale differences of rapid flow in the fracture network versus diffusive migration into (or out of) matrix blocks are stark. Forward-diffusion of the amendment from the fractures into the matrix, as well as back-diffusion of contaminant mass from the matrix towards fractures must be simultaneously considered to appropriately account for the rate of destruction and/or changes to the volume of contamination within the impacted zone.
This investigation aims to use discrete fracture-matrix (DFM) numerical models informed by high resolution characterization data to assess the volume and mass of contaminant under different scenarios of amendment injection; it aims to quantify the complexity of fracture network flow, non-linear transport, and (in future work) reactive transport scenarios. DFM numerical models are well suited for this, as they explicitly represent the geometry of the fractures and the variably sized matrix blocks between fractures. Random fracture network realizations combined with Monte Carlo methods represent the average style of fracture network (e.g., spacings of joint and bedding plane fractures) and average response to amendment addition. Simulation of various modes/intensities of delivery aid intuition and generate quantitative estimates of amendment effectiveness and can be used for cost and time-to-cleanup forecasts. Hypothetical scenarios of amendment type, injection rate, injection volume, and monitoring duration can be evaluated; active remediation can be contrasted with natural assimilative processes.
Approach/Activities
Simulation scenarios representing two modes of amendment delivered are considered: A) fractures only, assuming a particle-based amendment that is suspended in an injection fluid (e.g., carbon or zero-valent iron); and B) to the rock matrix adjacent to fractures, assuming the amendment is an aqueous solute (e.g., lactate or persulfate). For tractability reasons, transient amendment invasion is simulated with no contaminant present; the fracture network and matrix porewater architecture and concentrations of the delivered amendment are then superimposed onto a “target zone of antecedent contamination.” The interaction of the amendment to the contaminant is mimicked by manipulating boundary and/or initial conditions of the simulation (depending on the scenario and for each respective fracture network realization) and resultant contaminant volumes and mass are tracked. Future work may include use of a multiphase model to better define an antecedent contaminant distribution that would arise from DNAPL migration, a first order decay within the targeted spatial zones of Scenarios A and B, and/or a DFM reactive transport model where the contaminant and amendment would be consumed in proportion to the stoichiometry of the reaction. Additional realism could be introduced by imposing an ambient gradient, whereby mass discharge to the downgradient plume could be tracked and contrasted in cost-benefit analyses.
Results/Lessons Learned
Findings of prior numerical simulations of amendment delivery indicate that much of the injected amendment mass escapes the target zone. The remaining amendment destroys at best 20% of the contaminant mass (subject to the mass destruction methods’ limitations). Much of the volume of contamination is unaffected by the amendment and continues to increase by diffusion. When natural degradation of the contaminant is considered, it becomes the dominant process of mass removal in the long term. Thus, this provides a platform to estimate and contrast the cost of contaminant removed by engineered versus natural processes over various time periods.