Presentation Information
[T9-O-7]Applicability of the Single-porosity Models for Describing the In-situ Tracer Test on the Fractured Diatomaceous Mudstone in the Horonobe Underground Research Laboratory
*Koraphat Fairat1, Eiichi Ishii1, Yusuke Ozaki1, Yuta Fukatsu1, Hirokazu Ohno1, Mochizuki Akihito1, Chika Kuribayashi1, Masaki Takeda2,1 (1. Horonobe Underground Research Center, Japan Atomic Energy Agency, 2. Nuclear Fuel Cycle Engineering Laboratory, Japan Atomic Energy Agency)
Keywords:
mudstone,parameter estimation,solute transport,tracer test,PFLOTRAN
1. Introduction
Reliable solute transport models are critical for geological disposal performance assessments. Modelling solute transport in poorly connected rock fractures where both fractures and the rock matrix serve as active pathways remains challenging. Although the conventional single-porosity model conceptually simplifies these systems into a homogeneous continuum, its applicability to poorly connected networks is not well understood. Specifically, clear guidance is lacking on whether these models should incorporate the matrix porosity or a reduced effective porosity to accurately capture transport behavior. To address these knowledge gaps, this study investigated whether a single-porosity model can effectively evaluate the solute transport of a non-sorbing tracer on such rock masses and whether the estimated porosity is consistent with the rock matrix porosity.
2. Methodology
Two dipole tracer tests using uranine (50 mg/L) were executed across shear fractures in the Neogene diatomaceous mudstone of the Koetoi Formation at the Horonobe Underground Research Laboratory (URL). The experiments generated breakthrough curves (BTCs) by utilizing a shared, single injection point paired with two separate pumping points located at different distances. The shorter-scale configuration (TT-2024-Uranine) featured a 0.8 m dipole length (L) with a 10-minute injection at 10 mL/min and a 5 mL/min pumping rate. The longer-scale configuration (TT-2025-Uranine) extended to a 1.9 m dipole length, sustaining a 1,380-minute injection at 20 mL/min against a reduced pumping rate of 0.5 mL/min. Numerical interpretation of these BTCs was conducted via 3D single-porosity advection-dispersion equation (ADE) models in PFLOTRAN. Two separate model simulations were run to compare the resulting parameter estimates and BTCs: an optimized effective porosity model and a 60%-fixed matrix porosity model. By coupling each model run with the LMFIT optimization package, inverse analysis was performed to independently quantify effective porosity (φ), longitudinal dispersivity (aL), and horizontal transverse dispersivity (aT) for both tests.
3. Result and Discussion
The ADE solute transport parameter estimation for the TT-2024-Uranine and TT-2025-Uranine BTC demonstrated that the effective porosity model successfully reproduced the measured data with a very high R-squared value, indicating its viability for capturing the solute transport behavior, whereas the matrix porosity model failed to do so (Fig. 1). Achieving this fit required estimated longitudinal dispersivity values (aL = 13-15 m) that significantly exceeds the established empirical relationship (which predicts 10% of the 0.8 and 1.9 m dipole lengths). This inflation likely points to the influence of matrix diffusion. Despite a 2.3-fold increase in dipole length, a 138-fold longer pulse duration, and execution across distinct locations, the dispersive parameters remained remarkably stable, with longitudinal and horizontal transverse dispersivity (aT = 0.13–0.15 m) shifting only marginally. The estimated effective porosities for TT-2024-Uranine and TT-2025-Uranine were 12% and 0.7%, respectively. This variation likely stemmed from the distinct locations targeted by their respective borehole intervals. While both values fell significantly below the Koetoi Formation’s total matrix porosity (φ = 60%), they remained orders of magnitude higher than its fracture porosity (φ = 0.01%). This indicates that active solute transport was constrained to a minor subset of preferential, well-connected pathways rather than utilizing the bulk matrix.
4. Conclusions
This study demonstrates that a classical single-porosity ADE framework can accurately simulate tracer transport in poorly connected media where both fractures and the rock matrix serve as active pathways. It therefore offers a reliable, computationally practical tool for evaluating transport in these complex networks. However, relying solely on bulk matrix porosity fails to account for the true active transport volume. Ultimately, these insights contribute to a deeper understanding of solute transport processes and support the development of predictive transport models.
Acknowledgments
The study was supported by the OECD-Nuclear Energy Agency Horonobe International Project (HIP).
References
Ishii et al., 2011; Gelhar et al., 1992; Miyakawa et al., 2024
Reliable solute transport models are critical for geological disposal performance assessments. Modelling solute transport in poorly connected rock fractures where both fractures and the rock matrix serve as active pathways remains challenging. Although the conventional single-porosity model conceptually simplifies these systems into a homogeneous continuum, its applicability to poorly connected networks is not well understood. Specifically, clear guidance is lacking on whether these models should incorporate the matrix porosity or a reduced effective porosity to accurately capture transport behavior. To address these knowledge gaps, this study investigated whether a single-porosity model can effectively evaluate the solute transport of a non-sorbing tracer on such rock masses and whether the estimated porosity is consistent with the rock matrix porosity.
2. Methodology
Two dipole tracer tests using uranine (50 mg/L) were executed across shear fractures in the Neogene diatomaceous mudstone of the Koetoi Formation at the Horonobe Underground Research Laboratory (URL). The experiments generated breakthrough curves (BTCs) by utilizing a shared, single injection point paired with two separate pumping points located at different distances. The shorter-scale configuration (TT-2024-Uranine) featured a 0.8 m dipole length (L) with a 10-minute injection at 10 mL/min and a 5 mL/min pumping rate. The longer-scale configuration (TT-2025-Uranine) extended to a 1.9 m dipole length, sustaining a 1,380-minute injection at 20 mL/min against a reduced pumping rate of 0.5 mL/min. Numerical interpretation of these BTCs was conducted via 3D single-porosity advection-dispersion equation (ADE) models in PFLOTRAN. Two separate model simulations were run to compare the resulting parameter estimates and BTCs: an optimized effective porosity model and a 60%-fixed matrix porosity model. By coupling each model run with the LMFIT optimization package, inverse analysis was performed to independently quantify effective porosity (φ), longitudinal dispersivity (aL), and horizontal transverse dispersivity (aT) for both tests.
3. Result and Discussion
The ADE solute transport parameter estimation for the TT-2024-Uranine and TT-2025-Uranine BTC demonstrated that the effective porosity model successfully reproduced the measured data with a very high R-squared value, indicating its viability for capturing the solute transport behavior, whereas the matrix porosity model failed to do so (Fig. 1). Achieving this fit required estimated longitudinal dispersivity values (aL = 13-15 m) that significantly exceeds the established empirical relationship (which predicts 10% of the 0.8 and 1.9 m dipole lengths). This inflation likely points to the influence of matrix diffusion. Despite a 2.3-fold increase in dipole length, a 138-fold longer pulse duration, and execution across distinct locations, the dispersive parameters remained remarkably stable, with longitudinal and horizontal transverse dispersivity (aT = 0.13–0.15 m) shifting only marginally. The estimated effective porosities for TT-2024-Uranine and TT-2025-Uranine were 12% and 0.7%, respectively. This variation likely stemmed from the distinct locations targeted by their respective borehole intervals. While both values fell significantly below the Koetoi Formation’s total matrix porosity (φ = 60%), they remained orders of magnitude higher than its fracture porosity (φ = 0.01%). This indicates that active solute transport was constrained to a minor subset of preferential, well-connected pathways rather than utilizing the bulk matrix.
4. Conclusions
This study demonstrates that a classical single-porosity ADE framework can accurately simulate tracer transport in poorly connected media where both fractures and the rock matrix serve as active pathways. It therefore offers a reliable, computationally practical tool for evaluating transport in these complex networks. However, relying solely on bulk matrix porosity fails to account for the true active transport volume. Ultimately, these insights contribute to a deeper understanding of solute transport processes and support the development of predictive transport models.
Acknowledgments
The study was supported by the OECD-Nuclear Energy Agency Horonobe International Project (HIP).
References
Ishii et al., 2011; Gelhar et al., 1992; Miyakawa et al., 2024
Comment
To browse or post comments, you must log in.Log in