Laboratory studies of partially saturated infiltration dynamics in analog fractured porous rock systems - under consideration of fracture-matrix interaction and network geometry
Cumulative thesis
Date of Examination:2023-10-10
Date of issue:2024-08-09
Advisor:Dr. Jannes Kordilla
Referee:Dr. John R. Nimmo
Referee:Prof. Dr. Thomas Ptak-Fix
Files in this item
Name:Dissertation_2024_FlorianRuediger.pdf
Size:97.0Mb
Format:PDF
Abstract
English
Fractured porous rocks are abundant groundwater reservoir formations often covered by a thick, unsaturated zone. The interplay between fractures and rock matrix can significantly impede infiltration, particularly when the matrix has high diffusivity. Despite extensive research on preferential fracture flow and retarding matrix imbibition, a unified theory remains elusive. This Ph.D. thesis employs laboratory infiltration experiments to assess (1) the controls of fracture-matrix interaction, which governs unsaturated infiltration, and (2) the role of fracture intersection dynamics in controlling the infiltration process and its transition toward more saturated conditions. Infiltration through an intersected vertical fracture of varying lengths exhibits substantial variability below a defined Representative Elementary Volume (REV) size, where intersection dynamics introduce erratic flow. In contrast, employed systems above the REV establish a more stable flow related to the number of consecutive intersections. Applying a dual-domain model with a transfer function demonstrates the challenge of accurately estimating the fracture-matrix interface area during the partially saturated fracture flow. Nevertheless, averaging calibrated parameters from experiments with an appropriate REV size recovers discharge dynamics within systems of sufficient size and highlights the scale-dependent nature of the infiltration process. However, this approach lumped the effects of (A) fracture flow modes and (B) intersection dynamics on wetting front propagation into the domain coupling parameter. Infiltration in non-intersected vertical fractures further delineates fracture-matrix interaction and implications for fracture flow configurations and resulting wetting front propagation. Most striking, complex flow dynamics involving discontinuous slugs and film flow resemble, on average, plug flow infiltration. An analytical solution recovers wetting front propagations if applied flow rates exceed a matrix-diffusivity controlled threshold. Flow rates below only require a simplified solution but underestimate the initial infiltration depth controlled by the inflow boundary and the timeframe for matrix storage saturation due to the partially saturated flow geometry. Building upon these findings that imbibition dominates unsaturated fracture-matrix systems without intersection dynamics (above REV size), fracture network infiltration experiments in glass analogs aim to comprehend their control over infiltration and discharge in saturated portions of the system. Applying a preliminary model utilizing local rules to constrain partitioning behavior at intersections only replicates breakthrough times if the wetting front infiltrates the same number of horizontal fractures as in the experiment. Notably, bypass mechanisms gained prominence, and lateral water redistribution induced transient discharge phases, particularly pronounced for low flow rates. Implementing these mechanisms, introducing time-scale divisions for water transfer (slugs, films), and integrating a (porous matrix) storage component will refine our understanding of the interdependences between (1) intersection dynamics, (2) fracture-matrix interactions, and (3) flow modes, in shaping infiltration processes.
Keywords: Preferential flow; Fractured porous media; Fracture flow; Matrix imbibtion; Fracture-matrix interaction; Laboratory experiments; (Semi-)analytical modeling; Transfer functions; Dual-domain models; Fracture intersection dynamics; Fracture networks; Infiltration dynamics; Unsatured flow