The conducted research was part of the international research project DECOVALEX 2019 (DEvelopment of COupled models and their VALidation against EXperiments). DECOVALEX 2019 consisted of seven different research projects called Task A to G. SSM were involved in Task G, which were composed of three different teams from Technical University of Liberec (TUL), Seoul National University (SNU) and geomecon GmbH (GMC), and focused on the excavation damage zone (EDZ) and it’s coupling to water flow.
The formation of EDZ and the associated change of rock mass hydraulic conductivity has been a continuous matter of research, not exclusive of, but prominently in former DECOVALEX tasks (Liu et al., 2011; Wang et al., 2011; Rutqvist et al., 2009a; Rutqvist et al., 2009b; Min et al., 2004; Min et al., 2009; Öhman et al., 2005; and many more). Within these studies, many aspects have been covered and among them the changes of boundary conditions affecting the formation and evolution of the EDZ have been analysed. However, the impact that EDZ development have on a repository for radioactive waste, and in particular for spent nuclear fuel, have not been fully analysed and understood. Parameters, such as hydro geological conditions and rock stresses are important parameters for understanding the evolution of EDZ and have potential effects on the repository performance.
The conducted research is primarily concerned with the transmissivity evolution throughout the lifetime of a repository for a spent fuel in a sparsely fractured and competent crystalline rock mass. The conducted work includes a review of measurement methods for understanding the status and evolution of a repository for spent nuclear fuel (Appendix 2).
The research project consists of an examination of different codes ability to simulate interference test data from the TAS04 tunnel in Äspö (Sweden). The fractured rock mass is represented by a distributed fracture network model (DFN), which was defined by Clearwater Hardrock Consulting. The model benchmarking is part of the work package 2 (WP 2) and will be described in detail within this report. It builds on a previous work package (WP 1; Appendix 3), which dealt with a comparison of different software packages for this purpose. Three different teams, Seoul National University (SNU), Technical University of Liberec (TUL), and geomecon GmbH (GMC), tried to model the interference test with two commercially available software packages, 3DEC and COMSOL Multiphysics®. SNU using 3DEC modelled the flow between the injector and monitoring wells by applying a stress dependent fracture flow, while TUL and GMC using COMSOL Multiphysics® allowed the fluid to flow through fractures and the matrix.
While different approaches have been applied, it can be summarized that the codes were not able to predict all the in-situ pressure responses, particularly those with significant pressure differences in different observation holes. Therefore, the numerical models may not be suitable for modeling the evolution of the rock in the vicinity of tunnels and holes. Possible reasons for this are that: a) not all provided data could be implemented; b) the provided data might be incomplete with reference to channels along the fractures (channeling); c) the differences between the regular and irregular tunnel geometry induce stresses that differ from the in-situ stresses. The numerical codes did however represent the inflow into the tunnel quite well. This might suggest that the DFN representation of the fractured rock mass is more suitable at a larger scale than smaller scale.
The development of EDZ around deposition tunnels and deposition holes has been a target of attention in the development of radioactive waste disposal concepts in crystalline rocks. The reason for aiming at understanding the formation of an EDZ is not primarily related to rock stability issues. More important for the long-term safety of the repository is the characteristics of the EDZ that changes the hydraulic properties of the rock mass in the vicinity of the excavations. Due to these changes, there is a potential for an increase in fluid-conductive pathways around the disposal canister and buffer. These induced pathways may act as transport channels for radionuclides away from the repository. As the EDZ may enhance the hydraulic conductivity between the excavations and the fractures in the rock mass, this could also impact the radionuclide transport from the spent fuel locations to the biosphere, and the resaturation timescales for the bentonite buffer.
The understanding of the transmissivity evolution after closure of a repository for spent nuclear fuel relies to a large extent on numerical models. Therefore, continuous testing of different models against field data is of great importance. Not only to be able to better understand the post-closure transmissivity evolution of a repository for spent nuclear fuel, but also for a better understanding of what limitations different models have in this regard and of the challenges when implementing field data in numerical models. Thus, this study is important for SSM:s continued review of SKB:s safety analysis reports for the spent nuclear fuel repository in Forsmark.
Need for further research
To what extent a continuation of the performed research, dealing with post-closure processes like the thermal phase, glaciation, and earthquakes, would yield reliable results is questionable. However, the results from such work could indicate relative changes, e.g. whether the transmissivity in the excavation damage zone (EDZ) is going to increase or decrease during the post-closure processes. In this regard, they could yield valuable insights and enhance the understanding of the evolution of the repository after closure.
Another way forward is to focus more on channelized fractured flow, since a local DFN model appears to be unable to represent the complexity of fractured flow on this detailed scale. Which is relevant when considering near-field modelling of flow and transport close to the depositions holes for spend nuclear fuel.