The BMT 2 test was concerned with the upscaling of the THM processes in a fractured rock mass and its significance for large-scale repository performance assessment. The work was primarily concerned with the extent to which various thermo-hydro-mechanical couplings in a fractured rock mass adjacent to a repository were significant in terms of solute transport, which was typically calculated in large-scale repository performance assessments. Since the presence of even quite small fractures may control the hydraulic, mechanical and coupled hydro-mechanical properties of the rock mass, a key aspect of the work was to explore the extent to which these properties can be upscaled and represented by ‘equivalent’ continuum properties for appropriate PA calculations. Based on the above guiding understanding, the aims of the BMT2 two closely integrated components were: 1) to understand how an explicit acknowledgement of the need for upscaling of coupled processes might alter the approach to performance assessment modelling and the analysis of the model results; and 2) to understand the uncertainty and bias inherent in the outputs from performance assessment models in which the upscaling of THM parameters was either implicit or explicit. From these general aims the task was set-up as a numerical study of a realistic large-scale reference problem that was concerned with a hypothetical heat producing repository placed at depth in a hypothetical fractured rock. However, in order to obtain realistically complex and spatially varying geologic data, the hydrogeologic and hydromechanical input items were loosely based on the basement geology around Sellafield, United Kingdom, and on the site characterisation data acquired by Nirex UK Ltd (Fig. 7). However, the geometry analysed was purely fictional and did not represent the actual conditions at Sellafield. The main research performed was concerned with the far-field groundwater flow and solute transport for a situation where a heat producing repository was placed in a fractured rock medium. A vertical fracture zone cuts both rock units but lies beyond the end of the repository tunnel. The relevant data for the rock formations and fault were based on Sellafield site characterisation data acquired by Nirex. The data are in the form of statistical distributions of properties. Typically, most of the data concerns measurements on a small scale, whereas the problem to be studied mainly concerns the large scale.
The funding organizations and their research teams for BMT2 of Task 3 can be seen in Table 1.
The funding organizations and their research teams for BMT1 of Task 3 can be seen in Table 1.
Main findings from the upscaling hydro-mechanical rock properties at small scales
In BMT2, the problem definition did not provide, on purpose, a definitive fracture system geometry model, so as to encourage different approaches to be applied for characterizing the equivalent hydraulic and mechanical properties of the rock mass concerned. This naturally led to quite different interpretations for the data to be provided, problem dimensions, and how to use the existing experimental data. This kind of benchmark tests was used for the first time in the DECOVALEX project and played a significant role in deepening understanding of the issues of property upscaling in fractured rocks. The results demonstrated that it was the uspcaling of hydraulic properties that were the main sources of uncertainty in a problem of this nature. The added disturbance, in relation to in-situ stress, was small in the far-field of a deep repository. Yet, understanding the stress/permeability relation was important for understanding the nature of the permeability field.
Main findings from the solute transport simulations at large scales
Good progress was made for the development of approaches for upscaling the coupled HM processes in fractured rocks during BMT2. The upscaled properties were applied to numerical simulations for assessing their impact on the results of numerical simulations of large scale flow and transport processes, and the impact of the heat source by waste decay. Similar values of the maximum temperature were predicted by participating teams. The main scientific findings regarding the impact of coupling processes on nuclide transport were: 1) negligible effect of the hydraulic process on stress or displacement variations between TH and THM results, but significant effects of mechanical processes on the discharge vectors and the pore pressure variations; 2) relatively small changes in permeability induced by TM processes; 3) significant thermal effect on flow process; 4) THM impacts on transit time were more important at the 50 m scale than at the km scale. The fracture density assumed in the modelling, and whether upscaling was performed or not, appeared to affect the transit time to final discharge more than the THM coupling; 5) long-term heating would affect the hydraulic aperture.
| Team | Permeability (m2) | Deformation Modulus (GPa) | REV scale (m) |
| INERI
S |
5.5x10-14 –1.6x10-13 | 38 – 73 | 2 – 4 |
| JNC | 6.25x10-18 | 10 – 250 | |
| KTH | 3.5x10-16 – 3.4x10-15 | 34–44 | 5 |
| LBNL | No upscaling | 1.56; 12.5 – 50 | |
| OPG | 7.4x10-18 – 7.2x10-17 | 13 – 17 (32 – 39) | 10 – 100 |
| STUK | 8.61x10-19 | 7.5; 50 | |
| UoB | 5.0x10-17 – 5.5x10-16 | 10 | |
| UPV | No upscaling | 5 | |
| Ref. | 5.7x10-18 | 65 | 50 |
| Team | τ50 at 50 m (in years) | τ50 final in years/m | β50 final in years/m | Comment |
| JNC | 9.8·105 (3.110 13 s) | 8.8·106 (2.81014 ) | 4.1·1015 | |
| LBNL | (3.9–6.8)·104 | (1.48–2.09)·104 | (7.09–9.5)·108 | Only H (years) |
| 6.61·106 | TH | |||
| OPG | 4.85·104 | 7.85·105 | 2.60·107 | H only, no upscaling |
| 5.95·104 | 7.77·105 | 2.61·107 | HM, no upscaling | |
| 3.64·104 | 5.85·105 | 1.94·107 | TH, no upscaling | |
| 4.34·104 | 5.79·105 | 1.94·107 | THM, no upscaling | |
| 8.36·102 | 2.53·104 | 6.75·105 | HM, upscaled property | |
| 7.70·102 | 2.00·104 | 4.70·105 | THM, upscaled property | |
| 4.81·103 | 1.77·105 | 4.69·106 | HM, upscaled property (1/3 active fractures) | |
| 5.37·103 | 1.40·105 | 3.82·106 | THM, upscaled property (1/3 active fractures) | |
| STUK | 500
(1.5·1010 s) |
1.1·103
(3.3·1011s) |
2.7·1010
7.54·1017s |
Without TM |
| 566
(1.7·1010s) |
1.2·103
(3.58·1011s) |
2.8·1010
8.12·1017s |
With TM | |
| UoB | — | 120 yrs
(3.8·109 s) |
— | H base case (constant hydraulic aperture = 10 μm). |
| 3.72·105 | HM with (UCS = 39.6 MPa) | |||
| UPV | 1.5·108 | Only H, σ=1detailed | ||
| 1.7·108 | Only H, σ=1upscaled |
(β50 is taken as the travel time with Darcy velocity)
Lessons learned and outstanding issues
Despite the progress made in developing numerical approaches for HM upscaling of fractured rocks, the problem has such a degree of complexity that the current work can only be regarded as a start for understanding both the nature of the problem and the demanding computational skills and resources for further research. Some of the lessons learned and outstanding issues need to be further investigated including: 1) importance of the initial aperture of fractures for HM coupling and permeability; 2) uncertainty in the relation between hydraulic residual aperture and maximum mechanical aperture; 3) impact of the variations of the mechanical properties that could be higher than that of the HM coupling for performance measurements; 4) limitations on computational resources demanded for such analysis not available for the present project; 5) need for more reliable fracture system characterization regarding proper constraints of fracture length and densities, especially develping different scaling laws for different types of fractures and their relation with hydraulic properties; 6) need for 3D modelling; 7) Differences in using DFN or stochastic continuum approaches for upscaling.
Besides one report generated for BMT2 of Task 3, our journal papers were published on the a special issue of the International Journal of Rock Mechanics and Mining Sciences, Volume 42, Number 5-6: