T. Chan et al., INTERNATIONAL COMPARISON OF COUPLED THERMO-HYDRO-MECHANICAL MODELS OFA MULTIPLE-FRACTURE BENCH-MARK PROBLEM - DECOVALEX PHASE-I, BENCH-MARK TEST-2, International journal of rock mechanics and mining sciences & geomechanics abstracts, 32(5), 1995, pp. 435-452
Citations number
21
Categorie Soggetti
Geosciences, Interdisciplinary","Mining & Mineral Processing
An international inter-code bench mark comparison has been carried out
under the auspices of DECOVALEX, an international cooperative project
for the development of; coupled (thermo-hydro-mechanical) models and
their validation against experiments in nuclear waste isolation, The s
ubject bench mark problem, designated BMT2, involves a system of nine
blocks of intact hard rock separated by two pairs of relatively soft f
ractures. The coupled processes of conductive, convective and dispersi
ve heat transport; nonisothermal fluid flow; and thermo-mechanical str
esses and displacements in both the intact rock and the fractures were
simulated, Five research teams representing four European and North A
merican countries participated in this bench-marking exercise using tw
o fundamentally different types of numerical models, i.e. finite-eleme
nt models and distinct-element models. All teams simulated the fractur
es as discrete features, For the coupled hydro-mechanical behaviour of
the fractures the modelling teams have used either the empirical, non
linear Barton-Bandis-Bakhtar constitutive relation or a simpler Coulom
b friction like stress-deformation relation in conjunction with Snow's
parallel-plate fracture flow model. For this bench mark problem it wa
s found that (1) heat convection significantly affects the distributio
n of temperature, thermal stresses and displacements; and (2) the pred
ominant coupled effect is fracture closure caused by thermal expansion
of the rock blocks, Comparison of the modelling results obtained by t
he various teams have indicated very good overall agreement, In partic
ular, the results calculated using finite-element and distinct-element
codes are quite similar, Practically all the discrepancies among simu
lation results can be ascribed to different idealization, constitutive
relations, or parameter values adopted by the participating research
teams, In-depth discussions among participants have enhanced our under
standing of the simulated processes and provided guidance for model im
provements.