Rough joints can be over-closed, and remain over-closed by a previous application of a higher normal stress. This is an exaggerated form of hysteresis. Rough joints in igneous and metamorphic rocks can over-close even due to temperature increase alone, due to better fit, which is something beyond hysteresis. The rock mass deformation moduli, thermal expansion coefficients, hydraulic apertures, and seismic velocitiesmay each be affected. Well-controlled laboratoryHTMtests, in situ HTM block tests, and large-scale heated rock mass tests, lasting several years at Stripa, Climax andYucca Mountain, have produced evidence for this extra fully-coupled response. Over-closed laboratory direct shear tests give elevated strength envelopes in the case of tension fractures and joint replicas. Heating alone also increases the shear strength of natural joints. The coupled thermal-OC effect in HTM numerical modeling will require, as a minimum, thermal expansion coefficients that include rather than exclude relevant joint sets, if these have marked roughness and if they originated at elevated temperature. Subsequently elevated deformation moduli that attract higher stress must be expected.
Hydro-thermo-mechanical HTM modelling of high level nuclear waste disposal scenarios has been actively sought in the last 30 years. In simplified form, the HTM (and chemical) effects of excavation, heating and cooling (with eventual seismic loading from major earthquakes in the very long term), have each to be simulated. The effects of heating and cooling on rock joints likely to exist in the ‘geological containment’ will be the focus of this paper. A phenomenon revealed almost 40 years ago, that has proved to have relevance for both HTM field experiments and HTMmodelling, concerns over-closure of joints. Under ambient conditions we may refer simply to hysteresis effects, but when heat is added, thermal over-closure appears to accentuate closure effects in the rock mass. This sounds ‘positive’ for waste isolation: in fact it may be adverse, due to the subsequent cooling that requires shrinkage in a rock mass that may have over-closed rough joint sets that remain closed despite cooling. Difficulties in obtaining excavation-induced failure of artificial rock slope models, each consisting of 40,000 blocks, reported in Barton, 1971 and 1972, has proved to have an unexpected link to the above concerns. Steep, gravity- and horizontally-stressed slopes with adversely-dipping sets of tension fractures ‘would not fail’, in relation to slope stability calculations based on strengths obtained from conventional 1:1 direct shear tests. When loading to 4 or 8 times higher normal stress, prior to unloading and shearing, successively steeper shear strength envelopes were obtained, as illustrated in Figure 1. The excessively stable slopes (Figure 2) were actually caused by over-closure of the rough tension fractures.As observed sometimes in real slope failures, therewas evidence in slope-failure debris, of ‘over-closed’masses of blocks, which might be interpreted as ‘discontinuous jointing’ or evidence of ‘cohesive strength’ in field observations.
