Improvement of saltrock creep laws requires microscopic mechanisms incorporation, adequate tests, appropriate equation forms, and better use of field data. A low-stress pressure solution mechanism, affected by grain size, may be important for field modeling, leading to a stress exponent of 1.0 insitu for saltrocks with sufficient moisture content.
The requirement that nuclear repository behavior be predictable with "reasonable assurance" (U.S. Nuclear Regulatory Commission, 1987) prior to licensing dictates that constitutive models must portray reai behavior more realistically than conventionally required. Whereas this applies to all rock types encountered, it is particularly demanding for salt because of its creep behavior. This paper discusses the current state of saltrock mechanics, identifies issues in behavior prediction, and targets research needs. Behavioral knowledge gained from repository studies can obviously be applied to engineering projects such as solution and conventional mines.
A constitutive relationship for saltrock that accounts for deformation as a function of stress, temperature and time is necessary. Isothermal deformation rate predictions, especially in the transient phase, remain a primary difficulty. In a repository, heat flux results in thermal stresses as well as stresses due to opening geometry and virgin stress fields. Furthermore, heat generation rates decay, thermal gradients are dissipated by conduction, and viscous creep behavior is strongly temperature dependent. Openings will be backfilled before abandonment, and closure will relithify the backfill, re-establishing primitive stresses as the thermal anomaly dissipates. Relithification time is short compared to heat dissipation time, especially if backfill salt is moist and stresses high. Macroscopic creep is manifestation of microprocesses: dislocation lide, defect migration, or dissolved mass transport within crystals fluid inclusions) or along grain boundaries (liquid films). Hence crystal structure, grain size, mineralogy and moisture content must be considered in a constitutive statement. Fully empirical creep models fail in predictions partly because these factors are ignored. Also, laboratory parameters may not faithfully describe insitu behavior, a central concern in rock mechanics. The major issue for saltrocks is not discontinuity behavior, but the ratio of crystal size to specimen volume, laboratory strain rates used, specimen damage, and changes in moisture content during sampling and preparation. Saltrocks behave largely as intact continua.
Until the 1960's (more recently at some mines) pillar and opening design in saltrock was based on conventional tributary area methods (Abel and Djahanguiri, 1984) with design recommendations based on uniaxial compressive strength. This fails to account for creep, confinement effect, pillar slenderness, and the yielded zone in the pillar. Pillar dimensions can also be based on uniaxial laboratory model pillar creep tests which can simulate confinement and slenderness. However, room/pillar interaction, crystal size, moisture state, and specimen disturbance are not adequately handled. Design can be based on behavioral laws from triaxial creep tests on 100 mm specimens at low strain rates (10-9-10-10 s-l), integrated with data from instrumented test adits (Mraz and Dus seault, 1986; Dusseault etal., 1987).