Salt is Not a Steady-State Creep Material

Steady-state creep is increasingly used to model structures hosted in salt and potash. This is presumably due to the absence of representative constitutive models available in commercial geomechanics software. We consider a standard suite of tests performed on salt. These include creep, dilation and tension. These tests are fit to a simple material model and evaluated. Simulations of a solution mine are performed with varying degrees of completeness with respect to the constitutive model. Volumetric closures are calculated. This exercise quantifies the ramifications on closure (deformations) according to the various material assumptions. It is observed that both the creep transient and dilation contribute to closure. Furthermore, if a failure criterion is considered, dilation is the most appropriate candidate. It is insufficient merely to post-process a creep calculation to evaluate the potential for dilation. By it's nature, dilation is a strain-softening response. It localizes and concentrates and therefore must be included in the constitutive calculation. 1. INTRODUCTION Numerous material models have been considered for describing salt and potash. Among these are MatMod (Miller, 1976), the WIPP creep law (Senseny, 1985) and more recently, the much researched and enhanced Munson-Dawson model (Reedlum, 2018). The principal shortcoming of these models is that each geological site is typically unique and has to be characterized anew, that is, there is not a universal constitutive model for salt nor for potash, as is typical for various metal alloys. The large suite of tests required to calibrate these models is impractical. These models typically have several tens of parameters. Tests must be diverse enough to span the basis represented by all of these parameters/freedoms and furthermore be consistent. For industrial operations, this becomes infeasible due to the expense of the expanded test regime and also the limited amount of core available, layered media, for example. Characteristic of the aforementioned models is the lack of attention to dilation. Salt and potash are complex materials, exhibiting a crystalline creep response much like metals, and also a dilational response (expansion as grains shear against one another) characteristic of granular materials, i.e., rock. Only the Munson-Dawson model addresses dilation but the corresponding state variable is a scalar. Anisotropy requires, at a minimum, a second-order tensor state variable.