A meshing technique is detailed which is suitable for large scale finite element modeling of solution mines in salt and potash. Two related criteria impacted the development of this technique. First is the selection of the reduced-integration, stabilized, 8-node brick finite element. The material point density of a mesh composed of these bricks is much lower than a mesh of equivalent nodes composed of tetrahedra. This becomes meaningful if the constitutive model used is representative of salt and/or potash as much of the total computational effort is spent in the constitutive evaluation routines. Second, the mesh should facilitate a rapid refinement/coarsening. This should be accomplished distant from the solution mines themselves as poorly-shaped elements are involved. The resulting technique produces a completely compatible mesh, that is, no constraints are required to deal with ‘hanging’ nodes.

1. Introduction

The type of problem under consideration is one or more solution mines, generally of an unstructured geometry, situated in either homogeneous salt domes (gulf states in the USA) or in layered media composed of salt or potash layers (Alberta and Saskatchewan, Canada). Below is an example graphic of the latter.

Identified in Fig. 1. are two solution mines (caverns) situated in layered media. The mine on the right is observed to have a roof-fall, thereby extending beyond the salt layer into the non-salt layer immediately above. The mines are generally a kilometer or more below the ground level. The larger mine has an approximate diameter of 160meters, a volume of 9.4 x 105meters3 and a surface area of 6.1x 104meters2 This geometry will be used to demonstrate the meshing technique described below. Note also in Fig. 1. that the computational domain is chosen to be an octagon in plan. Although we employ bricks, this poses no difficulty for the technique.

In practice, many more mines and correspondingly larger computational domains are encountered. Since calculations are routinely performed, simulating decades of time, computational effort becomes intractable unless certain devices are employed to reduce this effort.

This content is only available via PDF.
You can access this article if you purchase or spend a download.