Numerical modelling is a powerful tool that has become common in engineering use in recent years. With advancements in hardware and software, complex models can be developed to simulate numerous problems related to the mining industry, including seepage flows, solute transport, and deformations. Numerical modelling is often required for the planning, operation and decommissioning of minesites. However, even though the models used will provide solutions with high precision, it is important to recognize that the systems being modelled are probabilistic. The inputs to the model can not be established with certainty. This uncertainty in the inputs generates uncertainty in the model outputs, regardless of how sophisticated the numerical model is. In many cases, a probabilistic treatment of the problem, using a less sophisticated numerical model, yields cost effective, timely and insightful results.
As an example, a probabilistic methodology was used to quantify the potential for off-site brine migration from the tailings management facility of a potash mine. Conventional practice would have utilized a deterministic analysis with a (complex) seepage/solute transport model; however, after investigating the problem further, a simple (l-dimensional) solute transport model was adopted and implemented within a discrete probabilistic simulation framework. This methodology permitted the estimation of the potential for off-site brine migration.
The methodology is outlined in the paper within the context of a specific application conducted at the Agrium Inc. potash mine located near Vanscoy, Saskatchewan. It is simple and applicable to many problems requiring numerical simulation in mining and mine decommissioning.