With the financial support from CEMI (Center for Excellence in Mining Innovation) and its sponsors, an alternative approach called the Point Estimate Method (PEM), a computing efficient probabilistic approximation, has been implemented in Phase2. Uncertainties and variability are the rule when dealing with data from the natural environment such as in geomechanics. Thus, a reliable design approach must be able to consider uncertainties; evaluate the probability of occurrence of a given scenario and take measures that reduce the risk to an acceptable level. In order to assess the effect of uncertainty, one needs probabilistic tools. However, many probabilistic methods (e.g. Monte-Carlo Simulations) are difficult to introduce in numerical code due to limitations in computing capacities. Now we can look forward to a broader use of the probabilistic approach in the mining industry as well as a better assessment of the reliability level of the design of underground openings.
In order to develop a reliable design approach, one must use statistical methods to deal with the variability of the input parameters. However tools usually used in geomechanics, like stress analyses (e.g. Finite Element Analyses, FEM), are in essence deterministic (a single set of input parameters leads to a single answer). Also, these tools are often computing time intensive and are not well-suited for the multiple runs needed for systematic sensitivity analyses or statistical simulations (e.g. Monte-Carlo). That's the reason why the Center for Excellence in Mining Innovation (CEMI) recently contracted RocScience Inc. to introduce an alternate method, the Rosenblueth point-estimate method (PEM, Rosenblueth, 1975), a simple, computing efficient probabilistic method, into their FEM software Phase2 version 8.0. This paper presents the approach and discusses its applicability.