One of the peculiarities of working within a digital ecosystem is that it’s an environment highly focused on enhanced certainty – higher accuracy, improved timeliness, better decisions, for example – but it can sometimes also involve increased levels of uncertainty; after all, it’s a complex environment in which new and different technologies interact throughout an organization at various levels and in dynamic and possibly confusing ways. Nevertheless, some major mining companies are studying the possibility that a reasonable amount of uncertainty might lead to better results in production planning and value-chain optimization.
The most recent example is the product of a strategic partnership involving Montreal-based KPI Mining Solutions, McGill University’s COSMO Stochastic Mine Planning Laboratory, and a global consortium of mining companies composed of AngloGold Ashanti, Anglo American De Beers, BHP, IAMGOLD, Kinross Gold, Newmont, and Vale.
Matheus Faria, product manager/senior mine planning consultant at KPI Mining Solutions, explained to E&MJ that the recently introduced KPI-COSMO Stochastic Optimizer is part of a new generation of software solutions that leverage powerful AI methods to optimize mining operations by accounting simultaneously for multiple stochastic (randomly variable) scenarios. This differs from mostly currently available solutions that require the use of different algorithms or software tools to optimize the mining complex/mineral value chain.
KPI-COSMO will enable mining companies to create a digital model of their mineral value chains, representing the various flows that different material types go through from the mines, passing through many processing facilities, up to the generated products. In each part of the chain, stochastic scenarios can be inputted, from the simulated block models in the mines to the metal prices at the end of the chains. This detailed optimization model enables users to capitalize all synergies in the mining complex.
Faria explained that the major inputs for KPI-COSMO Stochastic Mining Optimizer are the conditionally simulated block models (orebody models). Different geological attributes can be simulated and integrated within the solution. A set of simulations allows the modeling and quantification of uncertainty and variability of the metal content of multiple metals, material types, weathered facies, geometallurgical and other properties. Currently, major geology software vendors already provide efficient geostatistical simulation algorithms to generate the simulated block models, and KPI Mining Solutions is a partner of geostatics software developer Geovariances.
Faria advised that using geostatistical simulation requires a shift in thinking by mineral resources modeling/geology teams: Mining companies conventionally generate only estimated block models such as Kriged models. However, these models only produce a smooth representation of mineral deposits and consequently misrepresent the proportions of high, medium and low grades within a deposit, and they ignore the unavoidable uncertainties in these models. As a result, the mine planning process generates production plans that can be less profitable, and also riskier.
A certain level of orientation is needed to effectively use KPI Mining’s stochastic mine planning solutions, according to Faria. First, stochastic mine planning principles such as the differences between the estimated vs. simulated model of mineral deposits, risk assessment/profiling and risk management need to be understood and these concepts are not usually covered in mining engineering academic programs. Secondly, the simultaneous optimization framework and the functionality of KPI-COSMO Stochastic Optimizer also require training, and KPI Mining has consulting and technical support teams with experienced mining planning consultants to help customers extract maximum value from its solutions.
KPI-COSMO will be offered primarily on an annual subscription basis with integrated software support and maintenance. Shorter-term rentals are also possible to fit into clients’ needs, said Faria, particularly for mining consulting firms that use the software solutions for specific projects.