Discover the latest team updates and company news
This Friday, immerse yourself in a 2-hour technical workshop with our seasoned experts. Dive deep into a hands-on example of constructing a geometallurgical prediction workflow. We'll lay down a comprehensive framework, explore varied application instances, and focus on a specific example for a step-by-step demonstration. The session wraps up with a look into future challenges and a discussion.
Join the APMT Python Workshop and enhance your Python skills in data manipulation, 3D dataset visualization, and abstract mining problem representation. This comprehensive six-hour workshop will take place on two days (3 hours/day), Sept 7 and Sept 8, from 9:00 am to 12:00 pm.
Localized uniform conditioning (LUC) allows the estimation of grades at SMU support when available data is sparse and insufficient for an adequate estimation by conventional geostatistical linear interpolation. Widely used for resource modeling on mineralization presenting a heavily skewed distribution. Expert knowledge should guide the hyperparameter of the method to mitigate spatial artifacts between panel contacts. Checkout our LinkedIn page for a full access to the document.
Services Mining Consulting Our team provides our clients with experience and expertise in medium and large mining projects. We focus on: Orebody modeling, Mine planning, Geometallurgical analysis, Data science in mining. Checkout our LinkedIn page for a full access to the document.
Annapurna is our service for the automatic development of geological models, resource estimation, and economic assessment of mining projects, including exploratory data analysis, compositing, geological domain definition, contact analysis, resource classification, reserve estimation, and mine scheduling optimization, among many other processes. Future case studies will be released. Applied and validated on real deposits ranging from massive porphyry Fe-Cu deposits to narrow Ni-Cu-(PGE) deposits.
Multivariate Geostatistical Simulation and Deep Q-Learning to Optimize Mining Decisions: (i) Multivariate geostatistical modeling provides the means to account for uncertainty in the spatial and statistical distribution of multielement ore deposits. (ii) The resulting block model, the geometric constraints, the mining-processing-metallurgical parameters, and the economic context represent the environment in which a neural net-based agent is trained via deep Q-learning to find the optimal sequence of extraction in terms of total Net Present Value. (iii) The resulting optimal policy maximized the total NPV conditioned to the uncertainty in grade distribution