Reinforcement learning for eco-efficient comminution circuit characterization

Comminution, the process of crushing and grinding rock to a reduced size as part of the process of liberating valuable minerals, has been reported to be responsible for up to 4% of the world energy consumption [Environment and Climate Change Canada, 2017]. Designing for reduced carbon footprint and energy consumption is possible. Several study cases have reported significant energy savings with innovative circuits. The goal of this project is to develop an AI-based system to support the design of mineral processing circuits by learning to efficiently characterize a circuit. More specifically, given an existing circuit and a given ore, our goal is to learn the optimal sequence of tests in order to characterize the circuit as quickly as possible, while minimizing the cost. This project is part of a greater initiative (IntelliCrush, for the Crush It! Challenge ) in which we want to learn to ask the right questions and perform the right tests throughout the steps of designing a mineral process.

Faculty Supervisor:

Audrey Durand;Jocelyn Bouchard

Student:

Partner:

COREM

Discipline:

Computer science

Sector:

Mining

University:

Université Laval

Program: