Program A

Increasing certainty on the mill feed for predictable and controllable performance

In this program, we use sensors to enable machines to identify and monitor ore characteristics at several points of the upstream mining chain, from in-place resources to the mill feed. The aim is to achieve feed control and blending in stages to provide a stable, predictable and controllable feed for the plant.

Benefits

  • Monitoring the ore attributes in the Run-of-Mine (RoM) ore can provide rapid feedback to mining operations for dilution control and reconciliation in short time periods.
  • Having a greater certainty on mill feed attributes will allow the value to be optimised in shorter time periods than is currently possible.

Research challenges

  • Resource knowledge updated in real time with sensor information
  • Near real time ore tracking and tagging from mine to mill
  • Near real time mineralogy identification in the mill feed. 

Partners

BHP logo
Maptek logo
Eka logo
Veracio
Bureau Veritas
Scantech logo
MZ Minerals logo

Research projects


Stage project A1: belt sensing and ore sorting

These project are part of Research Group 1 - Geostatistical Modelling and Value Chain Optimisation.


Stage project A2: ore tracking

These project are part of Research Group 3 - Optimisation Group


Stage project A3: ore tagging and fingerprinting

These project are part of research group 2 - Mineralogy Group

  • Research project 4-1 (A RP4-1): Correlations of elemental, mineralogical, hyperspectral with sensor data for mineral identification

    This project has concluded.


    Industry challenge: Uncertainty on the mill feed, ore tracking and fingerprinting.

    This project is linked to Translation Project 3 (A TP3).

    Project scope:

    1. Validate the concept the fingerprinting at lab scale simulations
    2. Use fingerprinting concept on resource data and mil feed for multi-elements
    3. Simulate potential for ore sorting for attributes of value and cost.
    • Student: TBC
    • Principal supervisors:  Prof Nigel Cook
    • Postdoctoral fellow: TBC
    • Research leads: University of Adelaide
    • Translation partners: Bureau Veritas, Boart Longyear, Scantech, Consilium Technology
  • Research project 4-2 (A RP4-2): Geometallurgical modelling: Predicting metal recovery accounting for non-additivity and preferential sampling designs

    This project has been completed.


    Industry challenge: Developing a new algorithm for predicting metal recovery accounting for non-additivity and preferential sampling designs.

    This project is linked to Translation Project 3 (A RP1).

    Project scope:

    1. Develop a new algorithm for predicting metal recovery accounting for non-additivity and preferential sampling designs.
    2. Run various tests of the algorithm using Prominent Hill data.
    • Principal supervisors:  Prof Peter Dowd
    • Postdoctoral fellow: Dr Amir Adeli
    • Research leads: University of Adelaide
    • Translation partners: OZM

Stage project A4: sensor information and sorting

 

  • Research project 6 (A RP6): Blend strategy optimisation

    This project has been completed.


    Yue Xie

    Yue Xie

    Industry challenge: There is a need to optimise batches of concentrate over the mine plan by linking data and decisions.

    Project scope:

    1. Understand existing methods of copper concentrate blend strategy optimisation
    2. Translate mill models into computer language
    3. Improve objective functions and existing algorithms.
    • Student: Yue Xie
    • Principal supervisors: Prof Frank Neumann
    • Postdoctoral fellow: Dr Aneta Neumann
    • Research lead: University of Adelaide

    This project is part of Research Group 3 - Optimisation Group
    This project is linked to Translation Project 4 (A TP4).

    View the YouTube video of Yue Xie’s presentation on ‘Blend strategy optimisation’

  • Research project 7 (A RP7): Mill feed sorting to optimise throughput and energy usage

    This project has been completed.


    Hu Wang

    Hu Wang

    Industry partner challenge: Hard rock types negatively affect mill throughput. Early identification of rock mineralogy that is hard to process in the SAG mill feed is required.

    Project scope:

    1. Visual and machine learning on gyratory crusher and SAG mill feed to identify mineralogy type
    2. Machine learning to tune mill parameters to optimise throughput and energy usage.
    • Student: Hu Wang
    • Principal supervisors: Prof Chunhua Shen, A/Prof Max Zanin
    • Postdoctoral fellow: Dr Junjie Zhang
    • Research lead: University of Adelaide

    This project is part of Research Group 4 - Machine Learning Group
    This project is linked to Translation Project 6 (A TP6).


Translation projects