Mineral Processing

Development of a Multivariable Fuzzy Logic Controller for a SAG Mill
    Raglan, Quebec, Canada
    Xstrata Nickel
    Control Strategies, Controller Design, Programming, Optimization, Commissioning, Monitoring
    Optimization and Advanced Control, Industrial IT, Automation, Mining and Metals


After a series of recommendations were made to improve system performance, BBA was mandated to stabilize the production of a semi-autogenous grinding (SAG) mill through fuzzy logic. A previous fuzzy controller was used but after process changes and different operation approach, it was decided to design a new and different fuzzy logic controller.The mandate included:

  • ­    Adjusting control settings for control variables
  • ­    Designing the multivariable controller and programming display of mill-related messages
  • ­    Reviewing all rules and developing monitoring tools
  • ­    Adding the SAG mill's speed control to the fuzzy logic


  • Preliminary work prior to the first visit and on-site work:
    • Review needs and conduct modelling of the "best operator" (Design of experiment)
    • Develop the necessary modifications
    • Program a simulation to test the fuzzy logic
    • Design a new multivariable fuzzy logic controller
    • Design, implement and adjust the logic for new control rules
    • Form two groups of operators and metallurgists
    • Measure and analyze performances
    • Draft a work report
    • Adjust and finetune rules
    • Provide technical support


  • The project was conducted in close cooperation with plant staff.
  • Multidisciplinary teams were involved in carrying out the mandate.
  • The fuzzy logic controller was created using a programmable logic controller (PLC).
  • Special tools were developed to simplify modifications and monitor the controller.

  • Plant personnel is autonomous and can modify the system and program directly in the PLC.