Use of MIMO models in spreadsheets for the control of a column flotation circuit

The objective was to determine operating protocols for a lead and silver co lumn flotation cleaner being considered for installation at Sunshine Precio us Metals Inc.'s Kellogg mill in Northern Idaho. In this column, tetrahedri te and galena were floated, and pyrite and other gangue minerals, including siderite and silica, were depressed. The procedure was to develop a dynami c simulation using a linear transfer function approach. The transfer functi on related cause-and-effect variables such as the influence of collector an d depressant dosages on recovery and grades. Bench-scale flotation tests fo llowed by pilot column testing provided the data used to develop a linear t ransfer function. The transfer function was derived using a spreadsheet. Th is model was applied to simulate how the flotation column could be made to respond to disturbances in feed grade. The results were that xanthate, zinc sulfate and sodium cyanide were found to be suitable controls for grade an d recovery. Simulations showed the best sequence of manipulations of these reagents, and these sequences are useful for future operator training. The conclusion was that linear transfer functions were able to simulate lead an d silver recovery, develop scenarios for operator training and provide a fo undation for future on-line control.