From Stabilization to Optimization; Process Control in the Flotation Process at Xstrata Nickel’s Strathcona Mill

"The complexity of a modern flotation circuit makes it susceptible to disturbances such as changes in feed, ore mineralogy, equipment degradation, poor measurements and faulty instruments. Good measurements and a robust control system are critical to create and maintain a stable and optimum flotation process.This paper focuses on the development and implementation of process control applications in the flotation circuits at Xstrata Nickel’s Strathcona Mill ( referred to as the mill hereafter ) in Sudbury, Ontario. Following the path of ‘stabilize-improve-optimize’, basic controls to stabilize process variables such as flotation cell pulp level, air and reagent addition rates, cleaning-column feed and wash-water rates have been established. Factors such as best practice in pulp level and valve position measurement, feedforward control and the application of an appropriate control philosophy, have contributed to the significant reduction in process variability. For instance, standard deviation reduction of 79% for the Ni impurity in the column Cu concentrate grade was experienced and is shown. The variability reduction of the flotation circuits’ key process-variables paves the way to further improve and optimize flotation performance. A grams-per-metal unit reagent addition control scheme has been improved upon, particularly for different operating scenarios and better robustness. In addition, following significant stabilization of the primary controls and columns feedrates, Cu/Ni separation cleaning-column bias-control (cascaded to the wash-water rate) and hold-up control (cascaded to air flowrate) have been developed and tested. This strategy – from stabilization to optimization - is expected to achieve a more consistent control of final concentrate grades and improve recoveries.INTRODUCTIONFlotation is considered the most important and value-added process in the milling metallurgical plant and is still undergoing continuous evolution to treat greater tonnage and different ores (Wills, 2006). Starting over three decades ago, with the advent of on-line slurry content analysis devices, flotation automatic control has received persistent interest and a large number of applications have been reported (Thwaites, 1983, 1986 ; Mckee, 1991). Bergh and Yianatos (2011) reviewed different aspects of process control in flotation circuits and have evaluated the applicability of model predictive control to the process. Mckee (1991) analyzed the difficulty in developing lasting/robust process control applications for flotation circuits and stated that the path of stabilization-refinement-optimization would more likely lead to successful flotation circuit process control. Best practice on column process control was discussed by Bouchard, etc. where a multi-layered approach was proposed, provided that the measurement of column process variables ( i.e. bias, froth depth, hold-up, etc. ) are measured reliably (Bouchard & Thwaites, 2008). Wills (2006) stated that the best practice for flotation automatic control starts with fulfilling basic control objectives, such as stabilized flotation cell level, air flow, pH, reagent flow etc., before considering more advanced, or even optimal, control of metallurgical targets, which are more closely linked to economic performance."