Quantitative Analysis Of The Accuracy Of Steady State Simulation Of Mineral Processing Plant

Steady-state simulation is now in current use to assist mineral processing plant design and optimization. However, the reliability of results generated by simulators has not until now been seriously investigated. A simulator supervision algorithm was designed and programmed to study both the influence of the reliability of experimental data on the accuracy of unit operation model calibration, and the influence of unit operation model accuracy on the results of direct plant simulations. An application to a grinding flotation plant is described and discussed. The approach may open the way to the quantitative analysis of the risks taken in mineral processing projects, which would be similar to the application of geostatistics to mining projects.