Grinding mill process optimization algorithm

The efficiency of the grinding mill process plays a significant role in the overall performance of a mine circuit. Most existing control systems only consider measurements of the quality of the product when it is output from the grinding process leading to delayed responses. The reaction to deviations from quality set points may be calculated with large input error. The proposed solution uses a set of implicit correlation functions defined for pre-specified Fragmentation Size Distribution curves of the input ore feed. This approach is used to determine the future behavior of the grinding process without introducing unnecessary delays into the control loop. The fragmentation size may be obtained via an additional sensor system based on the video image of the material transported to the mill by the conveyor belt (line scan camera, laser profiler, line 3D camera). The dedicated tool - Grinding Mill Optimizer - was developed and used to perform an analysis of a dynamic model of the grinding process. Data obtained from an experimental system is used to validate the model.