Modelling a SAG Grinding System through multiples regressions

Due to the constant growth of the copper industry, the increase in production costs and complexity in the composition of the feed of the production processes that make up the industry, the analysis of alternatives that improve efficiency by studying the dynamics of the processes represents a significant cost reduction. Then, the generation of analytical models that represent the dynamic behaviour of production processes has the potential to contribute to generating a better understanding of the operating parameters that have a greater impact on the response (s), in addition to identifying operating restrictions and optimal levels of operation. The present work developed a digital model of the SAG milling process by generating multiple regression and quadratic regression models. The relationships between 22 operational variables with production in tons per hour were sampled, and after analysing the impact of the independent variables on the response, water feeding, sump level, percentage of solids in feeding, pebbles and hardness were maintained for fit the analytical models. The multiple linear regression model presents a good fit to the operational data (85.4%), however, the inclusion of the interactions and the quadratic effects of the variables increases the coefficient of determination (93.2%).