Optimization of the SAG Grinding Process Using Statistical Analysis and Machine Learning: A Case Study of the Chilean...

Area of Science:

• * Mining Engineering
• * Mineral Processing
• * Data Science

Background:

• * Rising production costs and the need for resource optimization are critical challenges in the copper mining industry.
• * Improving the efficiency of resource utilization in mining operations is a key strategic objective.
• * Semi-autogenous grinding (SAG) mills are crucial in mineral processing, but optimizing their performance is complex.

Purpose of the Study:

• * To develop predictive models for SAG mills using statistical analysis and machine learning (ML).
• * To investigate methods for improving productive indicators, specifically production rates and energy consumption.
• * To assess the potential of ML techniques in enhancing efficiency in mineral processing.

Main Methods:

• * Application of statistical analysis techniques.
• * Utilization of machine learning algorithms including regression, decision trees, and artificial neural networks.
• * Simulation of digital models to analyze process variables and their impact on production and energy.

Main Results:

• * Simulated digital models indicated a 4.42% increase in production as a function of mineral fragmentation.
• * Decreasing mill rotational speed showed potential for increased production and a 7.62% decrease in energy consumption.
• * Machine learning demonstrated effectiveness in modeling complex SAG grinding processes.

Conclusions:

• * Machine learning offers significant potential for increasing efficiency in mineral processing operations.
• * Implementing ML can lead to improvements in both production indicators and energy consumption.
• * Integrating ML into broader strategies like Mine to Mill could further enhance industrial-scale performance.