Application of Machine Learning in a Mineral Leaching Process-Taking Pyrolusite Leaching as an Example

Area of Science:

• Metallurgy
• Materials Science
• Chemical Engineering

Background:

• Pyrolusite leaching is a key step in manganese extraction.
• Optimizing leaching efficiency is crucial for hydrometallurgical processes.
• Electric-field-enhanced leaching offers potential improvements.

Purpose of the Study:

• To analyze process variables in electric-field-enhanced pyrolusite leaching.
• To predict manganese leaching rates using machine learning models.
• To compare the applicability of different machine learning models in hydrometallurgy.

Main Methods:

• Utilized several machine learning models for data analysis.
• Investigated the influence of leaching conditions (time, sulfuric acid, ferrous sulfate concentrations).
• Evaluated model performance using regression index (R^2) and mean square error.

Main Results:

• Identified leaching time, sulfuric acid, and ferrous sulfate concentrations as key factors.
• Support vector regression (SVR) model achieved the highest prediction accuracy (R^2 = 0.92).
• Gradient boosting regression model also showed strong performance (R^2 > 0.85).

Conclusions:

• Machine learning models are effective for optimizing manganese leaching.
• SVR is a highly suitable model for predicting leaching rates.
• The methodology is applicable to other hydrometallurgical processes for optimization and prediction.