Optimizing Bioleaching for Printed Circuit Board Copper Recovery: An AI-Driven RGB-Based Approach

Abstract

Featured Application: This AI-driven optimization research focuses on bioleaching for Cu recovery from e-waste, integrating RGB sensor data and machine learning to enable real-time monitoring and predictive maintenance. The IIoT-based approach improves recovery rates, lowers environmental harm, and promotes sustainable resource management across different operational scales in e-waste recycling plants. Recovering copper from end-of-life electronics, especially from printed circuit boards, provides significant economic benefits, reduces environmental impact, and supports a circular economy. This case study presents a data-driven approach to predicting copper recovery in the electrolysis stage of a bioleaching process by utilizing RGB sensor readings. We tested nine regression models using RGB values from experimental data. The gradient boosting model, optimized via response surface methodology (RSM), outperformed the others, with predictions matching 84% of observed patterns. These results demonstrate strong predictive capabilities, with scope for further accuracy enhancements. We offer an open-source, web-based digital twin designed specifically to monitor the bioleaching plant, enabling real-time and historical data analysis to support predictive maintenance. Our results underscore the potential to optimize the entire bioleaching process, marking a significant advancement for large-scale copper recovery. This study is the first to investigate predictive bioleaching continuous processes in a semi-industrial e-waste plant using RGB sensors, presenting a novel approach in the field.