Machine learning based metal recovery from the waste printed circuit boards of mobile phones for circular economy and sustainable environment

Abstract

The metal recovery from the waste printed circuit boards (WPCBs) of mobile phones presents reduced reliance on natural resources, savings in energy and emissions discharge complemented with economic incentives as well. Herein, we present a machine learning (ML) based model for Cu, Ni and Pb recovery from the WPCBs of mobile phones using low temperature roasting process. The ML model for the metal recovery is built using Artificial Neural Network (ANN) where three variables (roasting temperature, roasting time, NH4Cl dose) are used as inputs for the model. The model-based significance order of the input variables is obtained by computing partial derivatives of outputs with respect to the inputs. Temperature is the most significant input variable for Cu while time is the most significant variable for Ni and Pb recovery, contributing the percentage significance of 42%, 41.3% and 48.4% respectively. The proposed ML based analysis framework can be deployed for enhancing the insights about the metal recovery process contributing to the smart operation and operational excellence.