Data from research on the prediction of recovery rate of metals from spent Li-ion batteries using machine learning methods

Experimental methods (including hydrometallurgical methods) for recovering metals from spent lithium-ion batteries are complex, expensive, still inefficient, and generate secondary waste. Therefore, machine learning methods are used to optimize recovery methods while effectively managing waste.In our research on optimizing metal recovery using the hydrometallurgical method with sulfuric acid, we utilized an experimental dataset obtained in this project and literature data to determine the optimal leaching process parameters and maximize the metal recovery rate, mainly for lithium and cobalt. We employed the following machine learning algorithms: Random Forest, Gradient Boosting, Linear Regression, SVR (Support Vector Regression), MLP (MultiLayerPerception), KNN (k-nearest neighbors), and Decision Tree. The input data (training data) used were: sulfuric acid concentration, additive and reducer (H2O2) concentration, temperature, time, S/L ratio (mass of battery powder/volume of leaching bath), and metal recovery rate in percentage.