Abstract:
This study employed machine learning (ML) techniques to identify the applicability of biochar as a catalyst for the Fenton-like process using PMS as an oxidant. Acetaminophen (ACT) was selected as the contaminant to perform the experimental study. Different biochars were used to catalyze peroxymonosulfate (PMS) for experimental data generation. The biochars were produced by post-pyrolysis thermal treatment at different detention times. Then, experiments using different materials, a single catalyst dose and PMS concentration were employed for ACT degradation. The 24 h heat-treated biochar (24BC) had the highest ACT degradation efficiency. Accordingly, different experimental conditions were investigated, including different doses and PMS concentrations. Further, the influence of ionic strength was investigated for the best ACT degradation conditions using different ions individually and combined. ACT degradation was found to be enhanced by the presence of ions. The analysis of chemical oxygen demand showed that despite complete ACT degradation being achieved, by-products generated remain in solution, suggesting incomplete mineralization. Finally, various statistical and ML models, including Random Forest, Linear Regression, KNN, Ridge, Lasso Regression, Support Vector Machine, Decision Trees, and Adaptive Neuro-Fuzzy Inference System were applied to predict and analyze the degradation efficiency of ACT using Biochar/PMS processes and to identify the ML technique most appropriate for the given experimental conditions. This study presents a preliminary investigation which aimed to assess the feasibility of machine learning techniques in analyzing biochar-mediated ACT degradation. While the findings are promising, further research with larger datasets is necessary to confirm and to generalize the conclusions.