Application of Artificial Neural Network (ANN) as a predictive tool for the removal of pharmaceutical from wastewater streams using biochar: a multifunctional technology for environment sustainability.

Abstract

This study investigates biochar as an attractive option for removing pharmaceuticals from wastewater streams utilizing data from various literature sources and also explores the sensitivity of the characteristics and implementation of biochar. ANN 1 was designed to determine the optimal biochar characteristics (Surface Area, Pore Volume) to achieve the maximum percentage removal of pharmaceuticals in wastewater streams. ANN 2 was developed to identify the optimal biomass feedstock composition, pyrolysis conditions (temperature and time), and chemical activation (acid or base) to produce the optimal biochar from ANN 1. ANN 3 was developed to investigate the effectiveness of the biochar produced in ANN 1 and 2 in removing dye from water. Biomass feedstock with a high lignin content and high volatile matter at a high pyrolysis temperature, whether using an acid or base, achieves a high mesopore volume and high surface area. The biochar with the highest surface area and mesopore volume achieved the highest removal percentage. Regardless of hydrophobicity conditions, at low dosages (0.2), a high surface area and pore volume are required for a high percent removal. And with a higher dosage, a lower surface area and pore volume is necessary to achieve a high percent removal.