A novel hydrochar production from corn stover and sewage sludge: Synergistic co-hydrothermal carbonization understandings through machine learning and modelling

Abstract

The corn stover (CS) and sewage sludge (SS) were co-hydrothermal carbonized with three different mixing ratios were designed; 1/1, 1/2 and 1/3 calculated on CS/SS basis; temperature ranges (180, 200, 220 and 240) ⁰C and three different residence time of (1, 2 and 3) hour were selected by the current study. To understand the chemical characteristics synergistic influences of hydrochar attributes, van Krevelen diagram, principal component analysis, Pearson co-relation matrix, feature engineering and dep machine learning models were employed. To understand the combustion behavior of the hydrochar; thermogravimetric analysis was performed. The Tr-34; (mixing ratio 1/3, 240 °C and residence time of 1 h) was proved as the optimum with 23.21 MJ/kg of higher heating value, 76.25 % hydrochar yield and fuel ratio of 0.44. The FTIR spectrum of the same treatment had also confirmed the abundance of energy containing functional groups. Among feature engineering, fixed carbon was proved as the most important parameter with 76 % influences, governing the energy contents of the hydrochar. The linear, polynomial and ridge regression machine learning models had provided excellent fitness for the commercial hydrochar prediction with R² of 0.99.