Assessing the synergies between textile sludge and low-density polyethylene (LDPE) Co-pyrolysis: Experimental insights and machine learning

Abstract

Co-pyrolysis leverages complementary properties of diverse feedstock to improve conversion efficiency, offering a sustainable route for integrated waste management and energy production. This study investigates reaction kinetics, synergistic interactions, product analysis, and machine learning prediction for co-pyrolyzing textile sludge (TS) with low-density polyethylene (LDPE) at mass ratios of 25%TS:75%LDPE, 50%TS:50%LDPE, and 75%TS:25%LDPE. Thermogravimetric analysis was performed from room temperature to 1000 °C at 2.5, 5, 7.5, and 10 °C/min. Model-free methods like linear differential Friedman, linear integral Kissinger-Akahira-Sunose (KAS), and Ozawa-Flynn-Wall (OFW) were used to evaluate kinetic parameters. Average E_a (Friedman) was 262.61 kJ/mol for 25%TS:75%LDPE, 178.34 kJ/mol for 50%TS:50%LDPE and 267.80 kJ/mol for 75%TS:25%LDPE. Furthermore, positive synergistic interactions were most significant between 450–600 °C with the dominant peaks at 500 °C for all blended samples. Fixed-bed co-pyrolysis of the optimum blend (50%TS:50%LDPE) at 500 °C produced 22% pyro-oil, 34% biochar, and 44% gaseous products. Moreover, obtained pyro-oil comprises of hydrocarbons and long-chain aliphatic derivatives, N-containing heterocycles and amines, carboxylic and phenolic acids, ethers and acetal, phthalates, and the carbohydrate derivatives. Furthermore, machine learning models like Artificial neural networks (ANNs), Classification & regression trees (C&RT), and support vector machine (SVM) were developed to predict E_a. ANN performed best for 50%TS:50%LDPE and 75%TS:25%LDPE (R² = 0.999 and 0.997) while C&RT excelled for 25%TS:75%LDPE (R² = 0.988). Findings demonstrate pronounced synergistic interactions and integrated kinetic and machine learning prediction strategy for converting diverse waste into energy.