Investigation of the co-pyrolysis of coal slime and coffee industry residue based on machine learning methods and TG-FTIR: Synergistic effect, kinetics and thermodynamic

In this study, thermogravimetry and Fourier transform infrared spectroscopy (TG-FTIR) were used to analyze the co-pyrolysis characteristics of coal slime (CS) and coffee industry residue (CIR) at different heating rates. The CS and CIR were mixed according to five mass ratios of 1:0, 7:3, 5:5, 3:7 and 0:1. Through the detection of gas emission and mass loss rate with temperature changing, the results showed that the co-pyrolysis of CS - CIR revealed a synergistic effect, and blending of 30 % CIR in CS could reduce greenhouse gas emissions. Principal component analysis (PCA) was utilized to reduce the dimensionality of experiment and identify the main reactions of CIR-CS co-pyrolysis. Two non-isothermal methods (Kissinger - Akahira - Sunose and Flynn - Wall Ozawa) determined the law of kinetic parameter (E alpha) and thermodynamic parameters changing with the degree of conversion (alpha). Three input parameters (temperature, heating rate, and blending ratio) and one output parameter (mass loss percentage) were used in artificial neural network (ANN) model to predict CS and CIR copyrolysis TG data. ANN 11 was the best predictive model for the co-pyrolysis of CS and CIR.

Automated summary

Thermogravimetry and Fourier transform infrared spectroscopy were used to analyze the co-pyrolysis characteristics of coal slime and coffee industry residue, showing a synergistic effect and reduced greenhouse gas emissions with 30% blending of residue. Principal component analysis and non-isothermal methods identified reaction laws, while an artificial neural network model successfully predicted the co-pyrolysis data.