Prediction of ventilation air methane explosion in regenerative thermal oxidation based on hyperparameter-optimized random forest algorithm

Abstract

Regenerative Thermal Oxidation (RTO) is a key technology for utilizing Ventilation Air Methane (VAM), with safety assessments depending on accurate explosion predictions. This study develops a predictive model using a particle swarm optimization-random forest (PSO-RF) algorithm to determine whether methane will explode under various conditions in regenerative thermal oxidation. Experimental data were collected to determine the critical transition from oxidation to explosion. By integrating the ultra-lean methane oxidation kinetics model (GRTO) with Grey Relational Analysis (GRA), ambient temperature and methane concentration were identified as critical input features. The RF model's hyperparameters were optimized via the improved PSO algorithm to improve accuracy and computational efficiency. The dataset was randomly split into a training set and a testing set in a 7:3 ratio. The PSO-RF model was subsequently compared with Support Vector Machine (SVM), Decision Tree (DT), and Neural Network (NN) models. Results indicate that the PSO-RF model outperforms other models in key classification metrics on the test set, effectively predicting explosion risks under complex conditions.