Predicting Self-Heating Temperature and Influencing Factors in the Cement Composite Mixed with Multi-Walled Carbon Nanotubes Using Machine Learning

The self-heating temperature of the cement composite mixed with multi-walled carbon nanotubes (MWCNT–cement composite) is influenced by several factors, including the concentration of nano-material. However, conducting experiments to measure this temperature is time-consuming and expensive. Additionally, there are challenges in elucidating the correlations between the various influencing factors of the MWCNT–cement composite and its self-heating temperature. This study utilizes machine learning (ML) to predict the self-heating temperature of the MWCNT–cement composite and identify the correlation with influencing factors. ML techniques, including Random Forest (RF), eXtreme Gradient Boosting (XGB), and Gradient Boosting Machine (GBM), were employed. These ML models were optimized through hyperparameter tuning and k-fold cross-validation. The predictive performance of each model was evaluated using R2, mean absolute error (MAE), mean square error (MSE), and root mean square error (RMSE) metrics. All ML models exhibited high predictive performance, with the GBM model demonstrating the best thermal prediction capability, achieving an R2 value of 0.9795. Subsequently, the GBM model was used to analyze the major factors affecting the self-heating temperature of the MWCNT–cement composite. The analysis revealed that the concentration of MWCNTs, the amount of voltage, and the outdoor temperature are significant factors determining the self-heating temperature. Furthermore, it was found that the self-heating temperature of the MWCNT–cement composite increases as the concentration of MWCNTs and the amount of voltage increase and as the distance of the mesh decreases.