A consistent set of thermophysical properties of methane curated with machine learning

Accurately predicting thermophysical properties across different states of matter is essential for industrial and scientific applications. However, experimental data often suffer from variability and noise, among other problems, thus necessitating robust modeling approaches to create consistent sets. In this work, we employ machine learning (ML) models to predict multiple thermophysical properties of methane in liquid, vapor, and supercritical phases, including isobaric and isochoric heat capacities, density, volume, Joule-Thomson coefficients, enthalpies, sound speed, and viscosities. We explored different ML algorithms including Adaptive Boosting, Bagging, Decision Trees, Extra Trees, Gradient Boosting, Histogram-based Gradient Boosting Regression Tree, K-Nearest Neighbors, Light Gradient Boosting Machine, Nu-Support Vector Regression, Random Forest, Extreme Gradient Boosting, and Artificial Neural Networks across regions of the phase diagram. Combining ML techniques with previously available raw experimental data shows that ML models provide predictions closer to the statistically treated National Institute of Standards and Technology (NIST) data when compared to the original experimental datasets. Therefore, our approach shows the ML’s potential to identify and generalize complex patterns, smooth inherent data noise, and manage variability. The results indicate that ML models, particularly the Extra Trees and Gradient Boosting models, are a scalable alternative for thermophysical property predictions, offering consistency and efficiency over traditional methods in data processing and curation.