A reliable model to predict the methane-hydrate equilibrium: An updated database and machine learning approach

Gas hydrates are a type of crystalline compounds that consists of water and small gas molecules. A wide range of applications of gas hydrates in storing natural gas in the form of artificially created solid hydrates, known as solidified natural gas technology, gas separation processes, and seawater desalination technology, has attracted great interest in scientific and practical studies. Gas hydrate formation may also cause deleterious effects such as blockage of gas pipelines. Therefore, accurate prediction of equilibrium conditions for gas hydrates is of great interest. The purpose of this study was to propose machine learning based models to predict methane-hydrate formation temperature for a wide range of brines. In this regard, firstly, a comprehensive database including 987 data samples covering 15 different brines was gathered from the literature. After data cleaning and preparation, three different models of multilayer perceptron, decision tree, and extremely randomized trees were used and tested. The results showed that the extremely randomized trees is capable of predicting methanehydrate formation temperature with good accuracy. The root mean squared error for this model for the testing dataset was acquired as 0.6248, which shows its great accuracy. The findings of this study can be used as a reliable tool to predict the methane-hydrate formation PT curve in the pure water, single-salt brines, and multisalt brines.

Automated summary

Gas hydrates, composed of water and small gas molecules, have diverse applications such as storing natural gas, gas separation, and seawater desalination. However, they can also cause problems like gas pipeline blockage. This study developed machine learning models to accurately predict methane-hydrate formation temperature for various brines, and the results showed good accuracy with extremely randomized trees model.