Combustion phases of magnesium alloys based on predicted heating rate using machine learning

Magnesium alloys have achieved a highly substantial weight reduction in manufacturing industries, but various organizations have imposed strict restrictions on their usage due to the high flammability of magnesium. This research focuses on phase change during ignition testing to uncover insights into their changing properties in magnesium alloys. In this research, we propose a combustion framework that performed simulation work and utilized several machine learning models for extracting hidden features to predict new phases throughout the combustion process of magnesium alloys. We found a novel phenomenon: the heating rate continuously varied due to phases changing through all combustion processes. The results found that WE43 alloy proves superior resistance at 791 °C for ignition and 841 °C for flammability with the lowest heating rate at 9 °C/min and a most prolonged period of 90 min to ending combustion process as compared to AZ31 at 44.5 min and AZ91 at 39.6 min.