Open Access

Flammability characteristics of hydrogen mixtures have been extensively investigated at different initial conditions(temperature and pressure). Based on the available experimental datasets, empirical and semi-empirical models are commonly used to calculate flammability limits in dependance to initial conditions and mixture composition to reduce the experimental effort. However, unevenly distributed empirical data and the complex non-linear relationship characteristics of these data present significant challenges to empirical flammability limits prediction methods under various mixture initial conditions. Moreover, the empirical models and semi-empirical models only cover some influencing parameters, respectively. To address these issues, the present study adapts a machine learning (ML) approach for improving the hydrogen-air/oxygen-inert gas mixture flammability limits prediction at different conditions with a holistic approach. A Multi-Layer Perceptron (MLP) model was trained, validated, and tested using key input features such as flammability state, initial mixture temperature, equivalence ratio, inert gas concentration, adiabatic flame temperature, and Lewis numbers. Data augmentation techniques were conducted on experimental datasets to improve the predictive capability of the model. The models' performance was compared with empirical flammability limit prediction methods. The goal is to deliver fast, reliable, and more accurate predictions across different scenarios with a single prediction model. Most importantly, the machine learning approach offers a cost-effective and robust alternative to existing empirical flammability limit prediction methods, thus also reducing the experimental effort for explosion limits determination.

Flammability characteristics of hydrogen mixtures have been extensively investigated at different initial conditions(temperature and pressure). Based on the available experimental datasets, empirical and semi-empirical models are commonly used to calculate flammability limits in dependance to initial conditions and mixture composition to reduce the experimental effort. However, unevenly distributed empirical data and the complex non-linear relationship characteristics of these data present significant challenges to empirical flammability limits prediction methods under various mixture initial conditions. Moreover, the empirical models and semi-empirical models only cover some influencing parameters, respectively. To address these issues, the present study adapts a machine learning (ML) approach for improving the hydrogen-air/oxygen-inert gas mixture flammability limits prediction at different conditions with a holistic approach. A Multi-Layer Perceptron (MLP) model was trained, validated, and tested using key input features such as flammability state, initial mixture temperature, equivalence ratio, inert gas concentration, adiabatic flame temperature, and Lewis numbers. Data augmentation techniques were conducted on experimental datasets to improve the predictive capability of the model. The models’ performance was compared with empirical flammability limit prediction methods. The goal is to deliver fast, reliable, and more accurate predictions across different scenarios with a single prediction model. Most importantly, the machine learning approach offers a cost-effective and robust alternative to existing empirical flammability limit prediction methods, thus also reducing the experimental effort for explosion limits determination.

An understanding of Hydrogen-Oxygen/Air-Diluents gas mixtures combustion characteristics and their accurate prediction is crucial for ensuring the safety of hydrogen-related applications, reducing accidents risk, and protecting lives and property. Hydrogen detonation propagations are characterized by the detonation cell size, used to quantitatively predict for a mixture to detonate including, among others, the initiation energy, critical and minimum tube diameters. For the prediction of explosion limits, detonation run-up-distances and cell sizes, various empirical, semi-empirical and numerical models can be found in literature. These models are usually limited to a narrow range of explosion process or geometrical experimental parameters. Moreover, based on the limited availability of the detonation cell widths measurements, current estimation models are seemingly inaccurate. Machine learning models can be utilized to make justifiable prediction on the detonation cell sizes of hydrogen-air mixtures and other gaseous explosive mixtures cell sizes, explosion limits or run-up distance to detonation based on the mixture type, temperature, pressure, equivalence ratios as well as on geometrical parameters with consideration of highly diverse experimental data measurements uncertainties. Therefore, an up-to date database for explosion characteristics will be established and machine learning models will be developed, trained, tested, and validated using experimental data to predict explosion characteristics of hydrogen mixtures. The models predicted results will be validated against existing models. It will be tested whether machine learning models are able to predict the explosion characteristics of hydrogen mixtures with better accuracy and more comprehensively than conventional empirical and numerical models to be found in literature.