- Title
- Applications of machine learning to predict and evaluate the explosion of premixed methane- air mixtures
- Creator
- Namazi, Samira
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2024
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- Global warming, caused by the accumulation of greenhouse gases like CO2 and methane, have long-term impact on Earth's climate. Although methane is considered secondary to CO2 in terms of volume, it is considered 29.5 times more potent than CO2 on a 100-year scale. Underground coal mines contribute significantly to global warming through fugitive methane emissions, emphasising the need to capture and eliminate methane. Thermal oxidation, which involves burning methane into water vapour and CO2, is one of the most effective and practical method for reducing methane emissions. However, operating thermal oxidisers at high temperatures increases the risk of fire and explosions in mines. Understanding methane explosion characteristics is crucial for improving safety and minimising explosion risks in underground coal mines fitted with methane mitigation measures. However, conducting methane fire and explosion experiments under relevant conditions is expensive, risky, and time-consuming, requiring thorough preparation and safety procedures. The key aims of this study are twofold: i) to predict methane explosion properties for a typical underground coal mine using data mining approaches, and ii) to analyse the relationship between variables and the contribution of each variable in methane explosions, utilising data mining techniques. The data sets used in this study were obtained from extensive experiments conducted as part of the ventilation air methane abatement safety project conducted at the University of Newcastle, Australia, over the period between 2013 and 2018. The data sets include methane explosion properties such as pressure rise and flame propagation velocity, as a function of other variables such as methane concentration and quantity. In total, 253 data sets, each comprising over 2 million data points, were made available for investigation in this study. The developed data mining tool was validated against the available experimental data. The developed data mining approach in this study could effectively and accurately predict the likelihood of methane explosion as well as explosion properties, and the relationship between variables, in a typical underground coal mine. The outcomes of this research will benefit both researchers and the coal mining industry by providing valuable insights into methane explosion characteristics and consequences, eliminating the need for extensive experimental work. Additionally, the data mining tool will enable mine safety professionals to determine and implement appropriate prevention methodologies and countermeasures, effectively reducing the risk of methane explosions and their devastating consequences.
- Subject
- global warming; methane; mining; methane explosion
- Identifier
- http://hdl.handle.net/1959.13/1509980
- Identifier
- uon:56334
- Rights
- Copyright 2024 Samira Namazi
- Language
- eng
- Full Text
- Hits: 1056
- Visitors: 1057
- Downloads: 40
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | ATTACHMENT01 | Thesis | 2 MB | Adobe Acrobat PDF | View Details Download | ||
| View Details Download | ATTACHMENT02 | Abstract | 339 KB | Adobe Acrobat PDF | View Details Download |