Research on coal mine longwall face gas state analysis and safety warning strategy based on multi-sensor forecasting models

Chang, Haoqian; Meng, Xiangrui; Wang, Xiangqian; Hu, Zuxiang

Abstract

Intelligent computing is transforming safety inspection methods and response strategies in coal mines. Due to the significant safety hazards associated with mining excavation, this study proposes a multi-source data based predictive model for assessing gas risk and implementing countermeasures. By examining the patterns of gas dispersion at the longwall face, utilizing both temporal and spatial correlation, a predictive model is crafted that incorporates safety thresholds for gas concentrations, four-level early warning method and response strategy are devised by integrating weighted predictive confidence with these correlations. Initially tested using a public dataset from Poland, this method was later verified in coal mine in China. This paper discusses the validity and correlation of multi-source monitoring data in temporal and spatial correlation and proposes a risk warning mechanism based on it, which can be applied not only for safety warning but also for regulatory management.

Citation

Chang, H., Meng, X., Wang, X., & Hu, Z. (2024). Research on coal mine longwall face gas state analysis and safety warning strategy based on multi-sensor forecasting models. Scientific Reports, 14(1), Article 13795. https://doi.org/10.1038/s41598-024-64181-7

• Journal Article Type: Article
• Acceptance Date: Jun 5, 2024
• Online Publication Date: Jun 14, 2024
• Publication Date: Jun 14, 2024
• Deposit Date: Jul 10, 2024
• Journal: Scientific Reports
• Electronic ISSN: 2045-2322
• Publisher: Nature Research
• Peer Reviewed: Peer Reviewed
• Volume: 14
• Issue: 1
• Article Number: 13795
• DOI: https://doi.org/10.1038/s41598-024-64181-7
• Keywords: Time series analysis, Mining management engineering, Mining safety, Warning strategy
• Public URL: https://durham-repository.worktribe.com/output/2493202
• Additional Information: This article is available open access via https://doi.org/10.1038/s41598-024-64181-7