Gas monitoring data anomaly identification based on spatio-temporal correlativity analysis

• Published in: Journal of coal science & engineering, China Vol. 19; no. 1; pp. 8 - 13
• Main Authors: Zhu, Shi-song, Wang, Yun-jia, Wei, Lian-jiang
• Format: Journal Article
• Language: English
• Published: Heidelberg China Coal Society 01.03.2013
• Subjects: Earth and Environmental Science; Earth Sciences; Geochemistry; Geology; Geophysics/Geodesy; Geotechnical Engineering & Applied Earth Sciences; Mechanical Engineering; 异常识别; 时空; 气体传感器; 瓦斯监测; 监测数据; 相关性分析方法; 自动识别技术; 采煤工作面; anomaly pattern identification; spatio-temporal correlativity analysis; gas monitoring; algorithm
• ISSN: 1006-9097; 1866-6566
• DOI: 10.1007/s12404-013-0102-y

Based on spatio-temporal correlativity analysis method, the automatic identification techniques for data anomaly monitoring of coal mining working face gas are presented. The asynchronous correlative characteristics of gas migration in working face airflow direction are qualitatively analyzed. The calculation method of asynchronous correlation delay step and the prediction and inversion formulas of gas concentration changing with time and space after gas emission in the air return roadway are provided. By calculating one hundred and fifty groups of gas sensors data series from a coal mine which have the theoretical correlativity, the correlative coefficient values range of eight kinds of data anomaly is obtained. Then the gas moni- toring data anomaly identification algorithm based on spatio-temporal correlativity analysis is accordingly presented. In order to improve the efficiency of analysis, the gas sensors code rules which can express the spatial topological relations are sug- gested. The experiments indicate that methods presented in this article can effectively compensate the defects of methods based on a single gas sensor monitoring data.