Real-time Anomaly Detection for Gas Concentration
WU Haibo
SHI Shiliang
NIAN Qifeng
Abstract：In order to improve the accuracy of real-time risk assessment of gas disaster,it is necessary to analyze and detect anomalies in mine gas concentration streaming data in real time. The real-time detection model of gas concentration anomaly is con?structed by combining streaming linear regression algorithm with statistical analysis technique,and the anomaly real-time detection system of gas concentration is realized by the memory-based distributed streaming processing framework Spark Streaming. The ex?perimental results show that the method can periodically update the anomaly detection model,and real-time detects the anomaly in the streaming data. When the model update period is 45s and the abnormal threshold is set to 0.05,the number of anomaly detection is consistent with that of the boxplot anomaly analysis,and can improve the time-effectiveness of the gas risk assessment.
Keywords：gas concentrationstreaming dataanomaly detectionreal-timestreaming linear regressionSpark Streaming
Publication Date：2019-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：6( 1086-1090,1105 )