煤矿安全监控系统监测点数据异常识别技术研究

Study on technology of identifying data abnormality of monitoring points in colliery safety monitoring system

通过分析监控系统数据异常波动的影响因素,建立了异常识别原则,并设计了异常识别的模型。应用实践表明,综合运用关联分析、聚类分析及时间序列等方法,使异常识别准确率提高到78%以上,从而提高了监控数据的准确性和可靠性。

After analyzing the factors affecting data abnormal variation of the monitoring system, the abnormality identifying laws were determined, and the abnormality identifying model was designed. Practical application showed that the integrated application of correlation analysis, cluster analysis and time series made the identification accuracy increase to above 78%, thus the accuracy and reliability of the monitoring data improved.