Real-time Detection for Anomaly Data in Microseismic Monitoring System

Microseismic monitoring means to records microseismic activities caused by the changes of the rock physical properties continuously through the high sensitivity seismic sensor placed in mine. How to make real-time detection of abnormal data in mine microseisms positioning system is a extremely important task. Forecast model and mechanism of data stream in the mine microcosmic monitoring system are given through the linear self-regression analysis. Based on this prediction model, a detection method of abnormal data is proposed. This method detects whether real-time data is abnormal by calculating the ratio of real-time forecast error and average forecast error and making a comparison between the ratio and predefined threshold. Experimental results verified correctness and effectiveness of the prediction model to show that the model can realize real-time detection of abnormal event in mine earthquake.