地震预警系统中的算法研究

Study on Algorithms in Earthquake Early Warning System

通过对地震预警系统中基本算法的研究,在现有方法的基础上提出了一套地震参数快速估算的算法,以应用于地震预警系统中。该算法根据某些噪音的特殊性质,通过引入两个参数:触发时窗与取消阈值,对常用的P波初到判断方法STA/LTA(长时窗平均/短时窗平均)算法进行改进,以消除P波到达前的异常跳动点与脉冲噪音的影响,对P波到达进行识别。在识别到P波到达后,该算法实时的从P波初到后3s时窗内的竖向数据中提取出12类特征性质不同的计算参数作为估算震级,震中距等地震参数的依据,提高了震级、震中距的估算精度,减少了在大震下估算的离散性,取得了较好的结果。该算法可在地震P波到达后的几秒钟内,估算出地震相关参数,在S波到达前发出预警,并为采取相应的措施以及震后求援行动提供判断依据,达到防灾减灾的目的。

Earthquake early warning system is an effective method for earthquake prevention and reduction.This paper mainly researches the basic algorithms applied in the existing earthquake early warning systems,and presents the algorithm to estimate the earthquake parameters.This algorithm makes a modification to the STA/LTA algorithm,proposed by Allen,to pick the onset of the P-wave while removes the effect of the spike and small-pulses noise.After the detection of the onset of the P-wave,it calculates 12 kinds of the parameter from the first three seconds of the P-wave of the previous strong motion records,in order to estimate earthquake epicentral distance and magnitude.The proposed algorithm improves the accuracy of the estimation,as well as reduces the scattering in the large magnitude range.This algorithm can obtain the earthquake parameters shortly after the onset of the P-wave,and provide early warning alarm before the imminent destructive S-wave.Therefore,the results of this algorithm could guide the preventive actions and the rescue work after the shock.