论文部分内容阅读
利用单台站的信息在较短的时间内识别地震事件和干扰事件是地震预警系统面临的关键问题。研究了基于波形特征参数的决策树方法来提高地震预警系统的单台抗干扰能力。为了研究电磁干扰对地震预警系统的影响,进行了7种电磁干扰试验。使用事件触发后3 s内的波形数据提取5个特征(事件结束否、Speed、Sparse、Pration、对称性),用决策树建立了特征值和事件类型之间的映射关系。使用了大量的强震记录和干扰记录(6359条强震记录、163条干扰记录)对训练后的决策树进行了验证。结果表明,所提方法识别率高,易转换成IF-THEN规则,能方便地写到计算机程序或专家系统中。该算法可以作为地震预警或报警系统的前端地震事件检测模块,以提高系统的可靠性。
Using the information of a single station to identify seismic events and interference events in a short period of time is a key issue for the earthquake early warning system. The decision tree based on waveform feature parameters is studied to improve the single anti-jamming capability of earthquake early warning system. In order to study the influence of electromagnetic interference on the earthquake early warning system, seven electromagnetic interference tests were conducted. Five features (event no, Speed, Sparse, Pration, symmetry) were extracted from the waveform data within 3 s after triggering. The mapping between eigenvalues and event types was established using a decision tree. A large number of strong earthquake records and interference records (6359 strong earthquake records and 163 interference records) were used to verify the training decision tree. The results show that the proposed method has high recognition rate and is easy to convert into IF-THEN rules and can be easily written into computer programs or expert systems. The algorithm can be used as a front seismic event detection module in earthquake warning or alarm system to improve the system reliability.