地震预警系统中搜索引擎方法的优化研究

The earthquake damage can be mitigated if we can infer the source location, magnitude and focal mechanism as soon as possible when an earthquake occurs. Recently, the development of the earthquake search engine provides a new direction for determining the source location and focal mechanism quickly. Similar to the web search engine, the source parameters can be determined by matching the input waveform with the synthetic waveform database through the fast search method. In this study, we attempt to combine the search engine with the concept of the Tnow location method and improve the performance of the search engine method to adapt the application of earthquake early warning. The epicenter is generally near the location of the first triggered station. The source can be constrained in a location range by the triggered stations and the stations without P arrivals at current time Tnow; therefore, we can narrow the range of the search engine database. In this study, we attempt to build an index for multiple subdatabases according to the order of the triggered stations. When an earthquake occurs, we retrieve the subdatabase with the order of the triggered stations; then apply the Internet search engine method on the subdatabase to obtain the solutions of the source parameters; therefore, the efficiency of the search engine can be improved.

快速地从台站记录到的地震波数据中推断出震源的位置，量级以及震源机制，对于减轻地震灾害具有重要意义。近年来，借用互联网搜索引擎技术，通过快速匹配地震波形的方法搜索震源位置和震源机制，为地震实时监测提供了新的方向。为了更好地适应地震预警的应用，本研究提出结合现有的Tnow定位方法的思想，优化地震搜索引擎技术方案，使之更适应于地震预警台站触发的时效性要求。由于震中一般处于先触发的台站附近，利用当前Tnow时刻已触发和未触发的台站位置，可以限定震源位于一定区域，从而缩小搜索引擎数据库的范围。本研究中探讨根据台站触发顺序建立数据库结构索引的方法，地震发生后首先索引触发台站对应的子数据库，然后通过互联网搜索引擎算法搜索该子数据库，快速求解震源参数，提高搜索的效率和效果。