基于矩张量反演的传感器校准方法应用和对比

Application and comparison of sensor calibration methods based on moment tensor inversion

矩张量反演理论是研究深部岩体破裂机制的有效手段,而监测台站/传感器校准对获取准确的矩张量结果十分重要。为求得更加准确的校准系数和矩张量结果,提出一种新的传感器校准方法——搜索校准法。将该方法和监测网校准法分别应用于加利福尼亚州北部的Geysers地热田微震监测数据中,并综合考虑不同震源机制参数、预设校准系数、噪声添加方式及噪声水平等因素的影响,进行两种校准方法有效性对比的理论计算和模拟分析。结果表明:两种校准方法均可通过微震监测数据获得稳定的校准系数,且校准后的震源成分及应力状态分布均更加集中;所有模拟工况下,两种校准方法均可有效降低矩张量反演误差;低噪声条件下,两种方法对应的矩张量反演误差较小且十分接近;混合噪声和高噪声条件下,搜索校准法的准确性和稳定性多优于监测网校准法;进而优选出Geysers地热田微震监测数据更加可靠的矩张量结果。研究思路和结论为微地震矩张量研究提供了进一步指导。

Moment tensor inversion is an effective method to study the mechanism of deep rock mass failure,and station/sensor calibration is very important for obtaining accurate moment tensors.In order to obtain more accurate sensor calibration coefficients and moment tensors,a new sensor calibration method is developed:the search calibration method.The proposed method and the network calibration method are respectively applied to the microseismic monitoring data of the Geysers geothermal field in northern California.Then,based on the calibration results,a series of theoretical calculations and simulations are conducted to verify the effectiveness of the two different calibration methods.By comprehensively considering the influence of factors such as different focal mechanism parameters,preset calibration coefficients,noise addition methods and noise levels,the theoretical calculation and simulation analysis of the effectiveness comparison of calibration methods are carried out.The results show that both calibration methods can obtain stable coefficients from microseismic monitoring data,the components and stress state distribution of events are more concentrated after calibration;under all simulated conditions,the two calibration methods can effectively reduce the moment tensor errors;under low noise level condition,the moment tensor errors of two calibration methods are very small and similar;under mixed noise level and high noise level conditions,the accuracy and stability of the search calibration results are better than that of network calibration in most cases;based on the simulation results,the more reliable moment tensors of the Geysers microseismic data are selected.The research ideas and conclusions can provide further guidance for the study of microseismic moment tensor.