论文部分内容阅读
震源定位是微震监测的核心基础.目前使用较多的射线走时追踪定位方法主要基于L1和L2范数建立震源定位目标函数.然而,L1范数定位精度较低,L2范数易受P波初至大拾取误差影响.此外,传统定位方法易受初始点影响而得到局部最优,远距离P波初至数据质量较差.为此,本文提出了一种基于Log-Cosh函数及剔除远距离传感器到时的微震震源定位方法,其基本原理为:首先,建立基于Log-Cosh函数的震源定位目标函数,其具有L1范数的稳定性和L2范数的定位精度;其次,在矿山开采区域随机生成多个初始点,使用已建立的震源定位目标函数得到相应的多个初定位点,再以初定位点中数据场势值最大的50个定位点坐标均值作为初定位结果,计算初定位结果到各触发传感器的P波传播时间,剔除传播时间大于0.2 s的P波初至数据;最后,以去噪后的P波初至数据集重复上述定位步骤,得到高精度定位结果.以开阳磷矿用沙坝矿两个理论测试事件和八次爆破事件为例展开定位研究.结果表明,不管是否剔除传播时间较大的P波初至数据,基于数据场的Log-Cosh法定位误差都小于传统TT1法和TT2法定位误差,而在剔除传播时间较大的P波初至数据后,三种定位方法的定位精度都有所提高,表明本文所提出的定位方法具有较好的应用前景.“,”Source location is the core foundation of microseismic monitoring. To date, commonly used location methods have usually been based on the ray-tracing travel-time technique, which generally adopts an L1 or L2 norm to establish the location objective function. However, the L1 norm usually achieves low location accuracy, whereas the L2 norm is easily affected by large P-wave arrival-time picking errors. In addition, traditional location methods may be affected by the initial iteration point used to find a local optimum location. Furthermore, the P-wave arrival-time data that have travelled long distances are usually poor in quality. To address these problems, this paper presents a microseismic source location method using the Log-Cosh function and distant sensor-removed P-wave arrival data. Its basic principles are as follows: First, the source location objective function is established using the Log-Cosh function. This function has the stability of the L1 norm and location accuracy of the L2 norm. Then, multiple initial points are generated randomly in the mining area, and the established Log-Cosh location objective function is used to obtain multiple corresponding location results. The average value of the 50 location points with the largest data field potential values is treated as the initial location result. Next, the P-wave travel times from the initial location result to triggered sensors are calculated, and then the P-wave arrival data with travel times exceeding 0.2 s are removed. Finally, the aforementioned location steps are repeated with the denoised P-wave arrival dataset to obtain a high-precision location result. Two synthetic events and eight blasting events from the Yongshaba mine, China, were used to test the proposed method. Regardless of whether the P-wave arrival data with long travel times were eliminated, the location error of the proposed method was smaller than that of the L1/L2 norm and trigger-time-based location method (TT1/TT2 method). Furthermore, after eliminating the P-wave arrival data with long travel distances, the location accuracy of these three location methods increased, indicating that the proposed location method has good application prospects.