基于L-M算法的微地震定位方法

由于可引入复杂速度模型，基于振幅叠加网格搜索类定位方法大大提高了微地震定位的可信度。为保证定位精度，常规振幅叠加网格搜索定位方法需对目标区域进行较细网格划分，导致计算效率较低，无法满足压裂实时监测的要求。本文以振幅叠加网格搜索方法为基础，首先对目标区域进行粗网格剖分，然后在能量叠加较高的局部区域采用Levenberg-Marquardt（L-M）反演算法逐次迭代寻找能量叠加最大值点。该方案能够在保证定位精度的情况下大幅提高计算效率，模型试算及实际资料处理证明了方法的有效性。

A microseismic localization method based on L-M inversion algorithm

With the introduction of complex velocity model,the reliability of micro-seismic localization is greatly improved by the grid type searching based on the amplitude stacking.For the conventional grid searching and amplitude stacking,a finer gridding in target areas is needed to ensure the localization precision,which reduces the computational efficiency,and cannot satisfy the real-time monitoring in the hydraulic fracturing.In this paper,we propose an improved localization method based on Levenberg-Marquardt (L-M) algorithm.First a coarse gridding is performed on target areas based on the grid searching and amplitude stacking.Then the L-M inversion algorithm is used to find the maximum point of energy superposition in a high energy superposition zone.This scheme can greatly improve the calculation efficiency while the localization accuracy is guaranteed.Model and real data tests prove the effectiveness of the proposed method.