微地震位置和震源机制的快速波形反演及搜索引擎算法的研究

发布时间：2017-12-31 08:33

本文关键词：微地震位置和震源机制的快速波形反演及搜索引擎算法的研究　出处：《中国科学技术大学》2016年博士论文　论文类型：学位论文

【摘要】：徽地震监测已经被广泛应用于矿场,地热,和油气行业。比如,它是页岩气开发中用来成像注水压裂的裂缝分布的非常好的工具,微地震位置和震源机制等可以为现场工程师提供重要的信息来评估注水压裂的效力等等。本文主要目的在于利用全波形匹配的方法来推断微震位置和震源机制解,分别从两个不同的方法出来研究这个问题。一方面,从波形反演的角度来同时获得微震的位置和震源机制解。我们首先介绍一种基于梯度的方法来同时反演震源位置和震源机制,为了克服这种基于梯度的方法的局部极小值的问题,我们同时提出一种基于一种全局最优化算法的快速波形反演算法。另一方面,从快速搜索的观点来研究这个问题,我们提出应用计算机科学行业的搜索引擎的概念来解决微震波形的匹配问题。本文中主要研究的两个方法可以总结如下：1,微震位置和震源机制的快速弹性波全波形反演算法给定一个速度模型,我们首先在可能的震源位置网格点上计算好格林函数库。反演过程中我们计算一个基于离散和预先准备好的格林函数库的近似的目标函数。由于提前计算好了格林函数,合成波形的计算变得快捷,从而使得我们能够利用一种全局最优化算法,邻域算法,来实现同时反演震源位置和震源机制。该方法中的目标函数应用了包络相关的概念来匹配高频的波形数据。在事件检测后,该方法并不需要拾取P和S波走时,我们利用波形残差和相关系数来评价波形的拟合程度。我们利用合成和实际数据例子测试了该方法,在输入数据中包含噪音和速度模型存在误差的情况下,合成数据例子表明能够较好的恢复出真实模型。我们同时测试了452个实际数据微震事件,并且和走时网格搜索算法进行了对比,快速波形反演得到的微震位置分布能够和走时网格搜索算法结果一致。2.微震搜索引擎算法类似于互联网搜索引擎,该方法能够在1s内同时估计出微震事件的位置和震源机制,来检测注水压裂过程。对于给定的采集系统和速度模型,我们首先在所有可能的位置网格点上计算所有可能的微震事件波形,从而建立一个搜索数据库。然后通过计算机快速搜索技术,多个随机K维树方法,根据数据库中波形数据的振幅和相位信息来排列并建立一个索引。当一个微震事件发生时,和输入波形近似的最佳波形能够通过匹配数据库的特征很快地找到。该方法不仅仅返回一个最佳解,而是类似于互联网搜索中的一个解集,因此我们可以利用得到的解集来进一步研究结果的置信度和解析度。同样类似于互联网搜索引擎,微震搜索引擎不需要其他输入参数和处理经验：这样,对于任何用户结果都会一样。我们同样利用合成数据和实际数据例子验证了该方法,结果显示微震搜索引擎有很大的潜力应用于微地震的实时监测问题。
[Abstract]:Hui seismic monitoring has been widely used in mining, geothermal, and oil and gas industries. For example, it is a very good tool for imaging the distribution of fracturing fractures in shale gas development. The microseismic location and focal mechanism can provide important information for field engineers to evaluate the effectiveness of waterflooding fracturing. The main purpose of this paper is to infer the microseismic location and focal mechanism solution by using the method of full waveform matching. . Two different ways to study the problem. On the one hand. From the angle of waveform inversion, the location and focal mechanism of microearthquakes are obtained simultaneously. Firstly, we introduce a gradient-based method for simultaneous inversion of focal positions and focal mechanisms. In order to overcome the problem of local minima of this gradient-based method, we also propose a fast waveform inversion algorithm based on a global optimization algorithm. Look at this from the point of view of fast search. We put forward the concept of search engine in computer science industry to solve the problem of matching microseismic waveforms. The two main methods in this paper can be summarized as follows: 1. The fast elastic wave inversion algorithm for the location and focal mechanism of microearthquakes is given a velocity model. We first calculate the Green's function library on the grid point of the possible focal point. In the inversion we calculate an approximate objective function based on discrete and pre-prepared Green's function library. Lin function. The computation of synthetic waveforms becomes faster, which enables us to use a global optimization algorithm, neighborhood algorithm. The objective function of the method uses the concept of envelope correlation to match the high frequency waveform data. After event detection, the method does not need to pick up P and S wave travel time. We use waveform residuals and correlation coefficients to evaluate the fitting degree of waveforms. We use synthetic and practical data examples to test the method when the noise and velocity model errors are included in the input data. The synthetic data examples show that the real model can be recovered better. We also tested 452 actual data microseismic events and compared with the traveling time grid search algorithm. The microseismic location distribution obtained by fast waveform inversion can be consistent with the results of walking time grid search algorithm. 2.The microseismic search engine algorithm is similar to the Internet search engine. This method can simultaneously estimate the location and focal mechanism of microseismic events in 1 s to detect the fracturing process of water injection. For a given acquisition system and velocity model. We first calculate the possible microseismic event waveforms on all possible grid points, and then establish a search database. Then, through the computer fast search technique, several random K-dimensional tree methods are used. Arrange and build an index based on the amplitude and phase information of the waveform data in the database. When a microseismic event occurs. The best waveform similar to the input waveform can be quickly found by matching the features of the database. This method not only returns an optimal solution, but is similar to a solution set in Internet search. So we can use the solution set to further study the confidence and resolution of the results. Similar to the Internet search engine, the microseismic search engine does not need other input parameters and processing experience: so. For any user the results are the same. We also use synthetic data and real data examples to verify the method. The results show that the microseismic search engine has great potential to be applied to the real-time monitoring of micro-seismic problems.
【学位授予单位】：中国科学技术大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TE357.6;P631.4

【参考文献】