起伏地表地震波场定向照明分析与随机噪声压制

发布时间：2018-04-10 06:33

本文选题：起伏地表　切入点：定向照明　出处：《吉林大学》2017年硕士论文

【摘要】：无论油气勘探还是固体矿产勘探都进入了高精细的勘探阶段,从构造简单的地区向复杂的地区延伸,这些地区具有复杂的地质构造,地表形态多高低剧烈起伏,岩性变化快,人文环境情况复杂。地面观测系统的空间几何布局、波场在目标体上覆构造中的传播情况都将会直接影响地震资料的质量,基于水平观测界面的地震勘探方法已经越来越难以满足这类地区的开发工作,无法在地面采集到足够的反射记录。另外,在上述地表高低起伏的勘探地区采集的地震记录中,除了包含声波、面波等有规律的噪声之外,还会有各种无法预测的随机噪声,干扰有效地震信号,甚至会导致有效同相轴畸变,严重影响地震数据质量,给数据处理工作带来严峻挑战。本文针对复杂起伏的地表形态,研究了地震波场定向照明技术和地震数据随机噪声压制方法,为油气和金属矿的勘探与开发提供了理论支持。地震照明分析是对复杂地区进行观测系统设计、研究复杂构造对地震波场能量分布影响的重要手段。组合震源能激发定向的地震波场、对地下构造进行定向照明。但前人的研究都是基于水平的速度模型,在复杂的勘探区常常会存在照明阴影,影响偏移成像结果。本文在前人研究的基础上,采用变网格有限差分算法模拟复杂地表下的地震波场,研究并推导了非水平地表上组合震源激发定向地震波场及定向照明分析的方法。组合震源激发延时的计算是激发定向地震波场、进行定向地震照明的关键,主要运用了坐标变换的方法推导了不同地表形态下组合震源激发延时的计算方法。对多个速度模型进行了地震定向波场模拟和定向照明计算,这些速度模型地表或倾斜或高低起伏,均取得了理想的效果,照明方向明显。另外,从理论和数值模拟两个角度研究了震源间距、震源数量和检波器排列位置等采集参数对来自组合震源的地震数据的影响,以更好地指导复杂勘探区内地震观测系统参数的设置。来自复杂勘探区中的地震数据往往含有大量不规则的随机噪声,为了压制随机噪声对有效信号的干扰,研究了基于字典学习算法的地震数据随机噪声压制方法。在传统K-SVD字典学习算法去噪的基础上,从字典更新阶段和稀疏系数更新阶段两个方面对算法进行了改善:在字典更新阶段循环更新字典,以提高字典对复杂地震数据的适应性;在稀疏系数更新阶段,充分利用大稀疏系数代表有效信号的特点,采用部分大稀疏系数作为系数更新的初始值,可以提高算法的收敛速度。数值试验表明改善后的随机噪声压制算法能在压制随机噪声的同时保护有效信息,同相轴更清晰,显著提高了地震资料的信噪比。
[Abstract]:Both oil and gas exploration and solid mineral exploration have entered a high level of fine exploration. They extend from simple structural areas to complex areas. These areas have complex geological structures, and the surface forms fluctuate sharply, and the lithology changes rapidly.The human environment is complex.The spatial geometric layout of the ground observation system and the propagation of the wave field in the overlying structure on the target body will directly affect the quality of the seismic data.The seismic exploration method based on horizontal observation interface has become more and more difficult to meet the development of this kind of area, and it is unable to collect enough reflection records on the ground.In addition, in addition to regular noises such as sound waves and surface waves, in the seismic records collected in the above-mentioned exploration areas with high and low surface fluctuations, there will also be various unpredictable random noises that interfere with effective seismic signals.It can even lead to effective coaxial distortion, seriously affect the quality of seismic data, and bring serious challenges to data processing.In this paper, the directional illumination technique of seismic wave field and the random noise suppression method of seismic data are studied for the complex undulating surface morphology, which provides theoretical support for the exploration and development of oil, gas and metal deposits.Seismic illumination analysis is an important means to design the observation system in complex areas and to study the influence of complex structures on the energy distribution of seismic wave field.Combined focal can excite the directional seismic wave field and illuminate the underground structure.However, previous studies are based on horizontal velocity models, and illumination shadows often exist in complex exploration areas, which affect the migration imaging results.In this paper, based on the previous studies, the finite difference method with variable grid is used to simulate the seismic wave field under the complex surface, and the method to analyze the directional seismic wave field induced by the combined source on the non-horizontal surface and the directional illumination analysis are studied and deduced.The calculation of excitation delay of combined source is the key of directional seismic illumination. The method of coordinate transformation is mainly used to deduce the calculation method of excitation delay of combined source under different surface morphology.Seismic directional wave field simulation and directional illumination calculation have been carried out on several velocity models. These velocity models have achieved ideal results and the illumination direction is obvious.In addition, the effects of acquisition parameters such as source spacing, number of focal points and location of geophone arrangement on seismic data from combined sources are studied from two aspects of theoretical and numerical simulation.It can better guide the parameter setting of seismic observation system in complex exploration area.Seismic data from complex exploration areas often contain a large amount of irregular random noise. In order to suppress the interference of random noise to the effective signal, a dictionary learning algorithm based on random noise suppression method for seismic data is studied.Based on the traditional K-SVD dictionary learning algorithm, this paper improves the algorithm from the dictionary updating stage and the sparse coefficient updating stage. In order to improve the adaptability of the dictionary to the complex seismic data, the dictionary is updated periodically in the dictionary update stage;In the stage of sparsity coefficient updating, the characteristic of large sparse coefficient representing effective signal is fully utilized, and partial large sparse coefficient is used as the initial value of coefficient updating, which can improve the convergence speed of the algorithm.Numerical experiments show that the improved random noise suppression algorithm can protect the effective information while suppressing the random noise, and the cophase axis is clearer, and the signal-to-noise ratio of seismic data is improved significantly.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P631.4

【相似文献】