【摘要】：地震波阻抗反演是地震儲層預測的重要方面,壓縮感知由于具有高分辨率,低模型依賴優(yōu)勢,在地震信號處理領域中日益成為熱門的研究方向。壓縮感知在地震波阻抗反演中的應用主要包括基追蹤阻抗反演和匹配追蹤阻抗反演等,基追蹤方法是采用最小化第一范數實現信息重構,匹配追蹤方法是采用貪婪算法局部匹配最優(yōu)化字典實現信號重構。本文首先研究了壓縮感知中的匹配追蹤和基追蹤方法原理。分析了壓縮感知相對于傳統(tǒng)的數字信號處理方法如傅里葉變換、小波變換的優(yōu)勢。壓縮感知不同于傳統(tǒng)的數字信號處理技術,它通過構建相應的字典庫,實現了使用少量的觀測值對原始稀疏信號的表示。并分析壓縮感知的基追蹤和匹配追蹤方法如何應用于地震阻抗反演。以及研究壓縮感知中基追蹤和匹配追蹤方法在地震波阻抗反演應用的關鍵性因素。其次研究了匹配追蹤在地震波阻抗反演的實現方法。分別實現了基于單層(單極子)匹配追蹤阻抗反演、基于兩層調諧(雙極子)匹配追蹤阻抗反演、和基于四層調諧(四極子)匹配追蹤阻抗反演。根據匹配追蹤在實際地震記錄分解中存在的不穩(wěn)定性問題提出:在原子庫構建中先確定子波的相位、主頻等取值范圍。匹配追蹤分解策略在原有的地震信號投影到原子庫的最大值的基礎上考慮原地震信號與匹配合成地震信號的誤差值因素。第三,研究了基追蹤在地震波阻抗反演中的實現,先后實現了基于單極子和雙極子基追蹤阻抗反演。為進一步提高多層調諧下的反演分辨率建立了四極子基追蹤阻抗反演方法。最后建立相應模型試驗分析基追蹤的整體一致性優(yōu)勢和匹配追蹤的局部最優(yōu)化優(yōu)勢。分析了匹配追蹤阻抗反演和基追蹤阻抗反演各自存在的問題。通過綜合兩種反演方法的各自優(yōu)勢,即首先通過匹配追蹤獲得各反射層的最佳匹配地震子波,然后基追蹤使用匹配追蹤已經匹配好的地震子波構建分解的原子庫。再使用基于四極子基追蹤阻抗反演方法實現地震信息分解,從而得到更高分辨率的各反射層相對反射系數序列并求得各地層波阻抗。
[Abstract]:Seismic wave impedance inversion is an important aspect of seismic reservoir prediction. Because of its high resolution and low model dependence, compression perception has become a hot research field in the field of seismic signal processing. The applications of compression sensing in seismic wave impedance inversion mainly include base tracing impedance inversion and matching tracing impedance inversion. The basis tracking method is to realize information reconstruction by minimizing the first norm. The matching tracking method uses greedy local matching optimization dictionary to realize signal reconstruction. In this paper, we first study the principle of matching tracking and base tracking in compressed sensing. The advantages of compression sensing over traditional digital signal processing methods such as Fourier transform and wavelet transform are analyzed. Compression sensing is different from the traditional digital signal processing technology. It uses a small number of observations to represent the original sparse signal by constructing the corresponding dictionary. How to apply the base tracking and matching tracking methods of compression sensing to seismic impedance inversion is analyzed. The key factors of the application of base tracking and matching tracing in compression sensing in seismic impedance inversion are also studied. Secondly, the realization method of matching tracing in seismic wave impedance inversion is studied. The inversion is based on single-layer (monopole) matching tracing impedance inversion, two-layer tuning (bipolar) matching tracing impedance inversion, and four-layer tuning (quadrupole) matching tracing impedance inversion. According to the instability of matching tracing in the decomposition of actual seismic records, it is proposed that the wavelet phase and the main frequency range should be determined in the construction of atomic library. The matching tracing decomposition strategy takes into account the error between the original seismic signal and the matching synthetic seismic signal on the basis of the original seismic signal projected to the maximum value of the atomic library. Thirdly, the realization of base tracing in seismic wave impedance inversion is studied, which is based on unipolar and bipolar basis tracing impedance inversion. In order to further improve the inversion resolution of multilayer tuning, a quadrupole subbase tracing impedance inversion method is established. Finally, the global consistency advantage of base tracking and the local optimization advantage of matching tracking are established. The problems of matching tracing impedance inversion and base tracing impedance inversion are analyzed. By synthesizing the respective advantages of the two inversion methods, firstly, the best matching seismic wavelet of each reflection layer is obtained by matching tracing, and then the base tracking uses matching tracing to construct the decomposed atomic library. The seismic information is decomposed by using the method of tracing impedance inversion based on quadrupole basis, and the relative reflection coefficient sequence of each reflection layer with higher resolution is obtained, and the wave impedance of each layer is obtained.
【學位授予單位】：成都理工大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：P618.13;P631.4

【參考文獻】

相關期刊論文 前10條

1 印興耀;劉曉晶;吳國忱;宗兆云;;模型約束基追蹤反演方法[J];石油物探;2016年01期

2 劉曉晶;印興耀;吳國忱;宗兆云;;基于基追蹤彈性阻抗反演的深部儲層流體識別方法[J];地球物理學報;2016年01期

3 郝亞炬;文曉濤;李忠;李世凱;李天;;基于基追蹤分解算法的薄層波阻抗反演[J];科學技術與工程;2015年33期

4 楊昊;鄭曉東;李勁松;馬淑芳;;基于匹配追蹤的薄層自動解釋方法[J];石油地球物理勘探;2013年03期

5 黃捍東;郭飛;汪佳蓓;任敦占;;高精度地震時頻譜分解方法及應用[J];石油地球物理勘探;2012年05期

6 李雪英;陳樹民;王建民;裴江云;王元波;;薄層時頻特征的正演模擬[J];地球物理學報;2012年10期

7 武國寧;曹思遠;孫娜;;基于復數道地震記錄的匹配追蹤算法及其在儲層預測中的應用[J];地球物理學報;2012年06期

8 張繁昌;李傳輝;印興耀;;三角洲砂巖尖滅線的地震匹配追蹤瞬時譜識別方法[J];石油地球物理勘探;2012年01期

9 張繁昌;李傳輝;;基于正交時頻原子的地震信號快速匹配追蹤[J];地球物理學報;2012年01期

10 趙嵩;馬榮華;薛朝改;李恒建;;基于樹型冗余字典正交匹配追蹤的信號稀疏分解[J];揚州大學學報(自然科學版);2011年04期

相關碩士學位論文 前7條

1 屈冉;壓縮感知算法及其應用研究[D];南京郵電大學;2013年

2 文首先;壓縮感知匹配追蹤算法的研究[D];安徽大學;2013年

3 任曉馨;壓縮感知貪婪匹配追蹤類重建算法研究[D];北京交通大學;2012年

4 王方非;基于樹形結構回溯正交匹配追蹤的稀疏恢復算法研究[D];北京交通大學;2012年

5 胡軍;基于壓縮傳感稀疏重構方法的研究[D];湖南大學;2012年

6 李亞文;遺傳匹配追蹤算法的研究與改進[D];江南大學;2011年

7 趙玉娟;基于子空間匹配追蹤的信號稀疏逼近[D];西安電子科技大學;2005年

，

本文編號：2226248