| 纵向和横向同时约束AVO反演 |
| |
| 引用本文： | 霍国栋,杜启振,王秀玲,陈刚,郑笑雪.纵向和横向同时约束AVO反演[J].地球物理学报,2017,60(1):271-282. |
| |
| 作者姓名： | 霍国栋 杜启振 王秀玲 陈刚 郑笑雪 |
| |
| 作者单位： | 1. 中国石油大学(华东)地球科学与技术学院, 青岛 266580;
2. 中国石化石油物探技术研究院, 南京 211103;
3. 中石油新疆油田公司勘探开发研究院, 克拉玛依 834000 |
| |
| 基金项目： | 国家自然科学基金（41574125）、国家科技重大专项（2016ZX05018-005）、中国石油天然气集团公司（2016A-3302）和中央高校基本科研业务费专项资金（15CX08002A）联合资助. |
| |
| 摘 要： | 为了提高AVO(amplitude versus offset)反演结果的精度和横向连续性,本文提出了一种新的AVO反演约束方法,该方法结合贝叶斯原理和卡尔曼滤波算法实现了对反演参数纵向和横向的同时约束.文章首先结合反演参数的纵向贝叶斯先验概率约束和反演参数的横向连续性假设建立了与卡尔曼滤波算法对应的AVO反演系统的数学模型,然后将该数学模型代入卡尔曼滤波算法框架,利用卡尔曼滤波算法实现了双向约束AVO反演.二维模型测试和实际数据测试结果表明,相对于单纯的纵向贝叶斯先验概率约束,双向约束能更准确地刻画参数的横向变化,得到更准确、横向连续性更好的反演结果.
|
| 关 键 词： | 贝叶斯先验概率约束 卡尔曼滤波算法 双向约束 横向连续性 |
| 收稿时间： | 2015-05-04 |
AVO inversion constrained simultaneously in vertical and lateral directions |
| |
| HUO Guo-Dong,DU Qi-Zhen,WANG Xiu-Ling,CHEN Gang,ZHENG Xiao-Xue.AVO inversion constrained simultaneously in vertical and lateral directions[J].Chinese Journal of Geophysics,2017,60(1):271-282. |
| |
| Authors: | HUO Guo-Dong DU Qi-Zhen WANG Xiu-Ling CHEN Gang ZHENG Xiao-Xue |
| |
| Institution: | 1. School of Geosciences, China University of Petroleum(East China), Qingdao 266580, China;
2. Sinopec Geophysical Research Institute, Nanjing 211103, China;
3. Exploration and Development Research Institution, PetroChina Xinjiang Oilfield Company, Karamay 834000, China |
| |
| Abstract: | The prior probability distribution constraint, which is based on the Bayesian principle, is widely used in AVO (amplitude versus offset) inversion to make the inversion process well-posed. It assumes that the parameters of each trace obey some kind of prior probability distribution in vertical direction. And then the Bayesian principle is used to build the posterior probability density function and to maximize it to obtain the best inversion results in the probability theory sense. However, we cannot obtain satisfying inversion results just depending on the Bayesian constraint. We think that one of the reasons for this less-than-ideal inversion results is that the Bayesian constraint works only in vertical direction but ignores the lateral continuity of parameters.
To increase the accuracy and lateral continuity of inversion results, a new AVO inversion method is proposed to constrain the parameters in both vertical and lateral directions by combining the Bayesian principle and Kalman filtering algorithm. We assume that the parameters of two neighboring seismic traces are similar for earth media change gradually in the lateral direction. Based on this hypothesis, we think that the parameters of a seismic trace are determined by two parts:one is the information of parameters contained in the seismic record of that trace, and the other is the similarity with its neighboring trace. Building a mathematical model of the two parts above permits to obtain the measuring equation and the state transition equation. The measuring equation, which is established according to the Bayesian principle, is the maximum posterior probability density function of parameters. The state transition equation and the measuring equation are substituted into the Kalman filtering algorithm frame, and then the Kalman filtering algorithm is used to predict and modify the parameters of each trace. The parameters after predicting and modifying are the final double-direction constrained AVO inversion results.
The Marmousi2 model and actual data are used to test the proposed method. The inversion results show that the double-direction constrained method has better lateral continuity as well as accuracy than that with only vertical direction Bayesian constraint. And the double-direction constrained method is based on prior hypothesis, which does not depend on well logging or horizon information. So this method can be widely used even in the area without well logging or horizon information. |
| |
| Keywords: | Bayesian prior probability constraint Kalman filtering algorithm Double-direction constraint Lateral continuity |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《地球物理学报》浏览原始摘要信息 |
| 点击此处可从《地球物理学报》下载免费的PDF全文 |
|
