| 基于概率统计的地震岩相识别不确定性定量评价方法 |
| |
| 引用本文： | 袁成,李景叶,陈小宏.基于概率统计的地震岩相识别不确定性定量评价方法[J].地球物理学报,2015,58(10):3825-3836. |
| |
| 作者姓名： | 袁成 李景叶 陈小宏 |
| |
| 作者单位： | 1. 中国石油大学(北京)油气资源与探测国家重点实验室, 北京 102249;
2. 中国石油大学(北京)海洋石油勘探国家工程实验室, 北京 102249 |
| |
| 基金项目： | 国家自然科学基金项目(U1262207),国家科技重大专项课题(2011ZX05019-006)和中国石油大学(北京)校基金(2462012KYJJ0508)联合资助. |
| |
| 摘 要： | 地震岩相识别能够提供具有不同储层特征的岩相分布信息,对岩相识别的不确定性开展定量评价分析可降低后期油藏建模与储层评价的风险.考虑了地震岩相识别中测井岩相定义、岩石物理建模、井震尺度匹配及地震反演等环节的不确定性对岩相识别的影响,基于概率统计方法,引入熵函数实现了地震岩相识别不确定性定量评价,并结合岩相概率、重建率等多角度综合定量分析不确定性的构成及传递特征,系统地实现了地震岩相识别不确定性评价流程的整体连通.提出了结合属性交绘特征约束反演参数空间,提高地震岩相识别运算效率.模拟数据分析表明利用熵函数可精确实现岩相识别不确定性地定量表征,利用属性交绘特征约束参数空间既大幅度减少运算量,也可降低地震岩相识别的不确定性.
|
| 关 键 词： | 概率统计 熵函数 岩相识别 不确定性 定量评价 |
| 收稿时间： | 2014-10-14 |
Quantitative evaluation of uncertainties in seismic facies identification based on probabilistic statistics |
| |
| YUAN Cheng,LI Jing-Ye,CHEN Xiao-Hong.Quantitative evaluation of uncertainties in seismic facies identification based on probabilistic statistics[J].Chinese Journal of Geophysics,2015,58(10):3825-3836. |
| |
| Authors: | YUAN Cheng LI Jing-Ye CHEN Xiao-Hong |
| |
| Institution: | 1. State Key Laboratory of Petroleum Resources and Prospecting, China University of Petroleum, Beijing 102249, China;
2. National Engineering Laboratory for Offshore Oil Exploration, China University of Petroleum, Beijing 102249, China |
| |
| Abstract: | In the early stage of oilfield exploration, reservoir characterization is always considered to be a risky task, since few well-log data are available at the foremost stage of any production plan in the petroleum industry as well as the ambiguity of seismic data. Uncertainty in seismic reservoir characterization is generally quite large, especially in seismic facies classification which goes through multiple links. Therefore, quantitative uncertainty evaluation is valuable for seismic facies classification. It provides an important guiding sense for risk management as well as decision-making in any petroleum reservoir.
For evaluating the uncertainty propagation in seismic facies classification quantitatively, the uncertainty of well-log facies definition, rock physics modeling, scale change and seismic inversion has been taken into consideration in this case. We firstly compute the facies probabilities conditioned on different properties in each step of seismic facies classification. Then, the associated uncertainty and maximum a posteriori (MAP) of facies probabilities are assessed by means of entropy and reconstruction rate respectively, since the variable that represents facies is categorical. The influence of seismic noise on facies classification is also analyzed by synthetic seismic data with different signal noise ratio. In addition, the parameter spaces of well-log and upscaled elastic properties are restricted by the data distribution characters in cross-plot.
Synthetic example shows that uncertainty in seismic facies classification could be evaluated quantitatively with this methodology. By introducing entropy, the constitution and propagation of uncertainty can be evaluated quantitatively with the help of facies probability. The total flow chart of uncertainty evaluation in seismic facies classification is connected systematically. Furthermore, the influence of uncertainty propagation and seismic noise on facies classification is illustrated visually by entropy and reconstruction rate. Restriction of parameter spaces by the data distribution characters in cross-plot can not only reduce the computational cost but also the uncertainty in seismic facies classification dramatically since the parameter vectors which fall out of the restricted scopes are precluded.
Quantitative uncertainty evaluation brings many details of the uncertainty in seismic facies classification. It makes it possible to assess the propagation and accumulation of uncertainty quantitatively as well as its influence on the result accuracy, since the increment of entropy is considered to be the incoming uncertainty of the current step. It provides us with a unique angle of view to understand the uncertainty in seismic facies classification as well as the great value for risk management and optimal decision-making in the petroleum industry. |
| |
| Keywords: | Probability statistics Entropy Facies identification Uncertainty Quantitative evaluation |
| 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 |
|
