一种稳健的弹性阻抗反演方法

发布时间：2018-10-11 15:22

【摘要】：利用弹性阻抗分解,从弹性阻抗体中可以直接提取与储层性质更为密切的弹性参数体。常规弹性阻抗分解算法是基于最小二乘原理,通过求取弹性阻抗分解矩阵的广义逆提取弹性参数,由于该分解矩阵的条件数较大,导致在含噪声的情况下无法获取稳定的弹性参数。结合贝叶斯定理,引入弹性参数的先验分布,构建弹性阻抗分解的正则化项,从而有效提高了弹性阻抗分解的稳定性;考虑到弹性阻抗分解是个带限过程,并且同一时间采样点的弹性参数之间并非独立不相关,因此引入多变量高斯分布描述弹性参数的自然对数更为合理;最后结合基于低频软约束的叠后反演形成一套弹性阻抗反演的流程。模型试算和实际数据测试验证了该弹性阻抗反演流程具有较强的稳定性和较高的精度。
[Abstract]:By using elastic impedance decomposition, elastic parameter bodies which are more closely related to reservoir properties can be directly extracted from elastic impedance bodies. The conventional elastic impedance decomposition algorithm is based on the least square principle. The generalized inverse of the elastic impedance decomposition matrix is used to extract the elastic parameters, because the condition number of the decomposition matrix is large. In the case of noise, stable elastic parameters can not be obtained. Combining with Bayesian theorem, the regularization term of elastic impedance decomposition is constructed by introducing the prior distribution of elastic parameters, which effectively improves the stability of elastic impedance decomposition, considering that elastic impedance decomposition is a band-limited process. And the elastic parameters of the sampling point at the same time are not independent of each other, so it is more reasonable to introduce the multivariable Gao Si distribution to describe the natural logarithm of the elastic parameters. Finally, a set of elastic impedance inversion flow is formed based on low frequency soft constraint poststack inversion. The model test and actual data test show that the flow of elastic impedance inversion has strong stability and high accuracy.
【作者单位】：中国石化石油勘探开发研究院;中海油研究总院;
【基金】：国家油气重大专项“陆相页岩油资源和选区评价技术与软件实现”(2016ZX05049001-003)资助
【分类号】：P631.4

【相似文献】