利用五维数据直接提取裂缝型储层参数

现有裂缝型储层流体识别方法在构建流体敏感参数过程中没有考虑岩石骨架、裂缝、孔隙及油气水耦合影响，对五维数据中丰富的方位信息蕴含的流体信息挖掘不够，并且针对裂缝型储层油气识别以间接计算参数为主。宽方位地震五维数据反演本身存在“病态解”及“不确定性”等问题，多裂缝参数同时反演更易导致储层参数反演精度下降。因此，基于间接代数组合的流体因子在计算过程中不可避免地累积计算误差，影响油气识别效果。为此，结合各向异性Gassmann孔隙弹性理论和线性滑动理论，以“固液缝”解耦流体因子为基础，推导了由“固液缝”解耦流体因子直接表征的新的纵波反射系数方程，分析、讨论了新纵波反射系数方程的适用条件及敏感性；在此基础上，构建贝叶斯框架的地震反演目标函数，并基于时频资料研究五维数据流体识别方法，可直接获得裂缝型储层“固液缝”解耦流体因子，消除了累计误差的影响，提高了预测精度。模型试算及实际资料处理表明，所提方法可以合理、可靠地预测裂缝型储层油气分布。

Direct extraction of fractured reservoirs' parameters based on five-dimensional data

The existing fluid identification methods for fractured reservoirs do not consider the influence of rock skeleton, fractures, pores, and oil-gas coupling in determining sensitive parameters of the fluid and fail to fully excavate fluid information contained in a wealth of azimuth information in five-dimensional (5D) data. In addition, oil and gas identification of fractured reservoirs is mainly based on indirectly calculated parameters. The inversion of wide-azimuth 5D seismic data has problems of ill-conditioned solution and uncertainty, and simulta-neous inversion of multiple fracture parameters is more likely to make the inversion accuracy of reservoirs' parameters decrease. Therefore, the fluid factors based on indirect algebraic combination inevitably accumulate calculation errors during calculation, which thus affects the effect of oil and gas identification. In view of the above problems, by combining the anisotropic Gassmann theory of pore elasticity and linear sliding theory, a new P-wave reflection coefficient equation directly characterized by the fluid factor decoupled by solid-fluid-fractures is derived. Moreover, the applicable condition and the sensitivity of the equation are analyzed. On this basis, the seismic inversion objective function with Bayesian framework is constructed, and the fluid identification method by 5D data is studied based on the time-frequency data, which thus directly obtains the fluid factor decoupled by solid-fluid-fractures in fractured reservoirs, eliminates the influence of cumulative errors, and improves the prediction accuracy. Research on models and actual data show that the proposed method can reasonably and reliably predict the distribution of oil and gas in fractured reservoirs.