低渗储层叠前地震反演方法研究

Low permeability reservoirs account for more than half of new found reservoirs,it has great potential of development, but reservoirs are complicated and difficult to be exploited. In order to get a set of prediction and recognition technology about the loe permeability reservoirs, based on the analysis about low permeability reservoir, this project proposed a prestack seismic inversion method which effectively integrate inversion method based on bayesian theory and stochastic simulation besed on geological statistics. When processing the nonlinear prestack inversion problem based on bayesian theory, select the pre-conditioning matrix that corresponding to low permeability reservoir as constraint to carry out the optimization process, not only can effectively improve the efficiency of the computation of the solution, but also enhance the stability and reliability of the inversion solution; the integration of this inversion result with stochastic simulation based on geological statistics can break the limits of seismic frequency width and get high precision impedance result, with gamma、porosity and other non-impedance data, the inversion of lithology parameters ,such as porosity, can be processed, which can get more diverse underground parameters for the prediction of low permeability reservoir. These results can improve reservoir prediction accuracy, strengthen the cognition of reservoir engineers to subsurface reservoir and enhance oil recovery. It can provide technical support and guarantee for the exploration and development of low permeability reservoir and it has great meaning for the prediction and characterization of low permeability reservoir.

低渗透油气藏占新发现油气藏的一半以上，开发潜力巨大，但储层复杂、勘探难度大，针对以上问题，以获得一套面向低渗储层的预测和识别技术为出发点，本项目在分析低渗储层特性的基础上，形成有效融合基于贝叶斯理论的优化反演方法和基于地质统计学的随机模拟为一体的叠前地震反演方法。选取与低渗储层特征相吻合的预处理矩阵为约束条件，开展基于贝叶斯理论的叠前非线性反演的优化处理，不仅可有效提高解的运算效率，还可增强反演解的稳定性和可靠性；把该反演解与基于地质统计学的随机模拟相结合可突破地震频宽的限制，获得高精度的波阻抗数据体，借助伽玛、孔隙度等非波阻抗资料，还可进行孔隙度等岩性参数的反演，为低渗储层的预测提供更加丰富多样的地下参数信息，有效提高储层预测精度，加强油藏工程师们对地下目的储层的认识程度，提高采收率，为低渗储层的勘探和开发提供技术支持和保障，对低渗储层的预测和描述具有重要意义。

针对低渗储层复杂、勘探难度大的问题，为减小低渗油气田的勘探开发难度，提高低渗储层的勘探效率，本项目估算了低渗储层井位置处横波速度信息，进行了低渗储层叠前地震反演方法研究。首先，本项目在分析低渗储层特性的基础上，基于岩石物理基础理论模型构建了低渗储层等效岩石物理模型，进行了饱和岩石模量的求取，估算了与低渗储层特征匹配的横波速度，为后续低渗储层地震资料的叠前反演提供横波资料。然后，基于贝叶斯理论建立了正则化先验约束的低渗储层参数反演的目标函数，通过引入正演和共轭算子来模拟反演算法的正反演问题，以共轭解作为初始解，实现了储层参数叠前地震反演目标函数的非线性求解。同时，为突破地震带宽的限制，在保留常规随机反演方法优点的基础上，本项目引入了傅里叶滑动平均谱模拟进行频率域的地质统计模拟，有效提高了计算效率，并利用逐步变形优化算法进行扰动更新，确保模拟结果与实际地震数据的匹配，构建了基于地质统计先验信息的弹性参数随机反演方法。最后，基于最大后验概率估计理论，将储层参数的非线性反演结果与基于地质统计学的随机反演相融合，利用Metropolis抽样获得了可靠的高精度储层参数数据体，能够有效提高储层的预测精度，为低渗储层的勘探和开发提供高效可靠的技术支持。