多点地质统计学相控地震同时反演方法

Reservoir stochastic modeling is progressed from two-point geostatistics to multi-point geostatistics. However, the seismic inversion methods are mainly using the two-point geostatistics;Furthermore, the facies-controlled seismic inversion is always updating the elastic parameters but remaining the facies unchanged during the iteration. For giving a solution to this scientific question, This project will develop a new seismic inversion method based on multi-point geostatistics to update the elastic parameters and facies simultaneously. The whole replacement of faices in local zone defined by the search template is realized by the improved multi-point geostatistics Snesim.The sampling of elastic parameters is from the statistic distribution of the different facies-elastic models , which gives the solution of facies –controlled seismic inversion. The facies model is revised by the method of permanence of updating ratio to account for the elastic parameters from the previous iterative results,which a simultaneous inversion of both facies and elastic parameters is realized.The multiple traces comparison between the synthetic and real seismic trace is executed by the data template in the multi-point geostatistical realization to get a multi-traces simultaneous inversion, which will improve the precision of the inversion. The storage and access of the multi-point probability are studied for a rapid inversion. The theoretical and real model are designed to show the validity, efficiency and influence factors of the new facies-controlled seismic inversion method. The study will promote the seismic inversion from two-point geostatistical inversion to facies-controlled multi-point geostatistical inversion and enrich the theory and methods, which will give a new choice in real reservoir inversion.

油气储层建模已经由两点统计发展为多点统计学建模，然而在地震储层反演中仍然以两点统计学反演方法为主。此外，沉积相约束地震反演仅更新弹性参数，未能同时更新沉积相。本项目拟针对沉积相约束地震资料反演科学问题，研究多点统计反演中沉积相模型迭代更新及其约束于弹性参数同时反演的方法。通过改进的Snesim多点建模方法预测沉积相；利用沉积相---弹性参数统计关系约束弹性参数抽样；结合更新比率恒定理论将每次反演的弹性参数用于后期迭代过程中沉积相模式预测，从而实现迭代过程中弹性参数和沉积相的同时更新；通过多个地震道同时计算和对比，实现多道同时反演，提高反演预测精度；完善多点统计概率存储与访问策略，提高反演运算效率；利用理论模型和实际模型，对算法的有效性、运算效率、影响因素等进行评价。研究将推动地震储层反演由两点统计向多点统计相控同时反演方向发展，丰富地震储层反演理论和方法，并为地震资料储层预测提供技术支持。

本项目针对地震资料反演中考虑沉积相约束不足的问题，研究在多点地质统计反演迭代中增加沉积相-弹性参数相互约束，形成全新的多点地质统计学相控地震反演方法，推动地震反演从两点反演向多点统计反演进步，从而提高地震反演以及地质模型的精度。项目通过4年的攻关，完成了既定任务，取得了6个方面成果。①通过露头、现代沉积、水槽实验等获得储层原型模型，利用理论岩石物理参数以及子波，通过正演获得测试用训练图像库6套。②引入更新比率恒定理论，实现了多点地震迭代过程中沉积相与弹性参数互相约束，达到沉积相-弹性参数同时反演且一致的目的。完成了软件模块1套。③通过多道比较，实现了地震反演从单点估计到多道预测的突破，反演结果表明弹性参数连续性更好，地质体几何形态与叠置样式更符合实际。④提出并开发了利用二维地质剖面约束三维多点地质统计反演的新方法，通过斯坦福大学理论测试模型证明其反演误差小于两点统计与常规的多点地质统计反演，模拟计算效率高，并实现了地质基础研究到地震反演的一体化。值得进一步推广。⑤理论模型测试表明基于概率更新理论的迭代抽样30-50次效果与MCMC抽样效果几乎一致，但反演效率提升数百倍。⑥通过中石化新场油田三维储层模型反演对比结果表明，相比较于传统的两点地质统计反演，多点地质统计反演显著提高了反演精度，模拟误差降低6%以上。基于更新比率恒定理论的多点地质统计反演较MCMC多点统计反演效果略差，但比传统的基于Simpat的多点统计反演精度要高。研究丰富了地震反演理论，有力推动了地震反演从传统的基于两点地质统计反演方法向基于多点地质统计反演转变，提高了地震反演精度并服务油田生产开发。研究成果丰富，发表学术论文13篇，其中SCI收录10篇。获得授权美国发明专利1件，国内授权发明专利7件。获得湖北省科技进步一等奖1项，中国石油和化学工业联合会科技进步一等奖2项。培养博士生1名，硕士研究生11名。