多资料匹配融合储层建模技术

Reservoir model building with multi-data fusion

多资料匹配融合储层建模技术,是在贝叶斯理论框架下,将测井、三维地震及井间地震等多种资料进行优势互补的油藏建模新技术,可有效提高稀井区或无井控制区油藏建模的准确性,为中国东部老油田的精细油藏描述提供新的技术手段。该技术的核心算法为贝叶斯—序贯高斯模拟方法,以局部、少量的高分辨率(测井、井间地震等)数据为条件,建立与地震属性数据之间的条件关系实现地震属性的约束,有效利用了测井纵向分辨率高和地震资料横向分辨率高的优点,从而达到提高整个模型分辨率的效果。应用该建模方法建立了垦71区块的孔隙度模型,经对比发现,孔隙度模型井间趋势与约束数据趋势一致,纵向分辨率得到了显著提高,充分体现了多尺度资料联合物性建模方法的技术优势。应用实践表明,该方法模拟结果在井点附近与测井解释数据一致,在稀井或无井控制区的预测结果与地震约束数据一致,并有效提高了储层预测精度。

Based on Bayesian theory,reservoir model building with multi-data fusion takes advantages of logging,3Dseismic and cross-well seismic data to improve the accuracy of reservoir model building.This approach provides a new way for fine reservoir description in the Chinese eastern mature oilfields.The key algorithm of the new approach is Bayesian-sequential Gaussian simulation,which utilizes small amount of high resolution logging and cross-well seismic data as condition data to realizes seismic attribute constraint.This algorithm takes advantages of logging data high vertical resolution and seismic data high horizontal resolution. We build a porosity model of Block K71 with this method.It is found out that in this model crosswell tendency is well matched with constrained data,and vertical resolution is greatly improved.This illustrates advantages of the approach.Applications show that the proposed approach can effectively enhance reservoir prediction.