井间地层分形描述

用线性插值和分形插值方法建立了井间参数精细剖面模型。在相同的井点值下，修改随机数发生器的种子值，完成分形随机模拟的多次实现，然后用黑油模拟器对不同剖面模型进行动态模拟。对生成的精细剖面图进行了对比，结果表明，分形随机建模能同时考虑地层参数分布的结构性和间歇性，而分数高斯噪声（ｆＧｎ） 比分数布朗运动（ｆＢｍ） 更适合描述沉积环境变化剧烈、随机干扰强的地层。用多次实现的分形随机模型进行动态预测，对开发指标可给出最好、最坏和最可能的估计。间歇指数对ｆＧｎ 法所建模型的动态预测结果影响较大，对ｆＢｍ 影响较小， 在两端注、采井工作制度相同的情况下，用不同的间歇指数建立的ｆＢｍ 分形剖面模型的最终采收率基本不变

FRACTAL DESCRIPTION OF FORMATION BETWEEN WELLS

Fractal geostatistics and linear interpolation are used to generate highly detailed fractal distributions of porosity and permeability for vertical cross sections. By adjusting the value of random seed, the multiple realizations of fractal models are performed. A black oil simulator is used to predict the performance of all models. Both spatial structure and intermittence of property distributions can be taken into account in fractal models. The fractional Gaussian noise (fGn) method is more effective than fractional Brownian motion (fBm) method for describing the formation with more uncertainty. Flow simulations in multiple realizations of fractal models provide three kinds of dynamic predictions: the best, the worst and the most probable ones. Intermittency exponent influences the performance of fGn models more than that of fGm models.