三角洲前缘砂体骨架的随机模拟

Stochastic modeling of sand framework of deltaic front

鄂尔多斯盆地三叠系延长组长6段、长7段发育于大型正常河流三角洲前缘和湖相沉积环境。在陕西省延长县延河北岸的谭家河剖面,精细观察长6段、长7段出露较好的11条测线(间距为40m),建立原型模型和相应的地质知识库。分别将测线抽稀,间距为80m、120m、160m,用抽稀后的测线作为约束条件,采用顺序指示随机模拟方法,进行三角洲前缘砂体骨架随机模拟,每次抽稀模拟产生50个实现。将模拟实现的概率分布与实际岩性分布对比,结果表明,抽稀后建立的骨架模型是合理的。利用露头的原型模型建立地质知识库,在对井距较大的油藏建模时,利用原型模型给出不易求准的模拟参数,是建立储集层预测模型的有效手段。

Prototype model with 40 meters surveying line spacing and corresponding geological knowledge database are established according to the fine study on the outcrops of Chang 6 and Chang 7 stratum, which are emerged well in the Yanhe profile located in Ordos Basin. The reservoir framework model of Tanjiahe profile(key profile)is built under condition of not using some surveying lines, with a spacing of 80,120 and 160 meters, respectively. The stochastic simulation arithmetic used is sequential indicator simulation and 50 realizations are produced. The comparing with simulated and original profile by using layer number, average thickness and percent of thickness of different lithology as criterions, it is clearly that the model is reliable and has high forecasting precision and is a practical and useful tool for establishing a forecasting model of reservoir.