储层横向预测油气的一种模式识别法

PATTERN RECOGNITION FOR LATERAL OIL/ GAS RESERVOIR PREDICTION

本文分析了多种地震信息与含油气性之间的模糊关系,建立了各地震信息的隶属函数和模糊子集合。通过对储层各预测“点”与已知井模式贴近度的计算,按择近原则分类,预测了高产、中低产和非产区,为钻探提供可靠的定井资料。文中以ZC地区的6口井为例,用这种识别法对有关地震信息进行了检测;同时,观测了该地区15个参数点的含油气情况,取得了较好的效果。

In this paper, the fuzzy relationship between a great deal of seismic information and the state of oil / gas bearing was investigated, the subordinate functions as well as the fuzzy subgroups relevant to different seismic information were set up. According to the calculation for each 'PREDICTION POINT* on reservoir and the approximant fitted to the standard pattern of a known well and in accordance with the principle of 'BEST FITTINGS', areas of high-capacity, medium capacity and low capacity, as well as the area without expection are all predicted. Reliable data for well-location were porvided at the final. Take six wells in ZC district as examples, the 'might be' state of oil / gas bearing for 15 points were predicted with the technique recommended. Satisfactory results were achieved.