GeoEast软件特色技术在盆1井西凹陷北东环带砂质碎屑流储层预测中的应用

Application of GeoEast's characteristic techniques to prediction of sandy debris flow reservoirs in Northeast Ring of West Sag of Well Pen-1

砂质碎屑流储层普遍非均质性较强,纵、横向变化快,储层预测难度大。因此,需要精细描述砂质碎屑流储层,以拓展油气勘探新领域。利用多属性融合技术预测砂质碎屑流储层面临属性优选,很难得到明确的解释结果。为此,利用GeoEast软件的核主成分属性优化技术与神经网络反演技术定性与定量预测砂质碎屑流储层,落实砂质碎屑流储层发育区。首先,分析砂质碎屑流地震响应特征,利用核主成分压缩技术定性预测砂质碎屑流;其次,利用敏感曲线(GR曲线)对砂质碎屑流储层的敏感性,利用神经网络反演定量预测砂质碎屑流有效储层的分布范围。结果表明,利用神经网络反演结果在平面上确定了6个砂质碎屑流砂体,与钻井结果匹配较好。

It is difficult to predict sandy debris flow reservoirs because of their strong heterogeneity and rapid vertical and horizontal changes.Therefore,it is necessary to finely describe the sandy debris flow reservoirs to expand the new field of oil and gas exploration.The prediction of sandy debris flow reservoirs by the multi-attribute fusion technology requires attribute optimization,and thus,it is difficult to obtain clear interpretation results.Considering this,we use GeoEast's kernel principal component attribute optimization technology and the neural network inversion technology to qualitatively and quantitatively predict sandy debris flow reservoirs,respectively,to determine the development area of the reservoirs.Specifically,we analyze the seismic response characteristics of sandy debris flow and employ the kernel principal component compression technique to qualitatively predict the sandy debris flow.Given the sensitivity of the sensitivity curve(GR curve)to the sandy debris flow reservoirs,we use neural network inversion to quantitatively predict the distribution range of the sandy debris flow reservoirs.The results show that six sandy debris flow sand bodies are determined on the plane by use of the neural network inversion results,which are in good agreement with the drilling results.