基于多元线性回归分析的薄储层预测技术在胜利探区的研究与应用

Research and Application of Thin Reservoir Prediction Techniques Based on Multiple Linear Regression in Shengli Exploration Area

随着胜利油田油气勘探的逐步深入,薄储层成为重要的勘探方向之一,但其预测也一直是个勘探难题。因此针对其识别,开发了多元线性回归薄储层预测技术,研发了油藏表征系统软件。该方法在多种属性优选的基础上,建立了储层厚度与伪熵、混沌等优势地震属性间最优化定量计算关系,刻画了薄储层在平面上的分布趋势,进而掌握储层厚度分布特征。在阳信洼陷沙一段滨浅湖薄层砂体预测中,该项技术的应用取得了良好的勘探效果,较好地反映了生物灰岩储层平面上的展布特征,对胜利探区其它类型的薄储层油藏勘探,起到了很好的借鉴和参考意义。

Thin reservoirs become one of the important exploration directions with the grad- ual deepening of the oil and gas exploration in Shengli Oilfield, although its prediction has always been an exploration problem. Therefore, thin reservoir prediction techniques of multiple linear regression was developed for its identification, and reservoir characterization system software was also researched and developed. On the basis of a variety of attributes preferred, the method established the optimal quantitative calculation relationship between the reservoir thickness and the preferred seismic attributes such as the pseudo entropy, chaos etc.. It characterizes the thin reservoir distribution trends on the plane, thus the res- ervoir thickness distribution characteristics was mastered. The application of this technolo- gy in the prediction of shore -- shallow lake thin -- layer sandbodies of member1 of Shahe- jie Formation in Yangxin Sag has made good exploration results. It shows the distribution characteristics of the biological limestone reservoirs preferably on the plane, and provides a good reference for the thin reservoirs exploration of other types in Shengli exploration area.