基于多重分形的信息融合技术在琼东南盆地深水区油气勘探中的应用

为解决常规地震属性预测储层时效果差、精度低而导致的预测多解性问题，采用基于多重分形的信息融合技术预测琼东南盆地深水区有利储层区域。根据分形谱函数曲线形态提取优势属性，然后采用多重分形技术进行信息融合。融合后的谱函数计算结果显示，多重分形融合的曲线形态不同于任何一种单一属性的谱函数形态，说明信息融合的结果不是任何单一属性能够代替的。多重分形考虑了属性与属性之间的非线性边界问题及地震属性的非线性特征差异，有效地刻画了局部细节。将该技术应用于琼东南盆地陵南斜坡区崖城组储层预测，结果表明，有利储层区域分布于陵水29A、陵水30A、陵水29B、陵水35A和陵水35B区块，从属性融合结果得知，通过多重分形法圈出的这5个有利储层区域的储层优于永乐1区块，同时结合该区石油地质条件的综合分析，最后确定陵水29A、陵水30A和陵水29B区块为有利储层目标区，为下一步的靶区优选提供了依据。

Application of multi-fractal information fusion in oil and gas exploration in the deep water area of Qiongdongnan Basin

In order to solve the problem of multi-solution prediction caused by poor performance and low precision in predicting reservoirs with conventional seismic attributes, multi-fractal information fusion is used to predict favorable reservoir areas in the deep water area of Qiongdongnan Basin. The dominant properties are extracted according to the shape of the fractal spectrum function curve, and then the multi-fractal technique is used for information fusion. The results of the spectral function calculation after fusion show that the curve shape of multi-fractal fusion is different from the spectral function form of any single attribute, indicating that the result of information fusion cannot be replaced by any single attribute. Multi-fractal method considers the nonlinear boundary problem between attributes and attributes, and the nonlinear characteristic difference of seismic attributes, effectively depicting local details. The technology is applied to the reservoir prediction of the Yacheng Formation in the Lingnan slope area of the Qiongdongnan Basin. The results show that the favorable reservoir areas are distributed in Lingshui 29A, Lingshui 30A, Lingshui 29B, Lingshui 35A and Lingshui 35B. The results show that the reservoirs of the six favorable reservoir areas circled by the multi-fractal method are better than Yongle 1 area. On top of that, together with the comprehensive analysis of the petroleum geological conditions in the area, it is considered that Lingshui 29A, Lingshui 30A and Lingshui 29B are favorable reservoir targets, which provides a basis for the next target area optimization.