基于神经网络法的逐点渗透率测井解释研究

渗透率参数是储层解释与评价中极其重要的一个参数。采用常规测井解释方法逐点计算单井剖面中各小层的渗透率往往很难达到精度要求。鉴于单井剖面中解释点的地层渗透率与多种测井参数有关以及围岩点测井参数对其的影响，采用了BP神经网络技术通过构建合理的BP网络结构建立起多种测井信息与渗透率之间的逐点非线性预测模型，实现利用测井资料高精度地逐点解释单井剖面的渗透率。利用该模型处理了T3井等井的测井资料，逐点计算的渗透率不但与取心段的岩心渗透率较为一致，而且非取心段的处理结果令人满意。该法为测井解释地层渗透率参数找到了一条新的途径。

THE STUDY ON PREDICTING EACH POINT'S PERMEABILITY BASED ON NEURAL NET WORK AND LOG DATA

Permeability is an important parameter in reservoir interpretation and evaluation. It is very difficult to calculate this parameter from point by point conventional log interpretation for the layers in a single well section. Based on the relation between permeability and each log variable at each point on surrounding rock, this paper puts forward a predicting model, a non-linear relation between logging information and permeability by means of BP neural network, to calculate each point's permeability at high accuracy. The model is used to process the log data from well T3. The application result is satisfied. The method provides a new approach for permeability interpretation from logging data.