So far,the static quantitative data of well logging curves have been mostly used to identify the sedimentary microfacies in the existing oil fields,which is difficult to reflect the influence of logging depth accumulation on pattern recognition of sedimentary microfa-cies. Taken into the above shortcomings account,a discriminant model combining fuzzy inference and process neural network is estab-lished and proposed to realize the judgment of sedimentary microfacies,on the basis of selection of information on quantitative and quali-tative mixing process in the logging curve,which has been quantitatively processed to simplify the discriminant rule and extract the valid discriminant data so as to improve the accuracy of the sedimentary microfacies discrimination. Considered that the logging data is charac-teristic of the curve with depth the advantage in process input of the process neural network is introduced and then accuracy of sedimenta-ry microfacies discrimination by continuously optimizing the learning mechanism of process neural networks is improved. The experimen-tal results show that it has high accuracy and high rate,which is a practical method for the identification of sedimentary microfacies.%迄今,现有的油田进行沉积微相模式识别时大多选取测井曲线的静态定量数据,其难以反映测井相的深度累积效应对沉积微相模式识别的影响.针对上述不足,选取测井曲线中可处理的定量与定性混合过程信息,构建并提出了模糊推理和过程神经网络相结合的判别模型,以实现沉积微相的判别.该模型基于模糊集理论对测井相的定性信息进行定量处理,以简化判别规则,并提取有效的判别数据,从而提高沉积微相判别的精度;根据测井相数据随深度变化的特征曲线,采用过程神经网络的过程式输入优势,通过不断优化过程神经网络的学习机制来提高沉积微相判别的准确度.实验结果表明,基于模糊推理过程神经网络模型的沉积微相模式识别方法精度高、速度快,是一种比较实用的沉积微相识别方法.
展开▼
