论文部分内容阅读
正确识别岩性对储层参数的精确计算及流体识别工作具有重要意义.针对复杂岩性储层岩性难以准确判别,考虑到核极限学习机可收敛到全局最优解,将核极限学习机进行改进,提出基于归一化核极限学习机的岩性识别方法.通过对中东伊拉克M油田复杂岩性储层579块岩样进行建模,然后对未参与建模的井进行岩性识别,核极限学习机模型预测准确率达到80.03％,归一化核极限学习机模型不仅在预测准确率达到81.85％,且预测速度仅有0.001 1s,在预测准确率与速度上均优于传统主流模型.“,”The accurate identification of lithology is of great significance to the accurate calculation of reservoir parameters and further fluid identification.In view the of fact that the lithology of complex lithology reservoir is difficult to discriminate accurately,after improving the kernel extreme learning machine,the lithology identification method based on the normalized kernel extreme learning machine is proposed due to the consideration that kernel extreme learning machine can converge to the global optimal solution.After the modeling of 579 rock samples from complex lithology reservoirs in M oilfield,Iraq,Middle East,lithology identification was performed for the wells not involved in modeling.The prediction accuracy rate of kernel extreme learning machine is 80.03％.For normalized kernel extreme learning machine,not only the prediction accuracy rate reaches 81.85％ but also it takes only 0.001 1 seconds,which shows that it has advantages in the accuracy and speed compared with the traditional mainstream model.