摘要
在录井资料解释评价中岩性分类是一项必不可少的工作,利用测井曲线数据来实现精准的岩性识别对储层性质的评价至关重要。根据实际储层的非线性关系,作者基于k-means算法建立岩性分类模型,以松辽平原某X井测井曲线和岩性解释数据为样本进行聚类。在这个模型中,首先对选取井的数据进行数据预处理(数据清洗和数据标准化处理);然后,采用k-means模型进行聚类,并将模型聚类结果进行人工标签转换;最终,使用岩性剖面解释数据对模型分类结果进行验证。实验结果表明,本文提出的岩性识别模型岩性识别平均准确率达到87.3%,对于一些常见的岩性识别准确率高达96%,达到人工多次校验结果。
Lithological classification is an essential task in the interpretation and evaluation of logging data,and it is very important for the evaluation of reservoir properties to realize accurate lithological identification using logging curve data.According to the nonlinear relationship of the actual reservoir,this paper establishes a lithology classification model based on the k-means algorithm,and uses the logging curve and lithology interpretation data of a certain well X in Songliao Plain as samples for clustering.In this model,data preprocessing(data cleaning and data standardization)is performed on the data of the selected wells;then,the k-means model is used for clustering,and the model clustering results are manually converted into labels.Sex profile interpretation data are used to validate the model classification results.The experimental results show that the average lithology recognition accuracy of the lithology identification model proposed in this paper reaches 87.3%,and the accuracy rate for some common lithology identification is as high as 96%,reaching the results of multiple manual verifications.
作者
高雅田
范雄
GAO Yatian;FAN Xiong(Computer and Information Technology School,Northeast Petroleum University,Daqing 163318,China)
出处
《微型电脑应用》
2022年第8期113-115,共3页
Microcomputer Applications
