摘要
循环神经网络已广范应用于测井缺失曲线生成及孔渗饱等物性参数预测。本文以存储式测井自然电位曲线生成为研究背景,构建Y油区数据集,包括训练集、验证集和测试集,分别采用RNN、GRU和LSTM网络,进行了实验研究。通过实验数据的对比和分析,表明在本文的研究背景下,三种循环神经网络的性能基本一致。
The circulating neural network has been widely used in the generation of log deletion curve and the prediction of physical property parameters such as pore infiltration.In this paper,using the natural potential curve as the research background,we constructed Y oil area data set,including training set,verification set and test set,using RNN,GRU and LSTM network respectively.Through the comparison and analysis of experimental data,we show that the performance of the three recurrent neural networks is basically consistent in the context of this paper.
作者
邢强
王海霞
王志美
Xing Qiang;Wang Haixia;Wang Zhime
出处
《国外测井技术》
2023年第6期44-46,2,共4页
World Well Logging Technology
关键词
测井曲线生成
循环神经网络
实验对比
logging curve generation
recurrent neural network
experimental comparison
