基于主成分自组织神经网络法的测井曲线分层技术

Lithological Stratification from Logging Curves Based on Principal Component Self-Organizing Neural Network Method

在砂岩型铀矿找矿工作中,提高测井岩性分层效率和精度至关重要。为提高砂岩型铀矿岩性分层效果,本文采用主成分分析法对多个测井曲线进行降维处理,将主成分分析法的第一主成分、第二主成分、第三主成分作为自组织神经网络的样本数据,进行自组织神经网络训练,将训练好的网络模型用于砂岩型铀矿岩性的自动化分层。实验结果显示:主成分自组织神经网络法岩性分层精度可达到85%以上,高于传统自组织神经网络算法78%的分层精度,具有更好的测井岩性分层效果。因此,主成分自组织神经网算法的岩性分层方法有效减少了输入样本的种类,简化了自组织神经网络结构,其自动化分层效果要优于传统的自组织神经网络算法。本文的研究结果表明,主成分自组织神经网算法在砂岩型铀矿领域岩性识别工作中具有较好的应用效果。

It is crucial to improve the efficiency and accuracy of lithological stratification from logging curves in the exploration of sandstone type uranium deposits.In order to improve the lithological stratification effect of sandstone type uranium deposits,this work conducted principal component analysis to reduce the dimensionality of multiple logging curves.The first principal component,second principal component,and third principal component of the principal component analysis method were used as sample data for self-organizing neural network training.The trained network model was used for automated stratification of sandstone type uranium deposits.The experimental results show that the principal component self-organizing neural network method has a lithological stratification accuracy of over 85%,which is higher than the traditional self-organizing neural network algorithm's stratification accuracy of 78%,and has better logging lithological stratification effect.Therefore,the lithological stratification method of principal component self-organizing neural network algorithm effectively reduces the types of input samples,simplifies the structure of self-organizing neural network,and its automated stratification effect is better than traditional self-organizing neural network algorithms.The research results indicate that the principal component self-organizing neural network algorithm has good application effects in lithological identification of sandstone type uranium deposits.