应用CP网络进行岩性识别

为通过测井解决岩性识别问题，引入了具有分类准确、算法简练等优点的CP（Counter-Propagation)网络。在详细介绍CP网络的网络模型和算法的基础上，结合某油田的实际测井资料，进行了CP网络识别研究。应用结果表明：CP网络训练周期短、识别准确率高、不存在收敛问题。通过试验研究得出结论：CP网络完全可以用于解决岩性识别等问题，具有广阔的应用前景。

USING THE CP-NETWORK TO IDENTIFY THE LITHOLOGY

This paper introduces the counter propagation network for lithology identification of logging. The CP network is excellence of exact sorting and concise algorithm. The lithology identification based on logging data of an oil field wis carried out by using the CP network's algorithm. Our practice result indicates that the CP network's training period is short, the accuracy of identification is high and it wouldn't be fazed by the uneasy convergence problem. It can quickly identify the lithology and will have a very wide application field.