西南石油大学学报(自然科学版) ›› 1995, Vol. 17 ›› Issue (3): 66-70.DOI: 10.3863/j.issn.1000-2634.1995.03.009

• 油气田开发 • 上一篇    下一篇

利用神经网络技术建立岩石强度预测模型

刘向君 罗平亚



  

  1. 1.研究生部 2.完井技术中心
  • 收稿日期:1994-11-02 修回日期:1900-01-01 出版日期:1995-08-20 发布日期:1995-08-20
  • 通讯作者: 刘向君

Rock Strength Prediction Model Established by Nerve Network Technique

Liu Xiang-jun   

  1. (Dept. of Postgraduates)
  • Received:1994-11-02 Revised:1900-01-01 Online:1995-08-20 Published:1995-08-20
  • Contact: Liu Xiang-jun

摘要:

本文利用神经网络技术建立了岩石强度参数的预测模型,该模型的非线性关来非常复杂,是一般计算岩石强度的经念公式关来式及统计关系式难以表达的。在已知网络内部表达的情况下,将不同时刻测得的测井参数值直接代入已训练网络,便能够预测出其对应的岩石强度.计算结果表明,预测效果相当好,而且不需要对中问参数进行确定。因此,神经网络模型为岩石力学参数确定提供了一种新的方法和手段

关键词: 神经网络, 解释模型, 岩石性质, 机械强度

Abstract: In this paper,models which can be used to predict the rock strength parameters have been established by using the nerve network technique. The nonlinear relation of this model is very complex and is difficult to be expressed by the experimental formula and statistical expression to calculate the rock strength. With the internal- expression of the network known,the logging parameters measured at different time are taken into the network
trained. Then the corresponding rock strength can be predicted. The calculation shows that the prediction result is quite good. Besides,the middle parameters do not need to be deter-mined. So the nerve network model provides a new way for the determination of the rock me-chanical parameters.

Key words: Nerve network, Interpretation model, Rock property, Mechanical strength

中图分类号: