摘要
应用自适应算法对BP网络进行改进,可以提高BP网络的收敛速度和全局寻优性能。在此基础上,利用多种测井数据及岩心描述资料作为网络模型的学习样本,以测井解释渗透率的神经网络模型为例,通过网络的学习、训练,建立测井解释神经网络模型。并应用此模型,定量计算出多口井的渗透率值,与常规渗透率计算结果相比,BP的解释结果及精度均令人满意,同时还取得了良好的实际应用效果。
The self-adaptive algorithm is adopted to improve the accuracy and convergence speed of BP networks. Based on this algorithm ,we take the model of permeability as an example to show the establish of neural model and interpretation of the reservoir properties using laboratory data and multiple logging data. The difference between result of conventional method and the BP method indicates that the latter is of better performance.
出处
《物探化探计算技术》
CAS
CSCD
2004年第2期129-132,共4页
Computing Techniques For Geophysical and Geochemical Exploration
基金
国家财政部矿产资源保护项目(KB2000-07-01)
