人工神经网络在测井资料参数估算中的应用

Application of artificial neural networks for well-log informatiom estimation

利用人工神经网络(BP网络)的原理对测井资料进行了解释.建立了测井解释的BP神经网络模型,对BP网络进行了改进,提高了收敛速度,并防止陷入局部极小.利用多种测井数据及岩芯描述资料作为网络模型的学习样本,通过网络的学习、训练,建立了测井解释的神经网络模型.以渗透率的计算为例,应用此模型定量计算了川西某气田多口井的渗透率值,其解释结果及精度均令人满意,取得了良好的实际应用效果.

The principle of artificial neural network is utilized to interpret the well-log information and establish the model of ANN for interpretation of well-log information.The BP is improved for convergence speed and to avoid local optimization.The advanced neural network is employed to establish the network model for log interpretation.Various log data and core information are used in interpretation model of neural network by practice.Log data of several gas wells in west Sichuan area are interpreted by this model with accuracy and good application.