应用BP神经网络预测磨溪气田香四储层孔隙度

Application of BP Neural Network to Calculation of Reservoir Porosity of Moxi Gas Field

神经网络是一门新兴的信息处理技术，它可用来解决测并解释和油藏描述中的模式识别和参数估算等问题。本文利用取心井的储层孔隙度与测并数据，应用改进的ＢＰ神经网络模型建立了川中磨溪气田香四储层物性参数孔隙度的预测模型。与传统方法～回归方程、灰色方程和测井解释相比，其精度及实际预测效果均令人满意。该法值得推广应用。

Neural network is a new information processing technique, which can be used to solve the problems of log interpretation, pattern recognition and parameter estimation in reservoir description. Using the formation porosity and logging data of cored well of Moxi gas field, we have successfully applied Back Propagation Network to the establishment of a prediction model of calculating formation parameters-porosity. Compared with the traditional methods-Regression equation, Grey equation and log interpretation, its precision and predicted result are rather satisfactory. Thus this method is well worth popularizing.