论文部分内容阅读
非均质地层、储集层的物性参数级差较大，用传统统计方法很难准确建立储集层参数与测井响应之间的关系，而神经网络技术通过多次非线性映射可对二者关系达到最佳逼近。作者论述了BP网络的基本原理及其改进算法，并以克拉玛依某砾岩油藏为例，说明了神经网络储集层参数建模过程及应用Matlab工具包实现的方法，用该方法建立的储集层参数解释模型取得了较理想的结果。
Heterogeneity strata, reservoir properties of large parameters of the level difference, the traditional statistical methods is difficult to accurately establish the relationship between reservoir parameters and logging responses, and neural network technology through multiple nonlinear mapping can both Relationship to achieve the best approximation. The author discusses the basic principle of BP network and its improved algorithm. Taking a Karamay conglomerate reservoir as an example, this paper illustrates the process of neural network reservoir parameter modeling and the method of using Matlab toolkit to realize. By using this method, The layer-by-layer parameter interpretation model has achieved better results.