摘要
论述了用人工神经网络系统建立储层四维地质模型的原理和方法。具体步骤包括子模型的建立、子模型预测结果的合成、储层参数MIMO预测模型的建立、数据体的准备、输入信号的选取、模型结构与检验及模型的工作方式等。人工神经网络用于地质建模时 ,它可以把时间作为一个输入变量 ,可以较好地描述、研究储层参数在四维空间的特征和空间展布 。
The basic principle and method of 4D geological modeling using Artificial Neural Network System (ANNS) are discussed in the paper. The concrete procedures include setting up submodel, synthesis of predictive results from the submodel, building MIMO prediction model of reservoir parameter, preparing data pool, selecting input signal, model structure, its testing and working pattern etc. When ANNS is employed in geological modeling, time is regarded as an input variable, it can better describe and study the features and the distribution of reservoir properties in 4D space, then effectively predict the remaining oil distribution.
出处
《油气地质与采收率》
CAS
CSCD
2004年第3期4-6,共3页
Petroleum Geology and Recovery Efficiency
基金
国家重点基础研究发展规划项目"大幅度提高石油采收率基础研究"(G19990 2 2 5 0 9)资助
关键词
人工神经网络系统
储层
四维地质模型
石油地质
地质体
artificial neural network system, 4D geological reservoir model, reservoir parameters change, 4D multimodeling, MIMO model for reservoir parameters
