基于因素神经网络的混合型油藏数模辅助专家系统

介绍一种基于因素神经网络的混合型知识模式,在油藏数值模拟历史拟合辅助专家系统中的实现,体现了用多种人工智能手段协同解决复杂问题这一主流技术的应用。介绍系统的功能和分布式CLIENT/SERVER工作环境。分析问题的复杂性,提出相应的知识表示和混合推理模式(树型,解析型和模拟型因素神精网络)等问题求解策略。并描述了推理机制和知识库结构。

Assisted Expert System for Mixed Reservoirs Numerical Simulation based on Factor Neural Network

This paper presents a method to realize the mixing knowledge model based on the factor neural network on the history matching of reservoir simulation. It reflects a use of main technique solving complex problem with multi artificial intelligence methods. Some functions of the system and the working environment of CLIENT/SERVER distribution are given. The complexibility of the problem are analyzed. Some tactics are put forward to solve the problems about theses knowledge representations and the mixing inference models (tree, analytic and simulation inference model), some inference mechanisms and knowledge structures are described.