摘要
为实现油田生产管理和决策的现代化,使地层参数估值具有全局最优性,在研究油井井底压力分布的描述和有关地层参数辨识问题的基础上,提出了一种由二阶学习算法与GA(Genetic A lgorithm)构成的新型混合遗传算法,并给出一种新型神经网络。该网络把级数中的函数看成非线性神经元,建立油藏系统的函数型连接人工神经网络模型。由系统辨识理论中的F检验法确定网络模型的结构参数n,用二阶学习算法和新型GA交替辨识网络模型的权系数v和地层参数θ。应用表明,采用上述方法建模精度高,模型的平均相对误差在1%以内,并能求出地层参数的全局最优估值。
In order to realize the modernization of management and decision making for oil field, stratigraphic pressure distribution model and stratigraphic parameters identification are researched. To request global optional estimate values of stratigraphic parameters, a hybrid genetic algorithm consisting of second order learning algorithms and GA （Genetic Algorithm） is developed, and a new neural network is given. In the net, the functions are served as nonlinear neural units to establish function link artificial neural networks models of oil reservoir systems. The F-test in system identification theory is applied to determine the structure parameter n, and a second order learning algorithms and novel GA are used to identify the weighting coefficients of the networks v and θ respectively. It has high precision that our method is applied to build models of above systems, the average relative errors are within 1%. Moreover it can obtain global optimal estimate values of stratigraphic parameters.
出处
《吉林大学学报（信息科学版）》
CAS
2006年第6期624-628,共5页
Journal of Jilin University（Information Science Edition）
基金
黑龙江省自然科学基金资助项目(TF2005-26)
关键词
系统辩识
函数型连接神经网络
二阶学习算法
遗传算法
收敛性
system identification
functional link nets
second order learning algorithms
genetic algorithms
convergence
