摘要
油田开发优化是一种计算密集型任务,需要大量的油藏模拟,这种模拟非常昂贵,尤其对于大宗复杂的油藏模型。替代模型在历史拟合、生产优化和预测过程计算量少、成本低,能有效地评估新方案的目标函数值和大幅度减少计算过程中反复调用模拟模型所引起的庞大的计算负荷,从而加速油田开发优化过程。本文主要介绍常见的四种替代模型:多项式回归模型(PRM)、泛克里金模型(MKG)、薄板样条插值(TPSM)、人工神经网络(ANN)。
Oil field development is a kind of computing intensive task,which needs large numbers of reservoir simulation that is very expensive,especially for the complex reservoir models.The alternative model is less computational and costs little in history matching,production optimization and prediction process.It can evaluate the objective function value of a new program effectively and reduce computation caused by invoking the simulation model repeatly so that it can accelerate optimization procedure of oil field development.The paper presents four common alternative models.They are polynomial regression model(PRM),multivariate Kriging model(MKG),thin plate spline interpolation model(TPSM) and artificial neural network(ANN).
出处
《石油工业计算机应用》
2012年第2期17-19,3,共3页
Computer Applications Of Petroleum
关键词
替代模型
多项式回归模型
泛克里金模型
薄板样条插值
人工神经网络
proxy model
polynomial regression model
multivariate Kriging model
thin plate spline interpolation
artificial neural network
