广义回归神经网络技术在新油田快速评价中的应用

Application of Generalized Regression Neural Network in Rapid Evaluation of New Oilfields

新油田的动态和静态资料较少,油藏认识程度低,开发指标预测难度大。提出运用广义回归神经网络技术进行非线性优化回归,通过对比相似油田开发模式和开发效果,快速评价及预测新油田的开发方式和开采指标。以处于前期研究阶段的某油田为例,运用此方法进行指标预测,计算结果符合实际情况。

New oilfields have few dynamic and static data,so it is difficult to predict the development index due to the low level of reservoir recognition.Nonlinear optimal regression is carried out in this paper based on generalized regression neural network technology to fast evaluate and predict the new oilfield′s development program and production index data,by comparing the development modes and effects of similar oilfields.Take an oilfield in the early stage of development as an example,the prediction results of the method fit the actual situation.