基于FCM-BP神经网络的难采储量分类

目前储量的分类标准是通过划分指标值的范围来确定的,这就要求所有指标值恰好符合既定的指标范围,否则难以划分储量类别.为克服这一问题,结合模糊C均值算法和BP神经网络实现难采储量的分类.首先基于效益指标运用模糊C均值算法自动搜索储量的最佳类别,再利用BP神经网络建立储量效益指标类别与储量属性指标之间的关系表达式.在已知储量指标值的情况下,通过此关系式即可求得储量的类别.最后以大庆某油田为实例,对其难采储量进行了分类,有效指导难采储量滚动开发决策.

Classification of Difficult Recoverable Reserves Based on FCM and BP Neural Network

Currently,the classification and evaluation criterion of reserves were determined through the scope of the criteria value,which required all criteria values were just right in the existing range of criteria.Otherwise it would be difficult to divide the reserves category. To overcome this problem,this paper combined with Fuzzy C-Means clustering algorithm (FCM) and BP neural network method to classify difficult recoverable reserves.First use FCM to automatically search for the optimal category of reserves based on performance indicators.And then establish the relational expression between the reserves category and reserves properties by BP neural network.So in the case of the criteria value known,the categories of reserves can be obtained through this relational expression.Finally take the case of an oil field in the 10th Oil Production Plant of PetroChina Daqing Oilfield LLC, and evaluate the recoverable reserves,which conducts the rolling development of recoverable reserves.