基于BP神经网络的谱图形态提取在水淹评价中的研究

Feture of Chromatogram Based on BP Neural Network in Evalution of Water Flooded Layer

目前,我国大部分油田进入高含水后期开采阶段,调整挖潜难度增大,亟需一种高效的油层水淹评价方法,解决水淹评价参数单一的难题。因此提出采用BP神经网络对饱和烃气相色谱图进行形态拟合的方法,提取形态特征参数,应用于水淹层评价方法中。研究伽玛分布函数提取单峰谱图形态特征参数的方法,研究BP神经网络拟合色谱图原始形态的方法,将反映谱图形态特征与反映含油性特征的参数结合,建立油层水淹评价图版,经过在油田实际生产应用,提高了该区水淹层精细评价水平,为剩余油分布提供理论基础。

At present,most of China＇s oil field stop into the late stage of high water cut,it increases difficulty for adjustment and tapping potential,it urgently need an efficient reservoir flooding evaluation method to solve the problem of single parameters for evaluation of water flooded layer. By using the method of BP neural network to fit the form of chromatogram,extracting characteristic parameters from the chromatogram,the characteristic parameters are applied to the evaluation of water flooded layer. A method for extracting the characteristic parameters of chromatogram by gamma function and BP neural network is proposed,using the character-istic parameters and the original data to create the water flooded layer evaluation chart,the evaluation effect of water flooded layer is improved,theoretical basis for remaining oil is provided.