低渗气藏压裂井稳定产能预测方法研究  

PREDICTION OF CONSTANT RATE OF LOW PERMEABILITY FRACTURING GAS WELL BY USING BP NEURAL NETWORK

作　　者：苏建政[1] 

机构地区：[1]中国石油化工股份有限公司石油勘探开发研究院

出　　处：《钻采工艺》2008年第5期90-92,102,共4页Drilling & Production Technology

摘　　要：通过分析压裂气井稳定产能与其影响因素之间的相关性,应用目前流行的BP人工神经网络方法,建立了压后气井稳态产能预测模型,并且在Matlab软件平台上对网络模型实现。根据现场收集的近30口气井的压裂施工数据和压后产能数据,对网络进行训练,并将训练好的网络用于同区块的压裂井稳定产能预测分析。结果表明,与常规方法相比,该方法不需要复杂数值模拟计算,预测精度可以指导现场生产,为气田在区块开发过程中压裂气井稳定产能评价提供了一种分析方法。This paper analyzed the interdependency with the constant rate of fracturing gas well and its influence factors,estab-lished the prediction model of constant rate based on the BP neural network and realized this model using Matlab software platform.Based on collecting the construction data and deliverability data of nearly 30 wells,the neural network was trained,and after being trained,the network was used to predict the constant rate of frac-turing gas well in the same block.The results showed that the method didn' t need complex numerical simulation computation compared with other conventional methods,and its computation re-sult is practicable to predict the constant rate of fracturing gas well.

关 键 词：压裂气井 神经网络 产能评价 

分 类 号：TE377[石油与天然气工程—油气田开发工程]