摘要
了解油井生的井底流压大小是现场生产测试和分析中的一项重要工作。自喷井的井底流压值与产油量、含水、气油比、流体性质等参数呈复杂的非线性关系。神经网络具有表达任意非线性映射的能力，可以将其应用于建立自喷井井底流压预测模型。用一定数量的实测井底流压及相应的有关参数，根据BP神经网络学习算法对网络进行训练。训练后的网络就可用于预测一定生产条件下的井底流压，减少实测次数。实例计算表明：用神经网络建立自喷井井底流压预测模型是可行的，计算精度高。
It's a important task of production test and analyses in oil field to find out flowing bottomhole pressure (FBHP) of a flowingwell- FBHP of the flowing well has a complex nonlinear relationship with oil rate, water cut, gas-oil ratio and fluidproperties. etc.. Neural networks have a capability of expressing arbitrary nonlinear mapping' They are applied to developthe model for predicting FBHP of the flowing well in this paper. The BP neural network is trained based on its learningalgorithm with a number of sets of measured FBHP and other corresponding parameters of the flowing wel1s- The trainedneural network then can be used to predict FBHP in other production conditions. The calculation example is giyen and theresults show that the method is practicable and has high accuracy.
出处
《石油勘探与开发》
SCIE
EI
CAS
CSCD
北大核心
1997年第5期92-94,共3页
Petroleum Exploration and Development
关键词
自喷井
井底
流动压力
神经网络
Flowing well, Bottom hole, Flowing pressure, Prediction, Model
