摘要
以人工神经网络为手段 ,建立了油水分离旋流器设计模型。采用三层BP网络模型 ,成功地实现了根据处理物料物性参数和分离要求进行油水分离旋流器结构与操作参数全面设计的过程。通过设定足够大的神经网络训练次数 ,神经网络预测误差可逼近所需精度 。
A mathematical model based on artificial neural networks (ANN) was established for designing the geometric and operating parameters of oil water separation hydrocyclones. By using a three layer back propagation (BP) network, the comprehensive design of the geometric and operating parameters of oil water separation hydrocylcones was successfully carried out with only inputting the physical property of the oily water and the separation requirement parameters. By adopting enough run number for training the BP network, the prediction of the network can be approximated in arbitrary required accuracy, that is, the design modeling of the oil water hydrocyclones using the BP network can meet the requirement of the design.
出处
《油田化学》
CAS
CSCD
北大核心
2002年第3期250-252,256,共4页
Oilfield Chemistry
基金
中国石油江汉机械研究所委托项目"旋流油水分离器设计软件"(编号 981 0 2 8)部分研究成果
