应用数学和力学2017，Vol.38Issue(10)：1136-1145,10.DOI:10.21656/1000-0887.370339
小波神经网络模型预测二氧化碳+水溶液体系界面张力
Prediction of Interfacial Tension Between CO2 and Brine With the Wavelet Neural Network Method
摘要
Abstract
Interfacial tension ( IFT) between CO2 and formation water is one of the most im-portant parameters for CO2 capture and storage, for it controls the transport properties of both phases in the formation. In order to rapidly and accurately predict the IFT of the CO2-brine sys-tem, 1677 sets of measured IFT data from previous studies were acquired. A wavelet neural network ( WNN) prediction model was proposed in view of 6 parameters including the pres-sure, the temperature, the CH4 molality and the N2 molality in CO2 gas, the monovalent cation ( Na+ and K+) concentration and the bivalent cation ( Ca2+and Mg2+) concentration. The simula-tion results show that a 3-layer ( 6-16-1) WNN model comes out of 839 data as the training data-sets. The mean absolute error (MMAE) , the mean relative error (MMARE) , the root mean squared error (MMSE) and the determination coefficient (R2) of the WNN model were 1.23 mN/m, 3.30％, 2.30 mN2/m2 and 0. 988, respectively. The performance of the WNN model was further com-pared with one newly proposed multivariate fitting model and the BP neural network model. The comparison results suggest that the WNN model is better than the other 2.关键词
小波神经网络/界面张力/二氧化碳+水溶液体系Key words
wavelet neural network/interfacial tension/CO2-brine system分类
能源科技引用本文复制引用
江安,刘平礼,李年银,张云飞,杜新伟..小波神经网络模型预测二氧化碳+水溶液体系界面张力[J].应用数学和力学,2017,38(10):1136-1145,10.基金项目
国家自然科学基金(面上项目) (51574197) The National Natural Science Foundation of China (General Program ) (51574197) （面上项目）
