| 211 | 1 | 142 |
| 下载次数 | 被引频次 | 阅读次数 |
水盐体系相平衡的研究是高盐卤水资源化利用的理论指导。本文利用优化的Elman神经网络(ENN)对海水体系Na+, K+, Ca2+, Mg2+//Cl~-, SO42-, CO32-, HCO3~-, B_4O72--H_2O的15个三元子系统进行了研究。通过比较溶解度预测值与文献值,确定模型的最优拓扑结构为双隐含层,隐含层神经元拓扑结构为(2, 3),学习函数选择动量自适应学习速率梯度下降法(GDX)。由NaCl-CaCl2-H_2O体系液相线的预测结果可知,ENN-(2,3)GDX模型的拟合精确度和运算速度均优于热力学HMW模型。对未发表的四个三元体系液相线和共饱点进行了预测,结果能够较好地符合随温度变化趋势,其中,15.0℃下NaCl-KCl-H_2O体系包含1个NaCl、KCl共饱点;87.5℃下Na_2SO4-MgSO4-H_2O体系包含3个共饱点,分别是:硫镁矾(M1)、钠镁钒(Low)共饱点,Low、无水钠镁钒(Van)共饱点和Van、Na_2SO4共饱点;60.0℃下KCl-K2-SO4-H_2O体系包含1个KCl、K_2SO4共饱点;60.0℃下NaCl-Na_2CO3-H_2O体系包含1个NaCl、Na_2CO3·H_2O共饱点。
Abstract：Research on the water-salt system can guide the utilization process on high salinity water. In this paper, an optimized Elman neural network is used to compare the predicted solubility with literature, those are from the fifteen ternary subsystems of seawater system Na+, K+, Ca2+, Mg2+//Cl~-, SO42-, CO32-, HCO3~-, B_4O72--H_2O. The optimized topology consists with two hidden layers in which two neurons connect the first layer and three neurons connect the second layer. The adaptive gradient descent function(GDX) was selected as the learning function. The forecasting precision of ENN model is better than that of HMW model, by studying on the system of NaCl-KCl-H_2O, whose MSE is the highest of all the ternary subsystems. The liquidus and saturation points of the four unpublished ternary systems were predicted, which can well conform to the the trend of temperature variation. In addition, the ternary system of NaCl-KCl-H_2O at 15.0 ℃ contains one saturation point of NaCl and KCl; the ternary system of Na_2SO4-MgSO4-H_2O at 87.5 ℃ contains three saturation points, which are MgSO4·H_2O(M1) and Na_2SO3·MgSO4·2.5 H_2O(Low), Low and 3 Na_2SO3·MgSO4(Van), Van and Na_2SO4. The ternary system of KCl-K_2SO4-H_2O at 60.0 ℃ contains one saturation point of KCl and K_2SO4. The ternary system of NaCl-Na_2CO3-H_2O contains one saturation point of NaCl and Na_2CO3·H_2O.
[1] Domańska U.Experimental data of fluid phase equilibria correlation and prediction models:A review[J].Processes,2019,7:277.
[2] Li D,Huang B,Meng L,et al.Solubility measurement and solid-liquid equilibrium model for the ternary system MgBr2+MgSO4+H2O at 288.15K[J].Braz J Chem Eng,2014,31(2):561-570.
[3] Meng X,Wu X,Yuan J.Solid-liquid equilibrium of the quaternary system Ca2+,Mg2+//SO32-,SO42--H2O at 298.15 K[J].J Chem Eng Data,2018,63 (12):4339-4349.
[4] Debye P,Hückel E.The theory of electrolytes I Lowering of freezing point and related phenomena[J].Phys Z,1923,24:185-206.
[5] Pitzer K S.Thermodynamics of electrolytes I theoretical basis and general equations[J].Thermodynamics of Electrolytes,1973,77(2):268-277.
[6] Cui R Z,Li W,Dong Y P.Measurement and prediction of solid + liquid equilibria in the quaternary system LiCl-KCl-Li2SO4-K2SO4-H2O at 288.15 K[J].J Chem Eng Data,2019,64:4206-4213.
[7] Li D D,Zeng D W,Yin X,et al.Phase diagrams and thermochemical modeling of salt lake brine systems.II.NaCl+H2O,KCl+H2O,MgCl2+H2O and CaCl2+H2O systems[J].Calphad,2016(53):78-89.
[8] Li C,Wang D Y,Guo H F,et al.Phase diagrams of the quaternary system NaCl-MgCl2-NH4Cl-H2O at 333.15 K and their application[J].J Chem Thermodyn,2020,140:105872.
[9] Qiao C Z,Zhang J,Jiang P,et al.A molecular approach for predicting phase diagrams of ternary aqueous saline solutions[J].Chem Eng Sci,2020(211):115278.
[10] 王艳,吉晓燕,陆小华,等.存在反应的电解质溶液相平衡研究[J].南京工业大学学报,2006,28(1):107-112.Wang Yan,Ji Xiaoyan,Lu Xiaohua,et al.Progress in study of phase equilibrium in electrolyte solution involving chemical reaction[J].Journal of Nanjing University of Technology,2006,28(1):107-112.
[11] Goodfellow I,Courville A,Bengio Y.Deep Learning[M].Cambridge:MIT Press,2016.
[12] Jiang X,Xue H,Zhang L.Nondestructive detection of chilled mutton freshness based on multi-label information fusion and adaptive BP neural network[J].Computers and Electronics in Agriculture,2018,155:371-377.
[13] Whittington J C R,Bogacz R.Theories of error back-propagation in the brain[J].Trends Cognit Sci,2019,23(3):141-153.
[14] Li X Y,Zhang L,Wang Z P,et al.Remaining useful life prediction for lithium-ion batteries based on a hybrid model combining the long short-term memory and Elman neural networks[J].J Energy Storage,2019(21):510-518.
[15] Li W H,Jiao Z P,Fan W Y,et al.An indirect RUL prognosis for lithium-ion battery under vibration stress using Elman neural network[J].International Journal of Hgdrogen Energy,2019,44(23):12270-12276.
[16] Li F,Gu Z,Ge L,et al.Application of artificial neural networks to X-ray fluorescence spectrum analysis[J].X-Ray Spectrom,2019,48(2):141-153.
[17] Akhil K,Ashok K.Vector-time-series-based back propagation neural network modeling of air quality inside a public transportation bus using available software[J].Environ Prog Sustainable Energy,2016,35(1):7-13.
[18] 高隽.人工神经网络原理及仿真实例[M].北京:机械工业出版社,2004.Gao Jun.Artificial Neural Network Principle and Simulation Example[M].Beijing:Mechanical Industry Press,2004.
[19] 贾文铜,周瑞祥,张忠,等.基于Elman神经网络的齿轮故障诊断研究[J].计算机测量与控制,2012,20(5):1173-1175.Jia Wentong,Zhou Ruixiang,Zhang Zhong,et al.Research on gear fault diagnose based on Elman neural NET[J].Computer Measurement & Control,2012,20(5):1173-1175.
[20] 邓天龙,周桓,陈侠.水盐体系相图及其应用[M].北京:化学工业出版社,2013.Deng Tianlong,Zhou Huan,Chen Xia.Phase Diagram of Water Salt System and Its Application[M].Beijing:Chemical Industry Press,2013.
[21] Stephen H,Stephen T.Solubilities of Inorganic and Organic Compounds[M].Oxford:Pergamon Press,1963.
[22] 闻新.应用MATLAB实现神经网络[M].北京:国防工业出版社,2015.Wen Xin.Realization of Neural Network with MATLAB[M].Beijing:National Defense Industry Press,2015.
[23] Alpaydin E.Neural Networks and Deep Learning[M].Cambridge:MIT Press,2016.
基本信息:
DOI：10.16441/j.cnki.hdxb.20190352
中图分类号:TP183;O642.42
引用信息:
[1]孟宪泽,付云朋,袁俊生,等.利用Elman神经网络预测三元水盐体系相平衡溶解度[J],2020,50(08):102-108.DOI:10.16441/j.cnki.hdxb.20190352.
基金信息:
国家科技支撑计划项目(2016YFB0600504)资助~~
2019-10-21
2019
2020-01-06
2020
2