Modeling hydrometallurgical leaching process based on improved just-in-time learning algorithm

doi:10.11949/j.issn.0438-1157.20161803

Abstract

Abstract:

Least squares support vector machine (LS-SVM) based on just-in-time (JIT) learning algorithm was used to build prediction model of leaching rate, when considered multi-variable multi-mode nonlinear characteristics of hydrometallurgical leaching process. Time order was introduced into selection rule of JIT learning set for the determination of modeling neighborhood of current operating point, so as to improve modelling accuracy. A cumulative similarity factor was adopted to improve real-time performance of the model and an adaptive similarity threshold was used to determine necessity of updating local model of the current operating point. The simulation results for hydrometallurgical leaching process show that the improved modeling method has high precision and good real-time performance in leaching rate prediction, which can be used in hydrometallurgical industrial production.

Key words: hydrometallurgy, leaching process, just-in-time learning algorithm, least squares support vector machine, time order, cumulative similarity factor

摘要：

针对湿法冶金浸出过程中存在的多变量、非线性和多工况等问题，采用基于即时学习算法的最小二乘支持向量机建立浸出率的预测模型。将时间有序性引入到即时学习算法学习集的选取规则中以确定系统当前工作点的建模邻域，从而提高模型精度；引入累计相似因子以提高所建模型的实时性，并利用自适应相似度阈值来判定是否需要重新建立当前工作点的局部模型。将改进的建模方法应用到湿法冶金浸出过程浸出率的预测中，仿真结果表明，所建模型具有较高的精度和实时性，可用于湿法冶金工业生产过程。

关键词: 湿法冶金, 浸出过程, 即时学习算法, 最小二乘支持向量机, 时间有序性, 累计相似因子

CLC Number:

TP301.6

NIU Dapeng, LIU Yuanqing. Modeling hydrometallurgical leaching process based on improved just-in-time learning algorithm[J]. CIESC Journal, 2017, 68(7): 2873-2879.

牛大鹏, 刘元清. 基于改进即时学习算法的湿法冶金浸出过程建模[J]. 化工学报, 2017, 68(7): 2873-2879.

References 30

[1]	CHENG L, LIU W, HOU Z G, et al. Neural network based nonlinear model predictive control for piezoelectric actuators [J]. IEEE Transactions on Industrial Electronics, 2015, 62 (12): 1-1.
[2]	ZHAO C H, WANG F L, GAO F R, et al. Adaptive monitoring method for batch processes based on phase dissimilarity updating with limited modeling data [J]. Industrial & Engineering Chemistry Research, 2007, 46 (14): 4943-4953.
[3]	FENG K, LU J, CHEN J. Nonlinear model predictive control based on support vector machine and genetic algorithm [J]. Chinese Journal of Chemical Engineering, 2015, 23 (12): 2048-2052.
[4]	ZHAO C H, WANG F L, LU N Y, et al. Stage-based soft-transition multiple PCA modeling and on-line monitoring strategy for batch processes [J]. Journal of Process Control, 2007, 17 (9): 728-741.
[5]	LIU Z T, FENG Z H. Research on two-dimensional landslide model control system based on support vector machine modeling [J]. Applied Mechanics & Materials, 2014, 580-583: 947-953.
[6]	LIU R L. Modeling soft sensor based on support vector machine and particle swarm optimization algorithms [J]. Control Theory & Applications, 2006, 23 (6): 895-894.
[7]	ZHAO C H, GAO F R. Online fault prognosis with relative deviation analysis and vector autoregressive modeling [J]. Chemical Engineering Science, 2015, 138 (22): 531-543.
[8]	ZHANG X, LI Y, KANO M. Quality prediction in complex batch processes with just-in-time learning model based on non-Gaussian dissimilarity measure [J]. Industrial & Engineering Chemistry Research, 2015, 54 (31): 7694-7705.
[9]	綦欢, 叶贞成, 钱锋,等. 乙烯裂解炉燃料气消耗的实时优化 [J]. 计算机与应用化学, 2015, 32 (12): 1457-1461.QI H, YE Z C, QIAN F, et al. Real-time optimization for fuel gas consumption of ethylene cracking furnace [J]. Computers and Applied Chemistry, 2015, 32 (12): 1457-1461.
[10]	周丽春, 靳鑫, 刘毅,等. 即时局部建模在填料塔液泛气速预测的应用 [J]. 化工学报, 2016, 67 (3): 1070-1075.ZHOU L C, XIN X, LIU Y, et al. Just-in-time local modeling for flooding velocity prediction in packed towers [J]. CIESC Journal, 2016, 67 (3): 1070-1075.
[11]	罗健旭, 邵惠鹤. 应用多神经网络建立动态软测量模型 [J]. 化工学报, 2003, 54 (12): 170-177.LUO J X, SHAO H H. Developing eveloping dynamic soft sensors using multiple neural networks [J]. Journal of Chemical Industry and Engineering (China), 2003, 54 (12): 170-177.
[12]	DU W L, GUAN Z Q, QIAN F. The time series soft-sensor modeling based on Adaboost LS-SVM[C]//The 8th World Congress on Intelligent Control and Automation (WCICA),Jinan, 2010: 1491-1495.
[13]	WANG H Q, LI P, GAO F R, et al. Kernel classifier with adaptive structure and fixed memory for process diagnosis [J]. AIChE Journal, 2006, 52 (10): 3515-3531.
[14]	BONTEMPI G, BIRATTARI M, BERSINI H. Lazy learning for local modeling and control design [J]. International Journal of Control,1999, 72 (7/8): 643-658.
[15]	USHIDA S, KIMURA H. Just-in-time approach to nonlinear identification and control [J]. Journal of SICE, 2005, 44 (2): 102-106.
[16]	PAN T H, LI S Y, WANG X. A multi-model modeling approach to nonlinear systems based on lazy learning[C]//Proceedings of the 24th Chinese Control Conference. 2005: 268-73.
[17]	LIU Y, GAO Z L, LI P, et al. Just-in-time kernel learning with adaptive parameter selection for soft sensor modeling of batch processes [J]. Industrial and Engineering Chemistry Research, 2012, 51 (11): 4313-4327.
[18]	SU Q L, MARTIN W H, RICHARD D B, et al. Just-in-time-learning based extended prediction self-adaptive control for batch processes [J]. Journal of Process Control, 2016, 43: 1-9.
[19]	LIMA L R P D A, HODOUIN D. Simulation study of the optimal distribution of cyanide in a gold leaching circuit [J]. Minerals Engineering, 2006, 19 (13): 1319-1327.
[20]	LIAM L R P D A. Some remarks on the reactor network synthesis for gold cyanidation [J]. Minerals Engineering, 2006, 19 (2): 154-161.
[21]	HU G, MAO Z Z, HE D K, et al. Hybrid modeling for the prediction of leaching rate in leaching process based on negative correlation learning bagging ensemble algorithm [J]. Computers & Chemical Engineering, 2011, 35 (12): 2611-2617.
[22]	LIU Y, WAN H Q, LI P. Adaptive local learning based least squares support vector regression with application to online modeling for fermentation processes [J]. Journal of Chemical Industry and Engineering, 2008, 59 (8): 2053-2057.
[23]	GE Z Q, SONG Z H. A comparative study of just-in-time-learning based methods for online soft sensor modeling [J].Chemometrics and Intelligent Laboratory Systems, 2010, 104 (2): 306-317.
[24]	NIU D P, SUN Y, WANG F L. Optimization of advertising budget allocation over time based on LS-SVMR and DE [J]. Automation Science & Engineering IEEE Transactions on, 2014, 10 (4): 1076-1082.
[25]	LIN W Q, JIAN-ZHONG F U, YA-ZHOW X U, et al. Thermal error modeling & compensation of numerical control machine tools based on on-line least squares support vector machine [J]. Computer Integrated Manufacturing Systems, 2008, 14 (2): 295-299.
[26]	SUN L, YU F, ZHANG D, et al. Online modeling for wood drying based on least squares support vector machine [J]. Chinese Journal of Scientific Instrument, 2009, 30 (9): 1991-1995.
[27]	SERAFINI T, ZANGHIRATI G, ZANNI L. Gradient projection methods for quadratic programs and applications in training support vector machines [J]. Optimization Methods & Software, 2005, 20 (2/3): 353-378.
[28]	葛志强, 刘毅, 宋执环. 一种基于局部模型的非线性多工况过程检测方法 [J]. 自动化学报, 2008, 34 (7): 792-797.GE Z Q, LIU Y, SONG Z H. Local model based monitoring for nonlinear multiple mode process [J]. Acta Automatica Sinica, 2008, 34 (7): 792-797.
[29]	刘毅, 金福江, 高增梁. 时变过程在线辨识的即时递推核学习研究 [J]. 自动化学报, 2013, 39 (5): 602-609.LIU Y, JIN F J, GAO Z L. Online identification of time-varying processes using just-in-time recursive kernel learning approach [J]. Acta Automatica Sinica, 2013, 39 (5): 602-609.
[30]	孙维, 王伟, 朱瑞军. 即时学习算法在非线性系统迭代学习控制中的应用 [J]. 控制与决策, 2003, 18 (3): 263-266.SUN W, WANG W, ZHU R J. Iterative learning control for nonlinear system using lazy learning method [J]. Control and Decision, 2003, 18 (3): 263-266.