基于关联变量时滞分析卷积神经网络的生产过程时间序列预测方法

doi:10.11949/j.issn.0438-1157.20170197

摘要/Abstract

摘要：

生产过程通常具有大时滞、非线性、多变量耦合等特点，往往难以建立准确的时间序列预测模型。基于生产过程历史数据，提出了一种采用关联变量时滞分析卷积神经网络（CNN）的生产过程时间序列预测方法，首先选取合适的关联变量并对关联变量与输出进行时滞分析，然后利用时滞分析结果确定关联变量时间窗的大小，最后建立合适的CNN模型对时间序列进行预测。某反应精馏过程实验表明，此方法对于大时滞系统的长步长时间序列预测具有较好的准确性。

关键词: 神经网络, 预测, 关联变量, 时滞系统, 反应精馏

Abstract:

It is often difficult to establish models for accurate time series prediction, as a result of time delay, nonlinearity, and multi-variable coupling characteristics in production processes. A time series prediction method was proposed by applying correlated process variables to time delayed analysis of convolutional neural network (CNN). First, appropriate correlated variables were selected and time series were analyzed between these variables and outputs. Then, results of time series analysis were employed to determine length of temporal time windows of associated variables. Finally, appropriate CNN models were established for time series prediction. Experimental results of a reaction distillation process showed good accuracy in time series prediction of long time frame and large time-delay processes.

Key words: neural networks, prediction, correlated variable, time-delay system, reactive distillation

中图分类号:

TP183

张浩, 刘振娟, 李宏光, 杨博, 路洁. 基于关联变量时滞分析卷积神经网络的生产过程时间序列预测方法[J]. 化工学报, 2017, 68(9): 3501-3510.

ZHANG Hao, LIU Zhenjuan, LI Hongguang, YANG Bo, LU Jie. Process time series prediction based on application of correlated process variables to CNN time delayed analyses[J]. CIESC Journal, 2017, 68(9): 3501-3510.

参考文献 33

[1]	李宏光, 臧灏. 管式加热炉多模型神经网络预测控制[J]. 控制工程, 2012, 19(4):650-653. LI H G, ZANG H. Multi-model adaptive neural network based predictive control of furnace[J]. Control Engineering of China, 2012, 19(4):650-653.
[2]	ZHANG X X, XU J, ZHANG S, et al. Experiment on parameter identification of a time delay coupled nonlinear system[J]. IFAC Papersonline, 2015, 48(11):694-699.
[3]	SZMIT M, SZMIT A. Use of holt-winters method in the analysis of network traffic case study[J]. Communications in Computer & Information Science, 2011, 160:224-231.
[4]	ROSSI M, BRUNELLI D. Forecasting data centers power consumption with the Holt-Winters method[C]//Process of IEEE Workshop on Environmental, Energy and Structural Monitoring Systems IEEE, 2015:210-214.
[5]	PRAVILOVIC S, APPICE A, MALERBA D. Integrating Cluster Analysis to the ARIMA Model for Forecasting Geosensor Data[M]. Foundations of Intelligent Systems, 2014:234-243.
[6]	王锋, 李宏光, 臧灏. 基于Logistic和ARMA模型的过程报警预测[J]. 化工学报, 2012, 63(9):2941-2947. WANG F, LI H G, ZANG H. Process alarm prognosis based on Logistic and ARMA models[J]. CIESC Journal, 2012, 63(9):2941-2947.
[7]	LENG X Z, WANG J H, JI H B, et al. Prediction of size-fractionated airborne particle-bound metals using MLR, BP-ANN and SVM analyses[J]. Chemosphere, 2017, 180:513.
[8]	WANG L, TING M F, CHAPMAN D, et al. Prediction of northern summer low-frequency circulation using a high-order vector auto-regressive model[J]. Climate Dynamics, 2015, 46:1-17.
[9]	LIU Y K, XIE F, XIE C L, et al. Prediction of time series of NPP operating parameters using dynamic model based on BP neural network[J]. Annals of Nuclear Energy, 2015, 85:566-575.
[10]	LI G, SHI J, ZHOU J Y. Bayesian adaptive combination of short-term wind speed forecasts from neural network models[J]. Renewable Energy, 2011, 36(1):352-359.
[11]	ATIYA A F, EL-SHOURA S M, SHAHEEN S I, et al. A comparison between neural-network forecasting techniques-case study:river flow forecasting[J]. IEEE Transactions on Neural Networks, 1999, 10(2):402-409.
[12]	CHEVILLON G. Direct multi-step estimation and forecasting[J]. Journal of Economic Surveys, 2007, 21(4):746-785.
[13]	XIAO Q K, XING L, SONG G. Time series prediction using optimal theorem and dynamic Bayesian network[J]. Optik-International Journal for Light and Electron Optics, 2016, 127(23):11063-11069.
[14]	BONTEMPI G. Long term time series prediction with multi-input multi-output local learning[J]. Proceedings of European Symposium on Time, 2008:145-154.
[15]	SORJAMAA A, LENDASSE A. Time series prediction using dirrec strategy[C]//Esann 2006, European Symposium on Artificial Neural Networks, Bruges, Belgium, Proceedings. DBLP, 2006:143-148.
[16]	YANG F, DUAN P, SHAH S L, et al. Capturing connectivity and causality in complex industrial processes[C]//Springer Publishing Company, Incorporated, 2014:45-46.
[17]	YAN H, ZHANG Y P, YANG Q. Time-delay estimation based on cross-correlation and wavelet denoising[C]//Chinese Intelligent Automation Conference, 2013:841-848.
[18]	PERSON R S, MISHIN L N. Auto-and cross-correlation analysis of the electrical activity of muscles[J]. Medical & Biological Engineering & Computing, 1964, 2(2):155.
[19]	SCHMIDHUBER J. Deep learning in neural networks:an overview[J]. Neural Networks, 2015, 61:85-117.
[20]	KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks[C]//International Conference on Neural Information Processing Systems, 2012:1097-1105.
[21]	SOUTNER D, MÜLLER L. Application of LSTM Neural Networks in Language Modelling[M]. Text, Speech, and Dialogue. Springer Berlin Heidelberg, 2013:105-112.
[22]	MORTON J, WHEELER T A, KOCHENDERFER M J. Analysis of recurrent neural networks for probabilistic modeling of driver behavior[J]. IEEE Transactions on Intelligent Transportation Systems, 2016, 99:1-10.
[23]	LIU W, ANGUELOV D, ERHAN D, et al. SSD:Single shot multibox detector[C]//European Conference on Computer Vision. Springer International Publishing, 2016:21-37.
[24]	GIRSHICK R. Fast R-CNN[C]//IEEE International Conference on Computer Vision. IEEE Computer Society, 2015:1440-1448.
[25]	REN S Q, HE K M, GIRSHICK R, et al. Faster R-CNN:towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2016, 39(6):1137.
[26]	段艳杰, 吕宜生, 张杰, 等. 深度学习在控制领域的研究现状与展望[J]. 自动化学报, 2016, 42(5):643-654. DUAN Y J, LÜ Y S, ZHANG J, et al. Deep learning for control:the state of the art and prospects[J]. Acta Automatica Sinica, 2016, 42(5):643-654.
[27]	韩改堂, 乔俊飞, 韩红桂. 基于递归模糊神经网络的污水处理控制方法[J]. 化工学报, 2016, 67(3):954-959. HAN G L, QIAO J W, HAN H G. Wastewater treatment control method based on recurrent fuzzy neural network[J]. CIESC Journal, 2016, 67(3):954-959.
[28]	ZHU Q X, JIA Y W, PENG D, et al. Study and application of fault prediction methods with improved reservoir neural networks[J]. Chinese Journal of Chemical Engineering, 2014, 22(7):812-819.
[29]	POUL A K, SOLEIMANI M, SALAHI S. Solubility prediction of disperse dyes in supercritical carbon dioxide and ethanol as co-solvent using neural network[J]. Chinese Journal of Chemical Engineering, 2016, 24(4):491-498.
[30]	HE Y L, XU Y, ZHAO Z Q, et al. Soft sensor of chemical processes with large numbers of input parameters using auto-associative hierarchical neural network[J]. Chinese Journal of Chemical Engineering, 2015, 23(1):138-145.
[31]	安爱民, 刘云利, 张浩琛, 等. 微生物燃料电池的动态性能分析及其神经网络预测控制[J]. 化工学报, 2017, 68(3):1090-1098. AN A M, LIU Y L, ZHANG H T, et al. Dynamic performance analysis and neural network predictive control of microbial fuel cell[J]. CIESC Journal, 2017, 68(3):1090-1098.
[32]	SEDKI A, OUAZAR D, MAZOUDI E E. Evolving neural network using real coded genetic algorithm for daily rainfall-runoff forecasting[J]. Expert Systems with Applications, 2009, 36(3):4523-4527.
[33]	ZHANG G P. Time series forecasting using a hybrid ARIMA and neural network model[J]. Neurocomputing, 2003, 50:159-175.