基于迁移成分分析的发酵过程集成软测量建模

doi:10.16182/j.issn1004731x.joss.21-1165

摘要/Abstract

摘要：

青霉素发酵过程具有不确定性和多阶段等特点，不同批次发酵过程间存在工况差异，过程数据的分布不一定相同，使传统软测量模型性能下降。结合迁移学习策略和高斯混合模型，提出一种基于迁移成分分析的多模型集成软测量建模方法。该方法使用迁移成分分析求解样本间共享特征映射矩阵，适配建模过程数据与待测数据的边缘概率分布；并基于高斯混合模型对建模数据进行聚类划分，与偏最小二乘算法结合建立子模型的集成模型，完成对主导变量的预测。基于青霉素平台数据的仿真结果表明，所提方法不仅能够有效提高青霉素发酵过程软测量模型的精度，而且适应于变工况下青霉素浓度的预测。

关键词: 迁移成分分析, 集成, 青霉素发酵, 变工况, 软测量

Abstract:

The Penicillin fermentation process is an uncertain and multi-stage process. There are different working conditions among different batch fermentation processes, and the distribution of process data is not necessarily the same, which degrades the performance of the traditional soft sensing model. Combined with the transfer learning strategy and Gaussian mixture model, a multi-model ensemble soft sensor modeling method based on transfer component analysis is proposed. In this method, the transfer component analysis is used to get the shared feature mapping matrix between samples, and adapt the edge probability distribution of labeled dataset and unlabeled dataset; the modeling data are clustered based on Gaussian mixture model, then use partial least squares algorithm to establish an ensemble model to predict dominant variables. The simulation results based on penicillin platform data show that the proposed method can not only improve the accuracy of the soft sensor model of penicillin fermentation process effectively, but also adapt to the prediction of penicillin concentration under variable working conditions.

Key words: transfer component analysis, ensemble, penicillin fermentation, variable working conditions, soft sensor

中图分类号:

TP274

周阅昇, 熊伟丽. 基于迁移成分分析的发酵过程集成软测量建模[J]. 系统仿真学报, 2023, 35(3): 623-631.

Yuesheng Zhou, Weili Xiong. Integrated Soft Sensor Modeling of Fermentation Process Based on Transfer Component Analysis[J]. Journal of System Simulation, 2023, 35(3): 623-631.

图/表 9

图1

图2

表1

图3

图4

表2

图5

图6

图7

参考文献 21

1	刘聪, 谢莉, 杨慧中. 基于改进DPC的青霉素发酵过程多模型软测量建模[J]. 化工学报, 2021, 72(3): 1606-1615.
	Liu Cong, Xie Li, Yang Huizhong. Multi-model Soft Sensor Development for Penicillin Fermentation Process Based on Improved Density Peak Clustering[J]. Journal of Chemical Industry and Engineering, 2021, 72(3): 1606-1615.
2	Birol G, Ündey C, Çinar A. A Modular Simulation Package for Fed-Batch Fermentation: Penicillin Production[J]. Computers and Chemical Engineering (S0098-1354), 2002, 26(11): 1553-1565.
3	曹鹏飞, 罗雄麟. 化工过程软测量建模方法研究进展[J]. 化工学报, 2013, 64(3): 788-800.
	Cao Pengfei, Luo Xionglin. Modeling of Soft Sensor for Chemical Process[J]. Journal of Chemical Industry and Engineering, 2013, 64(3): 788-800.
4	Francisco S, Rui A, Jerome M. Review of Soft Sensor Methods for Regression Applications[J]. Chemometrics and Intelligent Laboratory Systems (S0169-7439), 2016, 152: 69-79.
5	Wang Z, He Q, Wang J. Comparison of Variable Selection Methods for PLS-Based Soft Sensor Modeling[J]. Journal of Process Control (S0959-1524), 2015, 26: 56-72.
6	赵超, 李俊, 戴坤成, 等. 基于自适应加权最小二乘支持向量机的青霉素发酵过程软测量建模[J]. 南京理工大学学报, 2017, 41(1): 100-107.
	Zhao Chao, Li Jun, Dai Kuncheng, et al. Soft Sensor Modeling for Penicillin Fermentation Process Based on Adaptive Weighted Least Squares Support Vector Machine[J]. Journal of Nanjing University of Science and Technology, 2017, 41(1): 100-107.
7	Mamo N B, Dennis Y A. 基于人工神经网络的注水开发油藏产量预测[J]. 石油勘探与开发, 2020, 47(2): 357-365.
	Mamo N B, Dennis Y A. Artificial Neural Network Based Production Forecasting for a Hydrocarbon Reservoir under Water Injection[J]. Petroleum Exploration and Development, 2020, 47(2): 357-365.
8	李元, 燕亚运, 唐晓初. 基于局部模型的多阶段在线产品质量预测[J]. 系统仿真学报, 2016, 28(4): 966-971, 978.
	Li Yuan, Yan Yayun, Tang Xiaochu. Online Product Quality Prediction for Multi-phase Based on Local Model[J]. Journal of System Simulation, 2016, 28(4): 966-971, 978.
9	Shi X, Xiong W. Adaptive Ensemble Learning Strategy for Semi-supervised Soft Sensing[J]. Journal of the Franklin Institute (S0016-0032), 2020, 357: 3753-3771.
10	赵帅, 史旭东, 熊伟丽. 一种分层集成的高斯过程回归软测量建模方法[J]. 系统仿真学报, 2019, 31(10): 2042-2051.
	Zhao Shuai, Shi Xudong, Xiong Weili. A Hierarchical Integrated Soft Sensing Modeling Method for Gauss Process Regression[J]. Journal of System Simulation, 2019, 31(10): 2042-2051.
11	Yao L, Ge Z. Scalable Semi-supervised GMM for Big Data Quality Prediction in Multimode Processes[J]. IEEE Transactions on Industrial Electronics (S0278-0046), 2019, 66(5): 3681-3692.
12	Liu Y, Yang C, Liu K, et al. Domain Adaptation Transfer Learning Soft Sensor for Product Quality Prediction[J]. Chemometrics and Intelligent Laboratory Systems (S0169-7439), 2019, 192: 103813.
13	Long M, Wang J, Ding G, et al. Transfer Feature Learning with Joint Distribution Adaptation[C]// 2013 IEEE International Conference on Computer Vision (ICCV). Sydney, Australia: IEEE, 2013: 2200-2207.
14	阎高伟, 贺敏, 汤健, 等. 基于最大均值差异多源域迁移学习的湿式球磨机负荷参数软测量[J]. 控制与决策, 2018, 33(10): 1795-1800.
	Yan Gaowei, He Min, Tang Jian, et al. Soft Sensor of Wet Ball Mill Load Based on Maximum Mean Discrepancy Multi-source Domain Transfer Learning[J]. Control and Decision, 2018, 33(10): 1795-1800.
15	贺敏, 汤健, 郭旭琦, 等. 基于流形正则化域适应随机权神经网络的湿式球磨机负荷参数软测量[J].自动化学报, 2019, 45(2): 398-406.
	He Min, Tang Jian, Guo Xuqi, et al. Soft Sensor for Ball Mill Load Using DAMRRWNN Model[J]. Acta Automatica Sinica, 2019, 45(2): 398-406.
16	Pan S J, Tsang I W, Kwok J T, et al. Domain Adaptation Via Transfer Component Analysis[J]. IEEE Transactions on Neural Networks (S0893-6080), 2011, 22(2): 199-210.
17	兰雨涛, 胡超凡, 金京, 等. 基于迁移成分分析的跨域轴承故障分类方法研究[J]. 机电工程, 2021, 38(5): 521-527.
	Lan Yutao, Hu Chaofan, Jin Jing, et al. Cross Domain Classification Method of Bearings Fault Based on Transfer Component Analysis[J]. Mechanical & Electrical Engineering Magazine, 2021, 38(5): 521-527.
18	Giona M, Michele V, Mikhail K, et al. Semi-supervised Transfer Component Analysis for Domain Adaptation in Remote Sensing Image Classification[J]. IEEE Transactions on Geoscience and Remote Sensing (S0196-2892), 2015, 53(7): 3550-3564.
19	Pan S J, Yang Q. A Survey on Transfer Learning[J]. IEEE Transactions on Knowledge and Data Engineering (S1041-4347), 2010, 22(10): 1345-1359.
20	Kaneko H, Funatsu K. Adaptive Soft Sensor Based on Online Support Vector Regression and Bayesian Ensemble Learning for Various States in Chemical Plants[J]. Chemometrics and Intelligent Laboratory Systems (S0169-7439), 2014, 137: 57-66.
21	Sheng X, Ma J, Xiong W. Smart Soft Sensor Design with Hierarchical Sampling Strategy of Ensemble Gaussian Process Regression for Fermentation Processes[J]. Sensors (S1424-8220), 2020, 20(7): 1957.

算法	RMSE	MAE	R²
LWPLS	0.031 5	0.021 9	0.995 3
GMM-PLS	0.025 4	0.019 6	0.996 9
TCA-GMM-PLS	0.006 1	0.004 1	0.999 8

算法	RMSE	MAE	R²
LWPLS	0.133 6	0.081 2	0.916 2
GMM-PLS	0.061 9	0.045 7	0.982 0
TCA-GMM-PLS	0.048 2	0.026 4	0.989 1

[1]	刘文正, 张和明. 面向复杂系统运动仿真的信息物理融合建模[J]. 系统仿真学报, 2022, 34(5): 954-963.
[2]	周思锦, 陈棣成, 涂耿, 姜大志. 基于个性化和记忆机制的多模态情感计算模型[J]. 系统仿真学报, 2022, 34(4): 745-758.
[3]	陈魁, 毕利, 王文雅. 柔性作业车间AGV与机器双资源集成调度研究[J]. 系统仿真学报, 2022, 34(3): 461-469.
[4]	刘闽东, 吴龙军, 汤明超, 孟梅. 项目型制造过程信息流集成建模仿真研究[J]. 系统仿真学报, 2022, 34(3): 555-563.
[5]	白燕, 武璐璐, 贺引娥, 王玉英. 基于动态温度调控的空调系统能耗预测[J]. 系统仿真学报, 2022, 34(2): 366-375.
[6]	奇格奇, 刘思劲, 何一康, 王猛, 黄爱玲. 基于仿真驱动的高速公路主动限速效用评价与推荐[J]. 系统仿真学报, 2022, 34(12): 2522-2534.
[7]	代学志, 熊伟丽. 基于概率选择的主动学习智能软测量建模[J]. 系统仿真学报, 2021, 33(6): 1350-1357.
[8]	张天瀛, 姬杭, 周军华, 陶栾. 云端工具集成和管理方法研究[J]. 系统仿真学报, 2021, 33(12): 2911-2918.
[9]	胡小曼, 王艳, 纪志成. 模糊信息粒化与改进RVM的滚动轴承寿命预测[J]. 系统仿真学报, 2021, 33(11): 2561-2571.
[10]	邓欣, 龙灿, 米建勋, 张博宪, 孙开伟, 王进. 基于多域特征与随机子空间集成的脑电分类[J]. 系统仿真学报, 2020, 32(9): 1787-1798.
[11]	杨文峰, 王艳, 纪志成. 基于RF-GA-BP神经网络的N-乙酰氨基葡萄糖含量预测[J]. 系统仿真学报, 2020, 32(10): 2034-2040.
[12]	金炜东, 陈春利. 基于集成深度学习的雷达信号分选研究[J]. 系统仿真学报, 2019, 31(9): 1868-1874.
[13]	谭亚新, 罗建华, 范锐, 董志明. 一种基于作战编成并行仿真方法及其实现[J]. 系统仿真学报, 2019, 31(11): 2350-2355.
[14]	吴迎年, 贺梦嘉, 项伟. 基于集成学习的人眼定位和头部状态检测研究[J]. 系统仿真学报, 2019, 31(11): 2366-2373.
[15]	赵帅, 史旭东, 熊伟丽. 一种分层集成的高斯过程回归软测量建模方法[J]. 系统仿真学报, 2019, 31(10): 2042-2051.