Data-driven soft sensor for animal cell suspension culture process based on DRVM

Abstract In order to solve the problem that key state variables (such as glucose concentration, lactic acid concentration and cell density) in the dynamic process of animal cell suspension culture are difficult to be measured in real time, a data-driven soft sensor based on dynamic relevance vector machine (DRVM) algorithm is proposed. The dominant variables of the soft sensor model are selected according to the mechanisms process. The c o r r c o e f ( ) function (belongs to the correlation coefficient command in MATLAB) is used to analyze the correlation among environmental variables, and the auxiliary variables of the soft sensor model are further determined. An improved method on the three edge location algorithm is used to optimize the dynamic weights of the DRVM model. Considering the influence of dynamic transition on soft sensor, the maximum likelihood distribution method under the Bayesian framework is used to train DRVM weight and super parameters, and the dynamic soft sensor model of animal cell suspension culture is established. The proposed method is applied to predict the key state variables in BHK-21 cell suspension culture Process. The experimental results show that compared with the traditional static soft sensing based on RVM, the data-driven soft sensor based on DRVM has higher accuracy, and the rationality and superiority of the method are verified. In order to further realize the real-time online prediction of key state variables, the monitoring interface of the suspension culture process is developed on the LabVIEW virtual instrument platform through its MATLAB Script node, and the data exchange of the DRVM soft sensor for the key state variables of the cell suspension culture process based on MATLAB and monitoring interface is realized.

[1]  Mika Liukkonen,et al.  Dynamic soft sensors for NOx emissions in a circulating fluidized bed boiler , 2012 .

[2]  Luiz Augusto da Cruz Meleiro,et al.  ANN-based soft-sensor for real-time process monitoring and control of an industrial polymerization process , 2009, Comput. Chem. Eng..

[3]  Dexian Huang,et al.  Novel Bayesian Framework for Dynamic Soft Sensor Based on Support Vector Machine With Finite Impulse Response , 2014, IEEE Transactions on Control Systems Technology.

[4]  U. Reichl,et al.  Toward Animal Cell Culture–Based Influenza Vaccine Design: Viral Hemagglutinin N-Glycosylation Markedly Impacts Immunogenicity , 2013, The Journal of Immunology.

[5]  T. S. Lee,et al.  A 3D analysis of oxygen transfer in a low-cost micro-bioreactor for animal cell suspension culture , 2007, Comput. Methods Programs Biomed..

[6]  Huang Yong-hon.  Soft Sensor Modeling Based on Biological Variables of Marine Protease Fermentation Process , 2013 .

[7]  Wang Youmin,et al.  Application of relevance vector machine in battery's remaining capacity prediction , 2012 .

[8]  Guohai Liu,et al.  Model optimization of SVM for a fermentation soft sensor , 2010, Expert Syst. Appl..

[9]  D. Simon.  Optimal State Estimation: Kalman, H Infinity, and Nonlinear Approaches , 2006 .

[10]  Xiong-lin Luo,et al.  Modeling for soft sensor systems and parameters updating online , 2014 .

[11]  Zhiqiang Ge,et al.  Supervised Latent Factor Analysis for Process Data Regression Modeling and Soft Sensor Application , 2016, IEEE Transactions on Control Systems Technology.

[12]  O. Merten,et al.  Introduction to animal cell culture technology—past, present and future , 2006, Cytotechnology.

[13]  Dimitri Solomatine,et al.  Experimental investigation of the predictive capabilities of data driven modeling techniques in hydrology - Part 1: Concepts and methodology , 2009 .

[14]  Hideshi Kawakami,et al.  Self-organization of polarized cerebellar tissue in 3D culture of human pluripotent stem cells. , 2015, Cell reports.

[15]  Michael E. Tipping.  Sparse Bayesian Learning and the Relevance Vector Machine , 2001, J. Mach. Learn. Res..

[16]  Xinghuo Yu,et al.  Distributed Optimal Consensus Over Resource Allocation Network and Its Application to Dynamical Economic Dispatch , 2018, IEEE Transactions on Neural Networks and Learning Systems.

[17]  J. Hamel,et al.  Suspension Culture Process of MethA Tumor Cell for the Production of Heat‐Shock Protein Glycoprotein 96: Process Optimization in Spinner Flasks , 2007, Biotechnology progress.

[18]  Chi-Man Vong,et al.  Modelling and prediction of automotive engine airratio using relevance vector machine , 2012, 2012 12th International Conference on Control Automation Robotics & Vision (ICARCV).

[19]  A. Tonso,et al.  Approach toward an efficient inoculum preparation stage for suspension BHK-21 cell culture , 2014, Cytotechnology.

[20]  D. Dixon,et al.  Ability of 16 priority PAHs to be directly cytotoxic to a cell line from the rainbow trout gill. , 1998, Toxicology.

[21]  N. Cho,et al.  Cancer‐associated fibroblast promote transmigration through endothelial brain cells in three‐dimensional in vitro models , 2014, International journal of cancer.

[22]  N. Koizumi,et al.  An evaluation of cultivated corneal limbal epithelial cells, using cell-suspension culture. , 2002, Investigative ophthalmology & visual science.