Mechanistic Mathematical Models as a Basis for Digital Twins
A future-oriented approach is the application of Digital Twins for process development, optimization and finally during manufacturing. Digital Twins are detailed virtual representations of bioprocesses with predictive capabilities. In biotechnology, Digit
- PDF / 1,066,942 Bytes
- 48 Pages / 439.37 x 666.142 pts Page_size
- 28 Downloads / 393 Views
Mechanistic Mathematical Models as a Basis for Digital Twins André Moser, Christian Appl, Simone Brüning, and Volker C. Hass
Contents 1 Introduction 2 Process Models as a Basis for Digital Twins 2.1 Submodel Framework of the Process Model 2.2 Submodel Framework of the Digital Twin 3 Modelling Approach 3.1 Model Development 3.2 Model Requirements 4 Model Types 4.1 Mechanistic Models 4.2 Non-Mechanistic Models 4.3 Hybrid Models 5 Model Based Process Optimization 5.1 Open Loop Control 5.2 Model Predictive Control 5.3 Adaptive Nonlinear Model Predictive Controllers 6 Case Study: A Generalized Structured Modular Model as a Basis for Digital Twins and Process Optimization 6.1 Structure of the Generalized Model
A. Moser Faculty of Medical and Life Sciences, Furtwangen University, Villingen-Schwenningen, Germany e-mail: [email protected] C. Appl and V. C. Hass (*) Faculty of Medical and Life Sciences, Furtwangen University, Villingen-Schwenningen, Germany Department of Biochemical Engineering, University College London, London, UK e-mail: [email protected]; [email protected]; [email protected]; [email protected] S. Brüning Thuenen Institute of Sea Fisheries, Bremerhaven, Germany e-mail: [email protected]
A. Moser et al. 6.2 Six-Compartment Model 6.3 Extension of the Model Structure 6.4 Adaption to Different Processes 6.5 Compartment Model as a Basis for Process Optimization 6.6 Basis for Digital Twins 7 Conclusions and Future Perspectives References
Abstract A future-oriented approach is the application of Digital Twins for process development, optimization and finally during manufacturing. Digital Twins are detailed virtual representations of bioprocesses with predictive capabilities. In biotechnology, Digital Twins can be used to monitor processes and to provide data for process control and optimization. Central and crucial components of Digital Twins are mathematical process models, which are capable to describe and predict cultivations with high fidelity. Detailed mechanistic models in particular are suitable for both use in Digital Twins and for the development of process control strategies. In this chapter the requirements that process models must fulfil in order to be used for process optimization and finally in Digital Twins will be described. Different types of models, including mechanistic as well as compartmentalized models, are outlined and their application in Digital Twins and for process optimization is explained. Finally, a structured, compartmentalized process model, which was specifically designed for process optimization and has already been used in Digital Twins, is highlighted. Graphical Abstract
Digital Twin
Control & Automaon Model
Mechanisc Model Reactor Submodel
Plant & Peripheral Model
Biological Submodel
PhysicoChemical Submodel
Keywords Digital Twin, Mechanistic models, Process optimization, Sixcompartment model
Mechanistic Mathematical Models as a Basis for Digital Twins
Abbreviations AA AI Amm ANMPC ANN CHO DNA DO DTP Et
Data Loading...
