3 Digital Twins as Training and Educational Tools
Digital Twins or “Digital Twin-like” simulators may also be used in industry to train
reactor and plant operators and in academia to educate future control and process
engineers. In this context, Digital Twins are usually referred to as OTSs [9–11, 57].
OTSs became increasingly popular since the mid-twentieth century, for the use in
various sectors, including the chemical and related industries [10, 54]. The reason
was the increasing complexity of process engineering plants with sophisticated
automation and process control strategies placing enormous demands on the skills
of the process operators [10, 54]. Several papers were published reviewing the
development and use of OTSs in the chemical process industry [54, 58, 59].
OTSs offer the possibility to train future reactor operators and bioprocess engineers in a very practical way without carrying out the real process. Even actions to
compensate process malfunctions may be trained safely. Impairments on ongoing
production processes due to training are avoided. OTSs can be described as “earlystage” Digital Twins.
The development and use of OTSs particularly for bioprocesses are beginning to
attract increasing academic interest [10]. Several research groups have investigated
the applications of OTSs for bioprocesses. The common premise of the presented
research works confirms experiences from the chemical industry. Model-based
OTSs are an efficient means to improve the training experience of students and to
increase plant operators skills in handling complex bioprocesses [13, 14, 16, 60, 61].
Table 3 gives an overview of already existing OTSs for bioprocesses.
Hass et al. [17] developed one of the earliest OTSs for a complex biorefinery
process. OTSs were created for the bioethanol fermentation and the distillation
process. Also, a separate biomass power plant training simulator was developed.
The mathematical process models were created and implemented using the FORTRAN programming language [65]. The process control software WinErs [20] was
used to link process control and the simulation models. PCS-like GUIs were
developed to obtain full operator training simulators. Functions were implemented
to simulate the processes at different speeds depending on the desired training target.
The different OTSs were designed for the training of students as well as industrial
operators in the handling of biorefineries and biomass power plants. Encouraging
training outcomes were reported [10, 17].
A research project by Gerlach et al. [61] presented an OTS for the training of
bioengineering students and plant operators on the operational procedures and
production skills required in recombinant protein production processes. To enable
the model to accurately represent the complex relations of factors in a recombinant
protein production process, the authors outlined that several metabolic interactions
affecting biomass yield, productivity and cellular viability need to be mapped in the
OTS model. To maintain numerical efficiency, a trade-off between model complexity and accuracy had to be found by capturing the most important metabolic
processes in the OTS model, without the model being cumbersome and numerically
Digital Twins for Bioprocess Control Strategy Development and Realisation
75
Digital Twins or “Digital Twin-like” simulators may also be used in industry to train
reactor and plant operators and in academia to educate future control and process
engineers. In this context, Digital Twins are usually referred to as OTSs [9–11, 57].
OTSs became increasingly popular since the mid-twentieth century, for the use in
various sectors, including the chemical and related industries [10, 54]. The reason
was the increasing complexity of process engineering plants with sophisticated
automation and process control strategies placing enormous demands on the skills
of the process operators [10, 54]. Several papers were published reviewing the
development and use of OTSs in the chemical process industry [54, 58, 59].
OTSs offer the possibility to train future reactor operators and bioprocess engineers in a very practical way without carrying out the real process. Even actions to
compensate process malfunctions may be trained safely. Impairments on ongoing
production processes due to training are avoided. OTSs can be described as “earlystage” Digital Twins.
The development and use of OTSs particularly for bioprocesses are beginning to
attract increasing academic interest [10]. Several research groups have investigated
the applications of OTSs for bioprocesses. The common premise of the presented
research works confirms experiences from the chemical industry. Model-based
OTSs are an efficient means to improve the training experience of students and to
increase plant operators skills in handling complex bioprocesses [13, 14, 16, 60, 61].
Table 3 gives an overview of already existing OTSs for bioprocesses.
Hass et al. [17] developed one of the earliest OTSs for a complex biorefinery
process. OTSs were created for the bioethanol fermentation and the distillation
process. Also, a separate biomass power plant training simulator was developed.
The mathematical process models were created and implemented using the FORTRAN programming language [65]. The process control software WinErs [20] was
used to link process control and the simulation models. PCS-like GUIs were
developed to obtain full operator training simulators. Functions were implemented
to simulate the processes at different speeds depending on the desired training target.
The different OTSs were designed for the training of students as well as industrial
operators in the handling of biorefineries and biomass power plants. Encouraging
training outcomes were reported [10, 17].
A research project by Gerlach et al. [61] presented an OTS for the training of
bioengineering students and plant operators on the operational procedures and
production skills required in recombinant protein production processes. To enable
the model to accurately represent the complex relations of factors in a recombinant
protein production process, the authors outlined that several metabolic interactions
affecting biomass yield, productivity and cellular viability need to be mapped in the
OTS model. To maintain numerical efficiency, a trade-off between model complexity and accuracy had to be found by capturing the most important metabolic
processes in the OTS model, without the model being cumbersome and numerically
Digital Twins for Bioprocess Control Strategy Development and Realisation
75