2. A stirred tank reactor to represent overflow metabolism compartment.
3. A nonideal stirred tank with a strong bypass flow of a nonideal plug flow reactor
with high dispersion to reproduce residence time distribution of limitation zone.
The implementation of such designs may face technical challenges in reality
finally causing well-balanced decisions between simulation agreement and hardware
constraints. In general, such designs are in line with multi-compartment bioreactor
settings already used in the labs [60–62]. Accordingly, the combination with CFD
calculations allows a quantitative estimate of how close lab performance mimics
large-scale conditions.
4 Scale-Down Examples and Methods from the Literature
As mentioned earlier, the scale-down procedure should be carried out with the end in
mind, and this results in various optimization targets based on the scenario of
interest. With this in mind, we briefly go through the literature to evaluate different
design goals addressed using this approach. One of the first applications of scaledown can be seen from the work of George et al. [63] that was a two-compartment
scale-down with one stirred tank reactor (STR) connected to a plug flow reactor
(PFR). The impact of the overflow metabolism was investigated and to ensure such
condition substrate was fed at PFR entrance. One of the design considerations in
such setups is to maintain proper dissolved oxygen (DO) levels in the scale-down
reactor to prevent deviating from an industrial case. For this PFR should also be
aerated without major distributions in the flow pattern. The same setup was later
used to evaluate the effect of substrate heterogeneity of acetate formation [7]. One
way to achieve this is to use oxygen gas in PFR compartment to regulate the DO
levels. On a side note, other scale-down studies were carried before, but they simply
were not titled, so an example of this is the work of Sweere et al. [64] and Sweere
et al. [40] which was concerned with DO fluctuations by using STR compartments
while investigating the impact of the tubing. Hewitt et al. [59] took interest in
imperfections occurring at industrial scale in E. coli fermentations with high cell
density and STR-PFR scale-down reactor design. Enfors et al. [65] also used an
STR-PFR configuration to shed light on by-product formation in yeast for different
feeding locations in the scale-down reactor. With a different perspective, Delvigne
et al. [66] presented a stochastic approach to reproduce hydrodynamic flow fields in
the scale-down reactor based on residence time distribution for four different
configurations, namely, two STR-PFR with various PFR diameters, STR-2PFR
(parallel) and STR-2PFR (series). High flexibility and attractiveness of scale-down
approach resulted in its application in more complicated bioprocesses like algal
bioprocesses where fluid flow was simulated by CFD and the illumination was
included using the Monte Carlo method [67]. Even with only one or two compartments at hand, a diverse range of designs can be generated from a single vessel with
fluctuating feed to an STR-PFR with mixing and aeration in PFR and, hence, provide
Euler-Lagrangian Simulations: A Proper Tool for Predicting Cellular Performance. . .
245
3. A nonideal stirred tank with a strong bypass flow of a nonideal plug flow reactor
with high dispersion to reproduce residence time distribution of limitation zone.
The implementation of such designs may face technical challenges in reality
finally causing well-balanced decisions between simulation agreement and hardware
constraints. In general, such designs are in line with multi-compartment bioreactor
settings already used in the labs [60–62]. Accordingly, the combination with CFD
calculations allows a quantitative estimate of how close lab performance mimics
large-scale conditions.
4 Scale-Down Examples and Methods from the Literature
As mentioned earlier, the scale-down procedure should be carried out with the end in
mind, and this results in various optimization targets based on the scenario of
interest. With this in mind, we briefly go through the literature to evaluate different
design goals addressed using this approach. One of the first applications of scaledown can be seen from the work of George et al. [63] that was a two-compartment
scale-down with one stirred tank reactor (STR) connected to a plug flow reactor
(PFR). The impact of the overflow metabolism was investigated and to ensure such
condition substrate was fed at PFR entrance. One of the design considerations in
such setups is to maintain proper dissolved oxygen (DO) levels in the scale-down
reactor to prevent deviating from an industrial case. For this PFR should also be
aerated without major distributions in the flow pattern. The same setup was later
used to evaluate the effect of substrate heterogeneity of acetate formation [7]. One
way to achieve this is to use oxygen gas in PFR compartment to regulate the DO
levels. On a side note, other scale-down studies were carried before, but they simply
were not titled, so an example of this is the work of Sweere et al. [64] and Sweere
et al. [40] which was concerned with DO fluctuations by using STR compartments
while investigating the impact of the tubing. Hewitt et al. [59] took interest in
imperfections occurring at industrial scale in E. coli fermentations with high cell
density and STR-PFR scale-down reactor design. Enfors et al. [65] also used an
STR-PFR configuration to shed light on by-product formation in yeast for different
feeding locations in the scale-down reactor. With a different perspective, Delvigne
et al. [66] presented a stochastic approach to reproduce hydrodynamic flow fields in
the scale-down reactor based on residence time distribution for four different
configurations, namely, two STR-PFR with various PFR diameters, STR-2PFR
(parallel) and STR-2PFR (series). High flexibility and attractiveness of scale-down
approach resulted in its application in more complicated bioprocesses like algal
bioprocesses where fluid flow was simulated by CFD and the illumination was
included using the Monte Carlo method [67]. Even with only one or two compartments at hand, a diverse range of designs can be generated from a single vessel with
fluctuating feed to an STR-PFR with mixing and aeration in PFR and, hence, provide
Euler-Lagrangian Simulations: A Proper Tool for Predicting Cellular Performance. . .
245