Biomedical Engineering Reference
In-Depth Information
TABL E 5.1. Operational Parameter Classification Based on Scale
Dependency (Assuming Equivalent Bioreactor Design)
Scale Factor
Scale-dependent parameters
Batch volume (L)
100
Feed-1 rate (g/h)
100
Feed-2 rate (g/h)
100
Agitation (RPM)
As per P/V
Scale-independent parameters
Seed density
1
Growth temperature (
C)
1
Induction temperature (
C)
1
pH set point
1
Induction duration (h)
1
Backpressure (psig)
1
modeling and qualification. The approach involves choosing a small-scale modeling
strategy, setting predetermined qualification criteria, performing small-scale runs, and
analyzing data to ascertain that qualification criteria were successfully met. Model
verification and refinement should occur over the life cycle of the product to incorporate
process changes made for process improvements and/or troubleshooting.
For the case study under consideration, the aspect ratios of vessel geometry with
respect to the diameters of the tank and impeller, liquid height, baffle width and number,
and sparger and impeller type were kept similar between scales. The operational
parameters that were selected and prioritized from the FMEA were divided into scale-
dependent and scale-independent operational parameters (Table 5.1). While the scale-
independent parameters were kept identical between scales, scale-dependent parameters
were scaled down linearly based on initial culture batch volume with the exception of the
agitation speed for which scaling was based on maintaining a constant power per unit
volume (P/V). Table 5.2 lists the various scaling strategies that can be used depending on
the application under consideration [15]. The P/V strategy is aimed at providing constant
k
L
a and thus similar gas transfer characteristics for controlling oxygen supply when
similar equipment geometry and sparger design are assumed. Since the characteristic
mixing times for a typical bench scale fermenter system are much shorter (measured in
seconds) than that for a manufacturing scale system (typically on the scale of minutes) at
similar P/V values, scaling down based on mixing time would require a significant
reduction in power per volume as described in Table 5.2, resulting in poor gas transfer
characteristics. The microbial cells in the case study are not sensitive to “shear” stress.
According to the literature, high local energy dissipation rates rather than tip speed are a
more reliable indicator of “shear” stress to cells [15]. Therefore, scaling agitation based on
tip speed relating tomanaging shear stress is not relevant in our case. Finally, scaling down
based on a constant Reynolds (Re) number does not apply when operating in the turbulent
flow regime in baffled fermenters. In this flow regime, the power number of the impeller,
which affects the delivered power per volume, becomes independent of Re number [16].
