Environmental Engineering Reference
In-Depth Information
reflect the common practices used by operators in everyday bioreactor operation. Numerous
kinetic expressions for bioreactors have been formulated using this approach [21, 22], and
fuzzy modeling and control are now regarded as promising methods for automating industrial
bioprocesses in which experienced operators play significant roles in their successful
operation [22].
2.2. Transport Phenomena and Models
Transport equations predict gradients of dissolved substances, temperature, etc. within
fluids. They are based on principles of conservation of mass, momentum, and energy. When
applied to bioreactor operation, they can be used both to discover and explain phenomena of
interest, as well as to guide bioreactor design and scale-up. Because of the central importance
of reactor uniformity, transport modeling is frequently used to address issues of mixing, mass
transfer between gases and liquids, heat removal and/or maintenance of optimal microbial
growth temperatures, and biofilm formation on reactor surfaces. The distribution of fluid
velocities, or velocity profile, within a bioreactor is especially important in the calculation of
gas transport and heat transfer patterns as well as shear stresses that influence locations of
biofilm formation.
The complete simulation of mass and energy transport throughout all parts of a reactor,
showing consequences of adjusting design variables, would be ideal for reactor design. While
the typical two-phase gas-liquid media composition, locally turbulent flows, and limitations
of kinetic models greatly complicate the calculations involved in traditional transport models,
new approaches are being developed that hold great promise. In recent years, computational
fluid dynamics (CFD) in particular has enabled the capture of salient features in bioreactors.
For example, the simulation of a bubble column fermenter is shown in Figure 2.
Instantaneous values are shown on the left, while time averages are shown on the right;
velocity profiles are plotted with arrows while oxygen distributions are shown in color. For
these calculations, the two-phase gas-liquid system has been treated as a homogeneous
medium with a variable density that depends on the gas retention in the column, resulting in
visibly turbulent flow at short timescales. In this example, quantification of the spatial
variation of the oxygen combined with a kinetic model based on oxygen provided the rate of
product generation.
Other reactor types present even more challenging modeling problems. For example, in
STRs, involving impellers that rotate rapidly through a two-phase flow, boundary conditions
on the transport equations require the use of moving boundary computational grids [23, 24].
Despite the challenges, describing and predicting details of mass and energy flow throughout
bioreactors mathematically-including the realistic representation of fluids that exhibit non-
Newtonian behavior due to the suspension of cells and particulates-remain important goals
due to their great potential to facilitate reactor design [25].
2.3. On-line Sensing
Real-time sensing of bioreactor conditions, involving spatially resolved measurements of
fluid velocities and reaction components, is essential for both the experimental validation of
bioreactor models and for monitoring of ongoing performance. Even the most perfectly-
designed and thoroughly-modeled bioreactor is expected to experience unforeseen conditions
occasionally, particularly given the presence of mutable microorganisms, with the result that
real-time or on-line sensing of bioreactor conditions during operation is essential. On-line
