Process optimization
With the help of our simulation model described in the previous chapter,
computational process optimization methods can now be used to
mathematically optimize process conditions and find the optimal
operating windows for the multi-enzyme cascade reaction sequence
implemented in our miniplant. More specifically, the Institute of
Process and Plant Engineering at Hamburg University of Technology has
developed the multi-objective process optimization tool “Advanced
Process Optimizer” (Adv:PO V0.7), which can be coupled with flow sheet
simulator tools such as Aspen Custom Modeler® . The
Adv:PO makes use of a genetic algorithm and has been successfully
applied to various chemical and biotechnological processes. The
implementation of this tool, together with our novel mathematical model,
allows for in-depth analysis and process improvement of the multi-enzyme
cascade reaction sequence to find optimal operating conditions.
During the optimization run, process variables, i.e. the concentrations
of each component, are varied between physically defined limits. Such
constraints are determined by component solubility and by the capacity
of the miniplant. Multi-objective mathematical functions are defined in
order to find the optimal solutions for any experimental variable of the
cascade reaction. In this paper we focus exemplarily on presenting our
results for maximizing CCI space-time yield in steady-state operation,
while at the same time minimizing the amount of the costly cofactors
NADH and NAD+. Such focus will broaden the
single-criterion-approach of increasing product space-time yield by the
economic factor of avoiding unnecessary amounts of costly substances.
The objective functions f1 and f2 in
equation (14) and (15) represent the maximizing of CCI space-time yield
and the minimizing of the cofactor input. They are complemented by
physical constraints. As shown in equation (16), the concentration for
each component must range within physically reasonable values. Moreover,
all dimensions, flow rates, temperatures, and kinetic parameters must
stay constant throughout the optimization run. Possible trivial
solutions of zero are not accepted: a minimum CCI space-time yield of
0.1 mmol/(l∙h) is set as a constraint for valid solutions.