Multiphase Bioreactor Design

In the work of Loeblein et al. (1999) a method is presented for evaluating the economic
performance of the on-line optimisation control in batch reactors under parametric
uncertainty. Since the integration of the dynamic process model is required in the
optimisation, and since the integration may be time consuming and may preclude the on-
line implementation, they suggested to approximate the set of algebraic differential
equations in a set of algebraic equations by orthogonal collocation. In this way the
dynamic optimisation problem reduces to a nonlinear programming (NLP) optimisation
problem more easily solved on-line. The NLP optimisation problem can then be solved
using standard techniques like Successive Linear Programming (SLP) and Successive
Quadratic Programming (SQP).

Software Support for an Integrated Approach

Advanced methodologies for bioprocess monitoring and control include an integrated
utilisation of a multiplicity of tasks, as depicted in Figure 3.11. Such an approach can be
brought into practice only through an appropriate software environment, which should
make it easy to integrate all those tasks and methodologies. It is obviously difficult to
develop such a support tool. This can be only achieved by classification of tasks and
adoption of software engineering procedures suitable for programming such complex
systems. Some important concepts are:

1. Object oriented modelling with object oriented programming.


2. Tree-like model hierarchies/Tree-like objects with object inheritance.


3. Standard interfacing of procedures like Single Input Single Output (SISO) and
   Multiple Input Multiple Output (MIMO) computation elements, optimisation
   procedures, adaptation schemes, etc...

Multiphase bioreactor design 84