Modelling Approaches for Bio-Manufacturing Operations - Measurement, Monitoring, Modelling, and Control of Bioprocesses - page 90

engineering equations to quantify technical performance on one side and to
explore issues of expenditure, facility fit, scheduling and risk on the other. Process
modelling has focussed on relating input variables to outputs through explicit
mathematical relationships in order to complete material balances or to analyse
process sensitivities. In some cases, such models have been conceptually simple
enough to require only spreadsheets or non-specialist simulation engines to solve
the necessary equations [ 9 ]. In other cases, more complex modelling frameworks
have been used, such as computational fluid dynamics, in which powerful software
is used to simulate momentum or mass transport properties. A key challenge lies in
linking different model types together to provide whole process/facility level
understanding and not just evaluating the technical outcomes of a specific unit
operation. Traditional software approaches for bioprocess development have ten-
ded to make limited provision for incorporating business information. Growing
cost pressures mean that it is now becoming increasingly important to make
manufacturing decisions from both financial and process-related perspectives [ 29 ].
Business-process modelling has therefore gained significant ground in recent years
as a way of managing process development activities from both technical and
corporate perspectives.
2 Setting Up a Model
2.1 Establishing the Mathematical Basis for Modelling
When creating a model, it is important to realise that not every feature of a
bioprocess will be critical, and where possible, simplifying assumptions should be
made. This should apply both for making the equations themselves more man-
ageable as well as for minimising the data requirement. Of clear importance then is
availability in the literature of a sufficient, accurate set of values and default
choices for bioprocess properties. With the types of highly complex biological
material processed industrially, however, the assumptions underlying these data
may not be valid and experimental studies will then be needed to acquire the
appropriate values. Depending upon the depth of information needed, various
types of modelling equations can be formulated for a range of unit operations, and
thus the type and quantity of data will bear a close relation to the nature of the
models. These range from the simple to the complex and are specified according to
end-user needs; For example, a straightforward capture chromatography model
may assume a given dynamic binding capacity and then use that to evaluate
outputs such as processing times or consumable costs. On the other hand, more
complicated models will calculate dynamic binding capacity by quantifying
uptake and adsorption/desorption events from first principles using mass transport
equations, calibrated using experimental studies.