Modelling Approaches for Bio-Manufacturing Operations - Measurement, Monitoring, Modelling, and Control of Bioprocesses - page 100

cost of the more expensive expanded bed matrices with lower resin lifetimes. The
effects upon cost of goods by parameters such as the yield emerging from the
expanded bed process as well as matrix cost, capacity and lifetime were examined.
Both direct product handling activities and ancillary tasks were modelled
explicitly to ensure that the costs included categories that might otherwise be
overlooked, and in this case, the expanded bed option was preferential owing to a
lower cost of goods and higher robustness.
Example 3
In a similar dynamic discrete-event example, Lim et al. [ 25 ] proposed another
hierarchically structured simulation which, alongside material balancing, cost
calculations etc., additionally included regulatory compliance activities such as
QA/QC, batch documentation and post-batch lot review. Animation features
enabled visualisation of batch (item) flow through the simulated process, thus
helping to identify and eliminate mistakes. Model outputs included annual
throughput, the total number of manufactured batches and resource utilisation
curves over time (thus identifying both resource under-utilisation and bottlenecks
hindering product throughput). The software was used to simulate the hypothetical
full-scale manufacture of monoclonal antibodies using perfusion culture to
determine the throughput, cost and resource impact of varying the culture broth
pooling frequency, assuming an unchanged downstream process. Determining the
optimal pooling interval is important to maximise resource utilisation but also to
minimise expenditure and contamination risks. One could either employ only a
small DSP capacity and then pool more frequently for purification or conversely
choose a plant with a larger DSP capacity with less frequent pooling of the
supernatant. Pooling more often makes greater use of DSP capital equipment since
there are larger numbers of downstream batches that need to be purified, although
with a smaller capacity the capital expenditure is lower in the first place. Using
equipment more often, however, increases the need for equipment preparation and
draws more heavily upon ancillary services, in addition to higher QC/QA costs
since these activities are completed more frequently. Conversely, less frequent
pooling makes lower demands for ancillary services or preparation of equipment
or regulatory requirements, but will lead to a higher capital investment. Product
stability is also an important factor since this affects the optimal choice of pooling
frequency. The authors used modelling approaches to examine the trade-offs
between such strategies.
3.5 Deterministic Versus Stochastic Modelling
As discussed earlier, modelling approaches should provide ideally not just a
deterministic analysis of feasible operating conditions but also an indication of the
impact of uncertainty within routine bio-manufacturing. Deterministic models will
assign just one value to every parameter, giving a 'single-point' output. In reality,
there are many possible risks that affect outcomes such as annual production levels