|
Process Understanding and Process Analytical Technology
Achieving dynamic process control, where process operations can be altered whilst maintaining regulatory committment, and where quality can be predicted at different stages of the manufacturing process, clearly requires a deep level of scientific understanding. Defining and understanding the multivariate connections that link together product design, process design, process parameters and raw material quality is the first step in achieving this. A crucial requirement is therefore a broad and deep scientific knowledge base.
LyraChem's mechanism-based, process modelling approach is ideally suited to achieve this level of understanding. Overall, our development strategy offers numerous advantages in chemical process development:
- Data obtained from early lab experiments allows an early assessment of critical engineering design parameters (e.g. batch-times, reactor volume, heat/cool/chill requirements) to be made.
- Different reactor operation modes (batch, continuous etc.) can be quickly explored non-experimentally. An early assessment can be made into, for example, the feasibility of a particular process step for continuous operation
- Data obtained from the lab and used in simulation experiments allows process chemistry to be matched to particular plant specifications. Technical specifications for most known reactors are contained in our simulation databases; reactors not in the database can be easily profiled
- Variability in product quality attributes resulting from changes to process operations can be discovered non-experimentally using simulation. Design and control space can be quickly defined and validated with minimal experimentation.
- Use of computer-based simulations allows parameter space to be investigated non-experimentally. Assumptions can be subsequently validated by a small number of experiments. This can greatly reduce development timescales.
- A parallel approach, where DoE and process simulation are coupled together, allows more shortcuts to be taken through design space in order to discover boundary conditions compared to a ‘DoE – only’ approach. This is because process models used in simulations are essentially mathematical representations of the fundamental chemical and physical properties that control the process. Optimal ranges for process input variables can be discovered quickly by simply varying input parameters within the simulation environment and then observing the effect on process outputs. DoE can then be applied to further optimise the process prior to experimental validation based on the outputs from process modelling studies, although this is often not necessary.
|
