Coupling Process Simulation and Computational Fluid Dynamics to Efficiently Design and Improve Multi-Scale Processes

Almost all real-world processes are multiscale processes, which mean that they have important features at different scales in time and/or space. Typically, not all scales can be resolved, so that for simulations of such processes certain assumptions or simplifications have to be made.

In the process industry for example, simulations are performed on a range of different scales and different purposes. Since most processes depend, more or less, on local conditions, there is an obvious benefit to integrating process models into CFD. In practice, however, the applicability of this approach is limited mainly for the following reasons:

• Depending on the system under investigation, the additional equations to be solved can generate a considerable overhead and lead to long computing times.
• The chemical-physical processes often happen on different time scales to the flow and mixing time scale.
• For design studies or variation calculations during design phase as well as during operation, results are required in a short time, usually in seconds or in a few minutes. This typically cannot be achieved with a CFD simulation.

Nevertheless, combining different scales can be very beneficial in terms of accuracy as well as getting a deeper understanding of the underlaying physics as previously shown e.g. for a wet granulation process [1].

In this contribution we show a new automated coupling method between the CFD tool Simcenter STAR-CCM+ and the process modelling tool gPROMS by PSE, where Simcenter STAR-CCM+ solves for all flow properties and gPROMS evaluates the fine scales based on the solution provided. The coupling between both codes is done in an automatic way. This includes the domain decomposition, the calculation of and the transfer of all relevant variables. The details of the coupling as well as the superiority of this approach will be demonstrated for a packed bed reactor.