In this entry on Computational Fluid Dynamics (CFD) we are going to study a cylinder system common in flow conditioning. This example shows how the geometry of a device could be optimised to achieve a flow with the necessary properties for certain processes.
For this study we will use Acusolve software with Hyperworks CFD as pre-processor. A tutorial is available in our Downloads section where the steps to follow from the preparation of the model to the analysis of the results are detailed.
The following picture shows a schematic of the system. By using CFD simulations, the effect of different combinations of M and d parameters on the flow properties at the outlet of the device can be checked without the need for experiments.
For this case we will again use symmetry simplifications to obtain a smaller two-dimensional domain. In this way we will be able to obtain realistic results with a model of lower computational cost.
Pre-processing with Hyperworks CFD
The final geometry to be considered can be seen in the following image:
To carry out the optimisation, different combinations of M and d parameters are tested. Once the geometry has been validated, from the physics menu we select the Standard K-Epsilon model. This allows us to manually enter the turbulence parameters at the input, corresponding to the results of an experimental case.
In this case we define for the meshing a suitable boundary layer, a refinement zone around the cylinder, and a global mesh size of 1mm. Finally, we define the zones to export results from the Solution tape.
Solving and Post-processing
As main results we show the velocity and turbulence kinetic energy profiles in the flow. These first images show results close to the cylinders.
However, in this device it is the properties when the profile becomes uniform again that are of interest, thus studying the effect that the conditioner has had on the flow at different distances, as shown in the picture.
The tutorial also shows a process to obtain the numerical values of the different output results integrated on the surfaces. These can be used as input to other simulations or as data for other software where you want to process the data.
Conclusion: Optimisation
In an application case, the tutorial would be followed by an optimisation phase of the dimensional parameters M and d to obtain the desired flow properties at the required distance.
This example shows how CFD simulations can be useful in process optimisation from the design phase. This type of study can lead to significant savings in testing costs, as it allows the optimum parameters of a preliminary design to be fine-tuned to a large extent before manufacturing begins.
