Image-based meshing: As easy as pie

The study analyses the coupled heat transfer and fluid flow around a hot mince pie. A 60 g mince pie was scanned using a T1 weighted MRI scan. Three segmentation masks, representing the crust, filling and air, were generated using ScanIP segmentation software. A fully volumetric finite element mesh was then generated in the ⁺ScanFE mesh generation module and exported directly to FLUENT.

Picture of the mince pie

MRI scan of the pie

Segmentation of crust, filling and air

CFD Analysis in FLUENT

In FLUENT the physical properties were assigned and the filler temperature was set to 450 K. A space was created around the pie of 30 cm downstream and 4 cm either side of the pie and the region was set as a fluid domain with a 2 ms^-1 and 300 K inlet condition, akin to a short term cooling of the pie in a breeze. This was then used as the initial condition for a transient calculation (using PISO and 1^st order differencing in time) during which the filler was allowed to cool naturally in a 0.5 m/s breeze over a time horizon of 300 s. The results shown indicate the temperature profile, internal mesh and air flow velocity vectors during the cooling process.

Pathlines by velocity magnitude

Contours of static temperature

Contours of static temperature

The study not only shows the ease with which complex geometries can now be tackled, but also illustrates the following more general points pertaining to multi-physics simulations on complex image-based domains:

• The ability to generate watertight surface meshes and fully volumetric internal meshes of complex structures.
• The ability to mesh multiple structures, and to allow for arbitrary boundaries between them (i.e. fixed or moving).
• The ability to assign material properties within a given structure based on signal strength. In our example, this corresponds to fat content.
• The ability to export the resulting meshes directly to FLUENT, with no further refinements or modifications required.
• The ability to complete such accurate studies in a fraction of the time taken by other approximate approaches. In our example, the time taken to go from the import of the 3D data through to the creation of the multi-part input mesh for Fluent is less than 10 minutes on a PC.

Publication

Simpleware, 2006. Image-based Meshing: Easy as Pie. Fluent News, Spring 2006, 21. Download the article from the Fluent website. PDF (375 Kb)