Source model preparation

ProTOp relies on external numerical FEA models, here called the source models, that have to be

• prepared in any suitable third-party modeler and
• imported into ProTOp for subsequent optimization.

• PTC Creo® (FNF format)
• SolidWorks® (GEO and INP formats)
• Simulia® Abaqus (INP format)
• SIEMENS NX™ (DAT format)

Additionaly, any other modeler, which offers FE model export to one of the above mentioned formats, can be used.

In order to prepare an adequate source model, three fundamental issues need to be addressed carefuly. These three issues are the

• importer and solver considerations,
• units of measurement, and
• domain partitioning.

Importer and solver considerations

When preparing the source model, the importer and FEA solver restrictions have to be taken into account. In short, use:

• Materials: linearly elastic; isotropic or orthotropic
• Finite elements:
  • 4-noded tetrahedron
  • 8-noded hexahedron
  • 10-noded tetrahedron
• Loads:
  • concentrated force
  • pressure
  • traction
  • total force
  • acceleration and centrifugal loading
  • non-zero displacements
  • thermal

NOTE. The availability of these options may vary in dependence of the chosen modeler and importer.

NOTE. The source model can be imported and optimized successfuly only after it is compliant with these restrictions.

Units of measurement

ProTOp does not recognize or handle any units of involved quantities. Therefore, all data in the model has to be defined in coherent units of measurement. For numerical reasons, usage of

• mm (length)
• N (force)
• s (time)

and consequently

• MPa or N/mm2 (stress, traction, pressure, elasticity modulus)
• ...

is typically recommended.

IMPORTANT. When preparing the source model, use only coherent units of measurement for all involved quantities.

Domain partitioning

In order to assign various optimization features, configurations, and constraints to the underlying model, the whole design domain (given by all finite elements of the FEA model) typically has to be partitioned into subdomains, such as

• volume regions (volume parts of the whole design domain) and
• surface regions (patches of the design domain boundary surface).

For example, if some part of the model has to be kept fixed during optimization, the corresponding subdomain, has to be marked to be kept fixed.

Figure. Regions are needed to define optimization goals like declaring a region to be either fixed (gray) or free for optimization (blue).

ProTOp is built on the assumption that the source model is the only data available. Thus, the geometrical data describing individual subdomains has to be made accessible to proTOp in some way. Currently, ProTOp makes the following assumptions to get these data:

• A particular volume region is defined by a particular material region. In other words, all finite elements, linked to a particular material, define the corresponding volume region.
• A particular surface region is defined by a particular fictive surface load. In other words, all finite element faces, linked to a particular fictive surface load, define the corresponding surface region.

NOTE. In ProTop volume and surface regions data are extracted from material and loading data of the source FEA model.