Rapid B-rep model preprocessing for immersogeometric analysis using analytic surfaces.

Area of Science:

• Computational fluid dynamics (CFD)
• Geometric modeling
• Numerical analysis

Background:

• Traditional CFD requires complex, boundary-fitted meshes, leading to extensive geometry preprocessing for intricate shapes.
• Existing immersogeometric methods rely on NURBS parameterization, which can cause issues with surface quality and computational cost.
• Conversion to NURBS for analysis can introduce errors and necessitate further geometry inspection.

Purpose of the Study:

• To develop an enhanced immersogeometric method that utilizes analytic surfaces directly for CFD simulations.
• To eliminate the need for converting B-rep models to NURBS, thereby reducing preprocessing time and potential errors.
• To improve the efficiency and accuracy of CFD simulations for complex geometries.

Main Methods:

• Developed quadrature rules for analytic surfaces (planes, cones, spheres, toroids).
• Implemented adaptive quadrature for trimmed analytic surfaces.
• Extended the immersogeometric method to directly generate surface quadrature points from analytic surfaces.

Main Results:

• Simulations of flow over a torpedo shape using analytic surfaces showed good agreement with NURBS-based immersogeometric and boundary-fitted CFD methods.
• The method demonstrated effectiveness in high-fidelity aerodynamic analysis of industrial-scale models, such as a truck.
• Reduced model-preprocessing time significantly by avoiding surface type conversion.

Conclusions:

• The proposed immersogeometric method using analytic surfaces offers a faster and more robust alternative to NURBS-based approaches for CFD.
• This approach maintains accuracy for aerodynamic quantities while streamlining the simulation workflow.
• It is particularly beneficial for real-world geometries and industrial-scale applications.