Methodology for Computational Fluid Dynamic Validation for Medical Use: Application to Intracranial Aneurysm

Area of Science:

• Biomedical Engineering
• Computational Fluid Dynamics
• Cardiovascular Flow Simulation

Background:

• Computational fluid dynamics (CFD) is valuable for diagnosing cardiovascular diseases but relies on simplifying assumptions.
• Existing CFD validation methods fail to isolate and quantify errors stemming from these modeling assumptions.
• Accurate error quantification is crucial for high-stakes clinical applications of CFD.

Purpose of the Study:

• To develop and demonstrate a novel validation methodology for CFD solvers.
• To quantify the specific model error introduced by CFD solver assumptions.
• To establish a streamlined workflow for determining the 'true' accuracy of CFD solvers.

Main Methods:

• Developed a validation methodology to identify and parse independent error sources in CFD and experimental data.
• Simulated patient-specific intracranial aneurysm flow using commercial CFD software (STAR-CCM+).
• Utilized Particle Image Velocimetry (PIV) for experimental validation data acquisition on orthogonal planes.

Main Results:

• The developed methodology successfully quantified CFD model error, averaging 5.63 ± 5.49% along the validation line.
• Demonstrated the superiority of the proposed validation method over three existing techniques using the same dataset.
• The method effectively isolates solver-induced errors from experimental and simulation uncertainties.

Conclusions:

• The novel validation methodology accurately quantifies CFD model error, enhancing solver reliability.
• This approach provides a more precise assessment of CFD accuracy compared to existing methods.
• The findings support the clinical translation of CFD by improving trust in simulation results.