Turbulence in blood

Transient turbulence in a Fontan patient's circulation.

Problem

Most Fontan CFD studies rely on turbulence models that smooth out fine-scale flow structures. I examined whether those missing scales affect clinically relevant hemodynamic indicators linked to thrombosis.

Approach

  • Perform extreme-resolution direct numerical simulation (DNS) with LBM in a patient-specific TCPC.
  • Compute wall shear stress, wall shear rate, elongation rate, and flow-type metrics.
  • Run matched LES cases and compare metric distributions against DNS.
  • Quantify where model-form assumptions alter risk interpretation.

Key finding

LES and DNS produced meaningful differences in regions relevant to thrombosis indicators. In particular, LES tended to overestimate low-WSS regions and underrepresent elongation-dominated flow structures.

Why it matters

High-fidelity simulation can change clinical interpretation of risk markers. The DNS setup also provides a benchmark for future Fontan hemodynamics research.

Outputs