Computational Fluid Dynamics Turbulence Model and Experimental Study for a Fontan Cavopulmonary Assist Device

Area of Science:

• Biomedical Engineering
• Fluid Dynamics
• Medical Device Design

Background:

• Fontan cavopulmonary assist devices (CPADs) require specialized computational fluid dynamics (CFD) simulations.
• Conventional Reynolds-Averaged Navier-Stokes (RANS) CFD models, typically used for ventricular assist devices (VADs), show limitations for CPADs due to their unique design (von Karman pump with large clearances).
• High-fidelity large eddy simulation (LES) is accurate but often cost-prohibitive for commercial optimization.

Purpose of the Study:

• To benchmark various RANS CFD turbulence models against experimental data and LES results for CPAD simulation.
• To identify the most suitable CFD models for different aspects of CPAD design and analysis.
• To provide guidance on selecting cost-effective CFD approaches for CPAD development.

Main Methods:

• Experimental head-flow (HQ) curves were measured for a CPAD using a blood surrogate in a mock circulatory loop.
• Multiple RANS turbulence models (standard k-ϵ, re-normalization group k-ϵ, realizable k-ϵ, SST k-ω, SST with transitional turbulence, Generalized k-ω) were employed for CFD simulations.
• Simulation results were compared with experimental measurements and high-fidelity LES data.

Main Results:

• The standard k-ϵ model demonstrated the best agreement with experimental pressure head predictions.
• RANS models showed varying accuracy for head and power predictions compared to LES.
• SST and LES models yielded near-identical torque values, while k-ϵ predictions were approximately 30% lower than LES.

Conclusions:

• For Fontan CPAD development, LES is recommended for final design simulations.
• The k-ϵ model is suitable for general flow and pressure head simulations.
• The SST model is recommended for predicting power, shear stress, hemolysis, and thrombogenicity.