Aortic Valve FSI Analysis
Aortic Valve FSI Analysis: Theoretical Foundations
The aortic valve opens and closes with each heartbeat, undergoing approximately 100,000 load cycles per day. The opening and closing behavior of the valve leaflets is determined by the strong interaction between blood flow and valve tissue, so analysis of fluid alone or structure alone is insufficient. This technology is essential for design optimization of prosthetic valves (mechanical valves, bioprosthetic valves, TAVI valves), predicting the progression of valvular disease, and surgical planning.
What is the mechanics like when the valve opens and closes?
When left ventricular pressure exceeds aortic pressure during systole, the valve opens and blood is ejected. During diastole, the valve closes due to the reverse pressure gradient. The leaflets are thin tissue about 0.5mm thick, undergoing repeated large deformations. Vortex flow forms in the Valsalva sinus downstream of the valve, and Leonardo da Vinci already observed that this vortex assists in valve closure.
Governing Equations
What equations are solved in valve FSI?
The fluid side uses the incompressible Navier-Stokes equations. The Reynolds number reaches about $Re \approx 5000$ to $8000$ at peak systole, and turbulent transition should sometimes be considered.
The structural side models the leaflets as a hyperelastic shell. Constitutive laws like the Fung-type or Lee-Sacks type, which consider fiber reinforcement, are used.
Here, $E_{11}$ is the Green-Lagrange strain in the fiber direction (circumferential), and $E_{22}$ is the strain in the orthogonal direction.
How is leaflet contact handled?
The three leaflets contact each other (coaptation) during valve closure. This is the biggest technical challenge in aortic valve FSI. The structural side needs to handle contact mechanics, and the fluid side needs to handle the limit where the gap becomes zero. The advantage of the Immersed Boundary method is that it can handle contact without changing the fluid mesh topology.
The Aortic Valve Endures 100,000 Openings and Closings Per Day
The aortic valve leaflets are tissue only about 0.5mm thick, opening and closing about 100,000 times a day while withstanding a pressure difference of about 120mmHg (16kPa) during systole. The design life is about 30 years. The anisotropic structure, where collagen fibers are arranged in a cross-hatched pattern, supports this incredible durability. To reproduce this fiber structure in FSI theory, a simple isotropic elastic model is insufficient, and hyperelastic models like the Fung-type are necessary. The complexity of the theory is a reflection of the amazing nature of biological tissue.
Computational Methods for Aortic Valve FSI Analysis
There are three main approaches.
| Method | Advantages | Disadvantages |
|---|---|---|
| ALE-FEM | High interface accuracy | Mesh breakdown during valve closure |
| Immersed Boundary (IB) | Strong for Large Deformation and contact | Interface smearing |
| Immersogeometric | IGA accuracy + IB flexibility | Complex implementation |
Immersogeometric? That's unfamiliar.
A method proposed by Kamensky, Hsu, Bazilevs (2015), which embeds a NURBS-based leaflet model into a fixed fluid mesh. It overcomes the weakness of the IB method (delta function smearing) with interface conditions using the Nitsche method. The Bazilevs lab at UT Austin leads this field.
IBAMR/IBFE Implementation
What specific software is used?
IBAMR (Immersed Boundary Adaptive Mesh Refinement) is the leading open-source example. Developed by Professor Griffith (UNC), it embeds the FEM structure of leaflets into the fluid on a SAMRAI adaptive mesh using the IBFE (Immersed Boundary Finite Element) method.
The fluid solves the Navier-Stokes equations with a penalty-based IB method.
The leaflet elastic force $\mathbf{F}$ is calculated by FEM and spread to the fluid grid via the delta function.
Time Step and CFL Condition
How small a time step is needed?
Valve opening/closing occurs in about 30ms (from systole onset to full opening). With the CFL Condition $\Delta t \leq h/|\mathbf{u}_{max}|$, peak velocity 1.5 m/s, and minimum grid spacing 0.1mm gives $\Delta t \leq 67\mu$s. In practice, calculations use $\Delta t = 10$ to $50\mu$s. Simulating one heartbeat (0.8s) requires 16,000 to 80,000 steps.
The computational cost is enormous.
Yes. Using AMR (Adaptive Mesh Refinement) to refine only near the valve is essential. In IBAMR, the region within a few mm of the leaflet surface uses the finest grid, while areas farther away use coarser grids, reducing computational cost to less than 1/10.
Choosing Between IBM and ALEโDecision Points for Aortic Valve Analysis
The choice between "Immersed Boundary Method (IBM) or ALE method" is always debated in numerical methods for aortic valve FSI. IBM is advantageous for handling large deformation and contact, but boundary accuracy tends to suffer. ALE has high accuracy but requires measures against mesh collapse during complete closure. In practice, some research groups adopt a hybrid strategy of "switching to ALE during opening and IBM during closure." Which one you choose greatly changes code complexity.
Aortic Valve FSI Analysis in Practice
A typical workflow is as follows.
1. Valve Geometry Model Creation: Create CAD from parametric shape (leaflet height, coaptation height, annulus diameter) or reconstruct patient-specific shape from echocardiography/CT.
2. Aortic Sinus (Valsalva Sinus) Model: Construct fluid domain from annulus to ascending aorta.
3. Mesh Generation: For leaflets (shell or solid) and fluid domain. For IB method, leaflet mesh is independent of fluid mesh.
4. Material Parameter Setting: Hyperelastic constants for leaflets. Identified from biaxial tensile test data.
5. Boundary Conditions: Left ventricular pressure waveform at inlet, aortic pressure waveform or three-element Windkessel at outlet.
6. Calculation Execution: Three or more heartbeats (to remove initial transients).
7. Postprocessing: Effective Orifice Area (EOA), pressure drop, regurgitant volume, leaflet stress.
What is EOA (Effective Orifice Area)?
Abbreviation for Effective Orifice Area, the most important performance indicator for prosthetic valves. Defined by the Gorlin formula.
$Q_{rms}$ is RMS flow rate (mL/s), $\Delta p_{mean}$ is mean pressure drop (mmHg). For TAVI valves, EOA > 1.0 cmยฒ is considered good.
Material Parameter Identification
How are the leaflet material constants determined?
Fit Fung-type or Lee-Sacks models to biaxial tensile test data. For native valves, $c_0 = 2$ to $10$ kPa is typical, and the fiber direction stiffness parameter $c_1$ is typically 5 to 20 times that of the orthogonal direction $c_2$.
| Valve Type | $c_0$ (kPa) | $c_1$ | $c_2$ | Source |
|---|---|---|---|---|
| Native Aortic Valve | 2โ10 | 10โ50 | 1โ5 | Billiar & Sacks (2000) |
| Bovine Pericardium (TAVI Valve) | 5โ20 | 30โ80 | 5โ15 | Varies by product |
| Porcine Valve (Bioprosthetic) | 3โ15 | 15โ60 | 2โ10 | Stella & Sacks (2007) |
How are calcified valves handled?
In aortic stenosis, calcification progresses on the leaflets. A common approach is to extract calcified regions from CT and set the elastic modulus of those parts 10 to 100 times higher. The degree of calcification directly affects paravalvular leakage (PVL) after TAVI valve deployment, so accurate modeling is crucial.
The On-Site Struggle of Reconstructing Valve Geometry from CT Images
In the practical work of aortic valve FSI analysis, "patient-specific geometry reconstruction" takes the most time. Segmenting leaflet shape from CT or MRI images is a difficult task, taking even experienced researchers several days per case. Recently, deep learning-based automatic segmentation has emerged, reducing time to a few hours. Still, it's a world where "80% of analysis quality is determined by geometry quality."
Experience the theory firsthand with the interactive simulator for this field
All Simulators