fsi-fluid-structure-interaction

name	fsi-fluid-structure-interaction
description	Fluid-structure interaction (FSI) — one-way and two-way coupling, ALE formulation, partitioned vs monolithic approach, flutter, Abaqus-Fluent coupling, mesh morphing, ANSYS System Coupling.
metadata	{"priority":7,"promptSignals":{"phrases":["fluid structure interaction","FSI","aeroelasticity","two-way coupling","ALE formulation","flutter FSI","Abaqus Fluent coupling"],"minScore":3}}

Fluid-Structure Interaction — Complete Skill

FSI Categories

One-Way (Decoupled) Coupling

Fluid loads applied to structure; structural deformation NOT fed back to flow When valid: small deformations; flow not sensitive to structural shape change Process:

1. Run CFD → extract pressure distribution on wet surface
2. Map pressure to FE mesh (interpolation — nearest node, RBF, or projection)
3. Run structural FEA with applied pressure loads → deformation + stress

Limitation: misses load redistribution due to structural deflection; errors if deformation > 5% of characteristic length

Two-Way (Coupled) FSI

Mutual interaction: flow field affects structure; structural deformation changes flow When required: large deformations, resonance near flutter, biomedical (arteries, heart valves), ship slamming

Partitioned approach (co-simulation):

• Fluid solver and structural solver run independently
• Exchange boundary conditions at fluid-structure interface
• Iterate until convergence at each time step
• Tools: ANSYS System Coupling, OpenFOAM + OpenFSI, Abaqus + CFD

Monolithic approach:

• Single solver handles both fluid and structural equations simultaneously
• Better stability for strongly coupled problems
• Tools: OpenFOAM-based (solids4foam), Kratos, few commercial options
• Higher computational cost; harder to implement

ALE (Arbitrary Lagrangian-Eulerian) Formulation

Problem: fluid mesh must move with structure (Lagrangian) but avoid excessive distortion

ALE solution: mesh moves at velocity w (arbitrary, intermediate between material v and Eulerian 0)

ALE Navier-Stokes: ρ [∂u/∂t|_χ + (u - w)·∇u] = -∇P + μ∇²u

χ = reference coordinates (ALE); u-w = convective velocity relative to mesh

Mesh morphing/smoothing: after interface deformation, smooth interior mesh:

• Laplacian smoothing: Δx_interior from boundary conditions
• Radial basis function (RBF) interpolation: x_mesh = Σ w_i φ(|x - x_i|)
• Spring analogy: treat edges as springs; solve for equilibrium

Re-meshing: when mesh quality degrades below threshold → regenerate; interpolate solution

Partitioned Algorithm

Conventional serial staggered (CSS):

1. Solve fluid → get loads F^n at interface
2. Apply loads to structure → solve structural → get displacements d^n
3. Move fluid mesh by d^n → advance to next time step

Stability: conditionally stable; added mass effect causes instability for ρ_structure/ρ_fluid ratio < 10 (e.g., thin shells, biological)

Strongly coupled schemes (subcycling + Aitken relaxation): Iteration within time step until ‖d^{k+1} - d^k‖/‖d^k‖ < tolerance Aitken relaxation: ω^{k+1} = -ω^k × (r^k · (r^{k+1}-r^k)) / |r^{k+1}-r^k|² Prevents divergence from added-mass instability

Flutter Analysis (FSI Mode)

V-g method (frequency domain): [K - ω²M + iωC_aero(V)]u = 0 Solve for ω and g (structural damping required for stability) vs. V Flutter speed: V_F where g = 0 → zero damping → onset of divergent oscillation

Time-domain FSI flutter: Run coupled simulation across range of dynamic pressures Flutter onset: small perturbation grows exponentially in time

ANSYS System Coupling Setup

1. Define FSI interface surfaces in both CFD (Fluent/CFX) and Structural (Mechanical)
2. System Coupling: create Analysis → connect Fluid System and Structural System
3. Define data transfer: pressure (fluid → solid), displacement (solid → fluid)
4. Coupling settings: max iterations per step (3–10); convergence criterion (0.001–0.01 RMS)
5. Time step size: Δt = min(Δt_fluid, Δt_structural) typically

Abaqus-Fluent FSI

Co-simulation engine:

• Abaqus: CEL (Coupled Eulerian-Lagrangian) or standard with surface coupling
• Abaqus-Fluent: abaqus_cosimulation → exchange at wet surfaces
• Control: Abaqus as master; Fluent as slave in time stepping
• Or Multiphysics via Simulia/3DS platform

OpenFOAM FSI (Open Source)

solidDisplacementFoam + simpleFoam (one-way): map p from fluid to FEA solver FSI using OpenFOAM + preCICE (partitioned coupling library): preCICE handles data mapping and convergence; compatible with OpenFOAM, CalculiX, FEniCS

Benchmark Problems

• 3D elastic flag in channel flow (Turek-Hron benchmark): validates partitioned schemes
• Oscillating cylinder in flow: aerodynamic force + structural mode coupling
• Heart valve dynamics: large deformation + contact + fluid (most challenging)
• Wind-induced vibration of long-span bridges: flutter + buffeting

Key Non-Dimensional Parameters

Strouhal number: St = f D / V_∞ (vortex shedding frequency vs. body frequency) Reduced velocity: U* = V_∞ / (f_n D); resonance near U* = 1/St Mass ratio: m* = ρ_s / ρ_f; governs added-mass effect significance Cauchy number: Ca = ρ_f V² L / (E t) → ratio of dynamic pressure to structural stiffness

Output

Provide: FSI type (one-way/two-way), method (partitioned/monolithic), ALE vs. re-mesh strategy, coupling convergence criterion, time step, pressure and displacement fields at critical instant, flutter velocity V_F [m/s] if applicable, structural fatigue from dynamic pressure loading.