The fluid applies a force to the structure, the structure deflects, and the changed shape alters the flow. Closing that loop inside the model is expensive, and it is not always necessary.
One-way is sufficient
The flow is solved, the pressure and shear stress are mapped to the structural analysis, and the structure is solved. There is no feedback.
The condition: the deflection must not meaningfully change the flow passage. A practical threshold is a deflection of the order of one percent of the characteristic opening. Thick-walled pressure vessels, casing stresses and pipe support loads fall into this class.
Two-way is required
If the deflection changes the passage. A thin blade, a flexible membrane, a check valve disc. See the moving boundaries note.
If instability is the question. Flutter, galloping, vortex-induced vibration. These are feedback phenomena by definition; a one-way analysis cannot even show they exist.
If added mass matters. A structure moving in a dense fluid has to accelerate the surrounding fluid as well. That raises its effective mass and lowers its natural frequency. A modal analysis in air therefore misleads for a structure that operates in water.
A numerical warning: added-mass instability
Where the fluid and structural densities are close (water-polymer, blood-vessel), a loosely coupled scheme can stay unstable even as the time step is reduced. This is counter-intuitive and often misdiagnosed. The remedy is not a smaller time step but a strongly coupled scheme.
What to report
If one-way was chosen, the justification: the expected deflection and its ratio to the flow geometry. Those two lines make the decision auditable.
References
- Ferziger, J. H., Perić, M. & Street, R. L. (2020). Computational Methods for Fluid Dynamics, 4th ed. Springer. DOI: 10.1007/978-3-319-99693-6
- Williamson, C. H. K. (1996). Vortex dynamics in the cylinder wake. Annual Review of Fluid Mechanics, 28, 477–539. DOI: 10.1146/annurev.fl.28.010196.002401