I am a new preCICE user with about 3 months of experience. I am performing a fluid-structure interaction simulation of the human aorta using OpenFOAM, preCICE, and CalculiX.
Problem description
As shown in the figure, after 0.025 s, the displacement of the watch-point on the aorta starts to grow exponentially. My arterial wall is defined as a linear elastic material.
At around 0.066 s, the maximum mesh displacement reaches 6.8 mm, while the aortic wall thickness is only 2 mm.
In the OpenFOAM log, the pressure Poisson equation outputs NaN, and ParaView shows that the simulation diverges due to excessive mesh deformation.
Physiologically, the maximum aortic displacement should be only 1–2 mm, not 6 mm.
What I have tried
I first thought the solid mesh in the thickness direction was too coarse. I increased the elements through the wall thickness from 2 layers to 3 layers, but the result did not improve. The displacement still blows up after 0.025 s.
The standalone OpenFOAM simulation runs stably without any issues.
I don’t think I described the problem clearly enough. The geometry of my aorta includes an inlet, which is the ascending aorta, and an outlet that comprises the three vessels above the aortic arch (BCA, LCCA, and LSA) and the descending aorta (DA). In fact, this is very similar to a three-dimensional elastic tube.
I also have some new observations and test results to share:
Although the long‑term displacement trend (exponential growth after 0.025 s) is similar for both meshes, their early‑time behavior is clearly different：
2‑layer mesh: displacement rises in an oscillatory, wave‑like pattern
3‑layer mesh: displacement increases almost linearly with time
If the pressure/force oscillates, I believe the displacement should also oscillates, but I previously used Rayleigh damping in my CalculiX input file.
When I removed Rayleigh damping, the displacement at the same point became completely non‑physical and erratic (I can hardly describe its behavior). And the simulation diverged on ~0.002 s.
I suspect that the inlet pressure oscillation may be causing the force oscillation, which then leads to the unphysical displacement growth and eventual divergence.I will continue testing and update my results.Any suggestions or possible causes are highly appreciated!
I have attached the Calculix input file, maybe it will help anyway.
I don’t think I can help much with the modeling, but:
Are you using a fixed or an adaptive time stepping in OpenFOAM? FSI with subcycling needs special treatment, and we do not have examples for that at the moment. Using the same solver time step size as coupling time window should help.
Any particular reason to still use preCICE v2? It should work as well, but preCICE v3 would give you some benefits.
