Oscillating pressure curves

[Giuseppe Dalla Vecchia]

Good evening, I am encountering an issue with the results of my simulations. I successfully ran the Tutorial1 simulation without any problems with the results. To continue my study with simplicity, I used the inflow data from Tutorial1 with my own problem, and there too I had no issues; the pressure/flow curves were normal and not oscillating. Now, when I used my own flow waveform at the inlet with a certain number of points, I am getting pressure curves (and subsequently also flowrate) with many oscillations (see figure).

pressure-result.jpg (277.85 KiB) Viewed 7975 times

I have already tried these solutions: initially, I had a series of points that created a discontinuity at certain points see figure below

old and new points.png (94.79 KiB) Viewed 7975 times

I thought that was the error. So, to facilitate the Fourier curve fitting process, I interpolated my data series with MATLAB and created a series of new points (red line). This didn't solve the problem; I tried increasing the points (decreasing the time interval between two points) as well as decreasing them, I also tried increasing the Fourier points, but nothing.

I attach some files so if someone can understand whether this is causing these oscillations or if there is something else. Thanks to anyone who can help me.

https://drive.google.com/drive/folders/ ... drive_link

Giuseppe

[Giuseppe Dalla Vecchia]

Update: I've tried another project, with a different model/mesh and with its own inflow conditions as well as RCR conditions. This time, the results are better. I can't figure out what might have caused the error I mentioned earlier. My doubts lie in these areas: the choice of the Fourier number, possible errors in the mesh, and optimizing simulation settings. Note that I've conducted other simulations with different meshes and conditions, always using the same simulation settings but varying the number of timesteps and delta_t.

Any suggestions on what could be causing these errors?

Thank you,
Giuseppe.

[Weiguang Yang]

Could you grant me the access to your files so that I can take a look?
Thanks!

[Giuseppe Dalla Vecchia]

Could you grant me the access to your files so that I can take a look?
Thanks!

I should have shared the folder with you. Giuseppe

[Weiguang Yang]

Regarding solver.inp, I don't notice anything particularly wrong. Could you please specify what parameters you set in the svpre file?

I may try to downsample your inflow.flow from 781 to approximately 100 data points while applying some smoothing.
For the svpre file, consider reducing the number of data points for bct.dat. For instance:
"bct_point_number 100
bct_fourier_mode_number 10"

In addition, is your geometry similar to the demo model?
Thanks

[Giuseppe Dalla Vecchia]

Good morning, and you're welcome for the response.
I could certainly try reducing the number of points in my inflow.flow file again, although I've already attempted that. I also tried setting the Fourier value to 10. However, I can give it another try.

As for the geometry: it differs from the base model (although I should ask for permission to show it). For this geometry, I conducted studies by imposing the same inflow condition as used in the tutorial (inflow3d.flow, 3201 data point, womerslay, 200 fourier mode)... it calculated pressure and flow rate (uncorrect for my model) but without the oscillations I'm mentioning.

Now that I've found an inflow curve for my model, I wanted to use that for the simulations, but I'm encountering these issues. As adjustments to the curve, I ensured that the first point is equal to the last one to have a continuous curve. I thought this would solve the problem, but nothing changed.

Do you think it's a problem with the parameters for the inflow? Because I thought that somehow the geometry or the simulation parameters could be involved.

P.S. I added the .svpre file on drive (name: 002_SCa.svpre).

Thanks,
Giuseppe

[Giuseppe Dalla Vecchia]

I found out what the problem was. It was a high tolerance on the residuals. By setting a lower value, the oscillations disappeared.

Thank you anyway,
Giuseppe