search

KeitaNakamura / Tesserae.jl

A Julia-powered toolkit for material point method
https://keitanakamura.github.io/Tesserae.jl/
MIT License
7 stars 0 forks source link

Possible bug report about fluid simulation #4

Open Darkhell101 opened 3 days ago

Darkhell101 commented 3 days ago

Dear Keita Nakamura, I greatly appreciate your Julia library "Tesserae.jl." And I have run some of the listed tests, but encountered errors during the fluid simulation test. The related files are attached for your review. My julia version is 1.11.1. System information: macos, ubuntu and windows11 (All give error information and erupt.) If you could check them, it would be very helpful for other users of your library. Additionally, I would like to learn more about implicit formulation cases. Could you please provide additional example cases? Thank you for your time and assistance. file.zip

KeitaNakamura commented 2 days ago

Thank you for the report. However, I was unable to reproduce the errors. Could you please confirm the following?

For the implicit formulation, please open a separate issue.

Darkhell101 commented 2 days ago

Sorry to tell you, code still erupt at approximately 50% progress when Δt = 1.0e-3. Could you send your running script of this example here? My running file is copied from tutorial. I want to check whether they are the same.

Darkhell101 commented 2 days ago

https://github.com/user-attachments/assets/c71110c2-6d6b-49a7-99df-b1e2439193f4

https://github.com/user-attachments/assets/c607f791-82cc-4cce-9412-6b9eabb96af5

I have uploaded videos on both Windows and Linux. Linux displays error messages, whereas Windows does not. The progress bar reaches 100% rapidly, but no actual processing takes place.

KeitaNakamura commented 2 days ago

I mean, I couldn't reproduce the errors using the original tutorial file. What happens if you visualize it using Preview?

Darkhell101 commented 1 day ago

When viewed in ParaView, only the first three seconds of the animation play; the rest is blank.

KeitaNakamura commented 1 day ago

I reproduced the error on CI and am now checking the details.

Darkhell101 commented 1 day ago

I found some information that might be helpful. First, changing the base parameters of the input doesn't alter the result. However, the parameters used in Newmark's method appear to differ from those recommended by Wikipedia. My understanding of integration algorithms is limited, and there could also be other issues at play.