Better way to initialize transient FSI simulation?

Hi all,

For my PhD project, I'm working on FSI models of abdominal aortic aneurysms and I'm facing a problem initializing my simulation.

Using a Fluent journal, I import an STL geometry in Fluent, mesh it and define my CFD model, including the dynamic mesh settings. Lastly, the solution is initialized using hybrid initialization. This causes my entire domain to be initialized with a pressure equal to the diastolic pressure (in my test case, this is 88 mmHg), just like I want.

After the hybrid initialization, my FSI simulation starts with solving the Transient Structural model, after which the data is transferred to Fluent and vice versa. When I look at the total mesh deformation in the first step, I observe that the deformation in the FSI model (right) is very small compared to the deformation in a structural model (left). I'm using the Quasi-Newton stabilization method to stabilize the forces transferred from Fluent to APDL. I've checked if the decrease in displacement in the FSI model is caused by the under-relaxation factor, but it turned out that the same results are obtained when the under-relaxation is set to 1 (so no under-relaxation at all).

I therefore wonder is its maybe better to start with a static 2-way FSI analysis and then restart this analysis as a transient FSI simulation. However, I don't have any experience with restarting FSI simulations, so I was wondering if you think this will solve my problem and have any tips on how to restart my FSI simulation.

Thanks in advance!

Kind regards,

Judith

Answers

Sign In or Register to comment.