[FLASH-USERS] Discrepancies in Z-Pinch MRT Instability Growth Rates

[Bowen Zhu (Student at CentraleSupelec)]

Dear Community,

I am reaching out for guidance on an issue I've encountered while simulating the magnetic Rayleigh-Taylor (MRT) instability using the Flash code, with pre-seeded perturbations. My goal is to validate my simulation results against a reference study conducted using the MACH2 code.

I've configured the Flash simulation based on the ZPinch example, modifying the initial conditions to match those of the MACH2 study. Specifically, I've altered the initial density and geometry to align with the reference. However, I'm observing that the MRT instability grows significantly faster in my Flash simulations.

My modifications to the Flash ZPinch example include:

  *
MGD, opacity and radiative transfer are disabled, as they are not critical in the run-in phase.
  *
I've switched to using eos_gam with the atomic number (A) and charge (Z) corresponding to tungsten, along with a step-like initial density profile of 1.3e-3 g/cm3 and a thickness of 1mm.
  *
The simulation domain is set to 1.5cm by 1cm with the inner radius of the z-pinch shell initially at 1cm.
  *
I've seeded density perturbations of various kinds along the z-axis and analyzed the growth by applying a Fourier Transform (FFT) to the density field, excluding the DC component.

To investigate the issue, I set the initial perturbation seeding amplitude to zero, anticipating no growth in non-DC modes. Contrary to expectations, these modes appear once the current reaches 5MA. A resolution study, using the FFT integral of these modes as an indicator, showed that higher resolution does not slow the growth but rather accelerates it. Also the non-zero initial integrals presumably suggest that the step-like density may be introducing RT instability via density (pressure) gradients.

I have also examined the impact of dtmax and concluded it is not the cause.

Given that this setup is inherently fragile regarding RT instability and that numerical errors could aggregate, leading to inaccuracies, I am seeking advice on parameter adjustments to counteract these unphysical results. Could you suggest any changes( to the solver, initial condition, boundary conditions, or artificial viscosity etc) that might help?

Your expertise and assistance are greatly valued, and I look forward to any recommendations you may have.

Best,
Bowen ZHU

[cid:e86abb76-b896-4b12-8259-0b5071aaa26e]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20240409/1bfe8ccc/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 80966 bytes
Desc: image.png
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20240409/1bfe8ccc/attachment.png>