Seeking Advice on Minimum Time Step Rejection Issue

1

Hello everyone,I am also encountering the same issue regarding the rejection of the minimum time step and would like to seek your advice. I am currently attempting to replicate a test case from a published paper by Marius Ruby and Holger Foysi (2025).

D8{N}DDVUUFLSRN$Q%81HL

The computational domain is a rectangular cavity with the following dimensions:

  • x: [0, 4π]

  • y: [-1, 1]

  • z: [0, 2π]

The mesh settings are as follows:

  • Grid number: Nx = 96, Ny = 64, Nz = 76 (To save computing resources,I have scaled each direction by a factor of 0.5).

  • The mesh is uniformly distributed in the streamwise (x) and spanwise (z) directions.

  • The wall-normal (y) direction follows a hyperbolic tangent distribution.

  • No-slip isothermal boundary conditions are applied at the walls.

  • Periodic boundary conditions are used in the homogeneous streamwise and spanwise directions.

I have attached my configuration (.ini) for your reference. I would greatly appreciate your guidance on this matter.

Best regards

[backend]
precision = single                   
rank-allocator = linear

[constants]
gamma = 1.4
Pr = 0.7
mu = 0.0003333333333333333  
cp = 1004.5
Tw = 293

[solver]
system = navier-stokes
order = 3

[solver-time-integrator]
scheme = rk45
controller = pi
atol = 1e-5
rtol= 1e-5
tstart = 0
tend = 5.00 
dt = 6e-05

[solver-interfaces]
riemann-solver = rusanov
ldg-beta = 0.5
ldg-tau = 0.1

[solver-interfaces-quad]
flux-pts = gauss-legendre

[solver-elements-hex]
soln-pts = gauss-legendre

[soln-ics]
rho = 1.0
u = 1.0      
v = 0.0
w = 0.0
p = 7.936507936507937  

[soln-bcs-bottom_wall]
type = no-slp-isot-wall
cpTw = cp*Tw

[soln-bcs-top_wall]
type = no-slp-isot-wall
cpTw = cp*Tw

[soln-plugin-writer]
dt-out = 0.5
basedir = .
basename = M03R150_1-{t:.2f}

[soln-plugin-nancheck]
nsteps = 50

[soln-plugin-dtstats]
flushsteps = 50
file = dtstats2.csv
header = true

[soln-plugin-residual]
nsteps = 50
file = residual2.csv
header = true
Summary

This text will be hidden

2

Have you tried looking at the flow just before it diverges? You may need a start-up procedure (ramping the polynomial order, for example).

Regards, Freddie.

3

Thank you, Freddie. As I am relatively new to this field, I haven’t yet tried examining the flow field just before it diverges. I’m also not entirely clear on what you mean by the “start-up procedure” you mentioned. Additionally, I attempted to set the order to 1 for the calculation, but the same issue still occurred. Would it be possible for you to provide more specific guidance on this, or perhaps point me to any existing examples or references that I could follow?

Regards

4

Most well-written papers will contain a description of how a simulation was started. A few examples are given in:

http://aero-comlab.stanford.edu/Papers/witherden_jameson_aiaaj_2019.pdf

However, you really need to start looking at the flow before it diverges to diagnose what is going wrong.

Regards, Freddie.

5

Thank you for your reply. I will read this article carefully in the near future and learn from its description of how a simulation was started, in the hope that it will help me resolve this issue independently.

Regards