NaN at low order

Errors
1

Hello,

I’m trying to simulate a simple building shape starting to simulate with P =1 for residuals and instabilies, but at this low order I’m geting the NaN error over and over agin. Am I missimg something?

Best

The ini file:

backend

precision = single
rank-allocator = linear

backend-cuda

device-id = local-rank
mpi-type = cuda-aware

constants

ac-zeta=40
;ac-zeta = 0.4
gamma = 1.4
Vin = 0
Win = 0
Pc = 101325
Pr = 0.72
nu = 0.000015

solver

system = ac-navier-stokes
order = 1
anti-alias = flux, surf-flux

solver-time-integrator

formulation = dual
scheme = bdf2
pseudo-scheme = rk45
controller = none
tstart = 0.0
tend = 5
dt = 0.5e-6
;dt = 0.0016
atol = 0.00001
rtol = 0.00001
pseudo-dt = 0.025e-6
;pseudo-dt = 4e-5
pseudo-niters-max = 20
pseudo-niters-min = 6
pseudo-resid-tol = 1e-4
pseudo-resid-tol-p = 1e-4
pseudo-resid-norm = l2
pseudo-controller = local-pi
safety-fact = 0.8
min-fact = 0.98
max-fact = 1.01
pseudo-dt-max-mult = 2.5

solver-interfaces

riemann-solver = rusanov
ldg-beta = 0.5
ldg-tau = 0.1

solver-interfaces-line

flux-pts = gauss-legendre
quad-deg = 3
quad-pts = gauss-legendre

solver-elements-tri

soln-pts = williams-shunn
quad-deg = 3
quad-pts = williams-shunn

solver-elements-quad

soln-pts = gauss-legendre
quad-deg = 3
quad-pts = gauss-legendre

solver-elements-hex

soln-pts = gauss-legendre
quad-deg = 3
quad-pts = gauss-legendre

solver-elements-pri

soln-pts = williams-shunn~gauss-legendre
quad-deg = 3
quad-pts = williams-shunn~gauss-legendre

solver-interfaces-tri

flux-pts = williams-shunn
quad-deg = 3
quad-pts = williams-shunn

solver-interfaces-quad

flux-pts = gauss-legendre
quad-deg = 3
quad-pts = gauss-legendre

soln-plugin-pseudostats

flushsteps = 10
file = pseudostats.csv
header = true

soln-plugin-nancheck

nsteps = 1

soln-plugin-writer

tstart = 0.0
dt-out = 0.005
basedir = /home/amir.shaulov/
#basedir = .
basename = Urban2-{t:.4f}

soln-ics

u =10
v = 0.0
w = 0.0
p = Pc

soln-bcs-walls

type = no-slp-wall
u = 0
v = 0
w = 0

soln-bcs-inlet

type = ac-char-riem-inv
ac-zeta = 40
rho=1.185
p = Pc
u = (0.49/0.41) * log((z+0.0032)/0.0032)
v = Vin
w = Win

soln-bcs-outlet

type = ac-char-riem-inv
ac-zeta=40
rho=1.185
p = Pc
u = (0.49/0.41) * log((z+0.0032)/0.0032)
v = Vin
w = Win

soln-bcs-slip

type = slp-wall
ac-zeta=40

soln-bcs-farfield

type = ac-char-riem-inv
ac-zeta=40
rho=1.185
p = Pc
u =10
v = Vin
w = Win

2

Do you have a rendering of the flow just before it NaNs?

Regards, Freddie.

3

This is the solution visualization (for second order solution) just before the error - the inlet boundary layer is almost interacts with the structure.

image
image712×515 48.9 KB

4

This just looks like your grid is extremely under resolved. How did you decide on your element count for the problem?

Regards, Freddie.

5

The elements near the wall are quite fine (y+~1) and the total cells number is about 350K. Basically this is a try to see if the results are reasonable.

6

Generally speaking, what is the recommended aspect ratio for the mesh? I’m sure that streched cells caused the simulation elapsing.

7

Generally, you want to keep the aspect ratio below 20:1, with 10:1 being preferred. This is substantially less than what you might encounter will RANS.

Regards, Freddie.

8

Thanks Freddie, that what I was aiming for.

Are the low aspect ratio regions shold be kept in regions far a way from viscous stresses (i.e. walls regions)?

best,

Amir