[Nek5000-users] Qustions about averaging for a 2D case

[nek5000-users at lists.mcs.anl.gov]

Hi,

I am running a 2D DNS simulation of backward facing step combustor with a
microjet. The microjet is placed on the step at 0.5H from the bottom wall.
The incoming velocity is U_o = 12.5 m/s and the microjet velocity is 25
m/s.

After running the case for 10 flow throughs, I started averaging over the
next 8 flow throughs. But the results for averaging don't exactly look
right. The geometry used and the averaged x-velocity colorfill contours
can be seen in the image at link below.

http://www2.uic.edu/~hkanch1/avg_mj10000.bmp

I have tried averaging twice, once over 5 flow throughs and other over 8
flow throughs. Both time getting similar results.

Here are the parameters that I am setting in the .rea file.

************************************************************************
 103   Parameters from file:/home/harish/nek5_examples/2D_bfs/mj_2d/mj1.rea
   1      1.00000         DENSITY
   2      1.556e-5        VISCOS
   4      1.00000         GTHETA
   7      1.00000         RHOCP
   8     0.100000E-01     CONDUCT
  11     128000.00          NSTEPS
  12     -0.0000010        DT
  15     128000            IOSTEP
  18     0.500000E-01     GRID
  19     -1.00000         INTYPE
  20      5.00000         NORDER
  21     1.000000E-07     DIVERGENCE
  22     1.000000E-09     HELMHOLTZ
  24     0.100000E-01     TOLREL
  25     0.100000E-01     TOLABS
  26     2.000000         COURANT
  27     3.00000         TORDER
  28     0.000000E+00
  49     0.000000E+00
  66     4.00000         write form (0), unform(1), universal(2)
  67     4.00000         read form (0), unform (1), universal(2)
  93     20.00000E+00      Number of pressure sol'ns saved
  94     5.000000E+00      start saving pressure soln's after this
  95     5.000000E+00      start saving pressure soln's after this
  99     3.00000            p99 = 3 --> dealias turned on
 103     0.01               103 filter weight

 IFTRAN   = T
 IFFLOW   = T
 IFHEAT   = F
 IFSPLIT  = T
 IFLOMACH = F
 IFUSERVP = F
 IFUSERMV = F
 IFSTRS   = F
 IFCHAR   = T
 IFCYCLIC = F
 IFAXIS   = F
 IFMVBD   = F
 IFMELT   = F
 IFMODEL  = F
 IFKEPS   = F
 IFMOAB   = F
 IFSYNC   = T
************************************************************************

At the outflow, I am using the "fill_div(usrdiv)" subroutine provided in
the "peris.usr" example. Is it possible the value of "vnmo_desire = 5.0",
that I set, is too high? I choose this value because at lower values the
simulation would blow up. Is there a way to calculate this value based on
incoming velocity?

Thanks for all the help,

Harish