Problems faced in 2-D poisuille flow

Hi Folks,
I have written a code in D2Q9 and I have wall on the top and bottom, my flow is basically in channel, where I have given inlet condition of constant velocity and outflow conditions at the outlet i.e.
at outlet:

X=Xmax+1
do y= 2,Ymax
f(3,x,y) = 2.*f(3,x-1,y)-f(3,x-2,y)
f(6,x,y) = 2.*f(6,x-1,y)-f(6,x-2,y)
f(7,x,y) = 2.*f(7,x-1,y)-f(7,x-2,y)
end do

The grid that I have taken is (500,100)
Uinlet =0.0667lu/ts
tau=0.8 (approx.)
Re=100.
I am using simple Bounceback boundary conditions at the walls.
My problem is at the outlet (exit) I should get Umax=0.1lu/ts(i.e. at the center) as inlet given is 0.0667 lu/ts, but I am getting maximum velocity as Umax = 0.092 lu/ts, hence the results are not matching with the theoretical results. I have checked the code many a times but I have not found any error in collision or bounce back or propagation.
I have just used constant inlet velocity and the outlet condition I have written in the above part of post. Kindly help me out with this, I am really puzzled.
regards
Prateek

Hi Prateek,

Does the mismatch persist when you scale it up to more nodes?

I’m no absolute expert on LB, only a humble user, but I could imagine that the bounceback has a bit too far reaching effects which might be decreased in a bigger lattice (currently it’s an 8% mismatch with only 50 nodes distance to the bounceback nodes, if I understand your setup correctly).

Or rather take one half the size and check if it becomes worse (saves computing resources ).

Best wishes,
Arne

PS: Could you post your code?