laminar velocity profile
 

Hi
I need urgent help
I simulate a pipe with 100 micron diagonal, there is water inside pipe. the inlet velocity is 0.4 m/s, it's on laminar regime. when I plot velocity profile on output, it shows maximum velocity is 0.6 m/s. it's wrong because in laminar regime maximum velocity is 2*(Average velocity). plz help me to solve this problem

best regard

Hi
refine ure mesh.
Maybe a double precision solver will help.
Regards

Hi dear friend
I use double precision solver and it still has error
Plz help me to solve this
best regards

Hi
Refine ure mesh. Use a finer mesh and compare with the first one.
u do the same thing untill u get the same solution.
If u want to have the real solution of Navier STokes, use a boundary layer mesh also.

Remember that a solution that is not independant of the mesh size, is never so usefel in CFD.

Regards

Hi
thanks for ur answer
I want to simulate 2D rectangular channel, when I draw velocity profile on outlet it shows wrong maximum velocity
can u help me?
regards

How can u simulate a rectangular channel in 2D?????????????????????????????

excuse me I'm new in fluent
my thesis is about simulation microchannel with rectangular cross section
Do u think it is not possible to simulate it in 2D?

Hi,
Regarding your case, i have two comments:
First, why do you want to simulte this problem? In fact analytical solution exists, if your case is single-phase, laminar flow.

send, when you can simplify a 3D to 2D, it means in third direction you can extend your geometry without limitation(no viscous effect), hence, for your case, you can not use 2D model.

Anyway, I hope it is useful to you.

yes as chaozhong says, u cant extend ure geometry.
Unless the width of ure channel can be considered as infinity...

Anyway, flow in rectangular channel is highly complex because u have secondary and corner vortices motion. It stills an actual theme of research.

U have analytical solution for the laminar case only.

Ask me if u still need something

Hi,

Thanks for your indication about this field.

I thought it was just a straigh channel with regular cross-section.

Good Luck!

Hi
thanks for ur answering, actually my M.s thesis is about microcahnnel. I have to simulate a microchannel with rectangular cross section. it has 4 velocity inlet and flow must be fully developed. when I simulate it with 0.4 m/s it showed maximum velocity is 0.6m/s and it must be 0.8
can u help me?
regards

Hi
yes i can help u. No problem.
U must get 0.8 m/s for a fully developped laminar flow in a CIRCULAR channel. Not in a rectangular one.
I think u have a very bad professor that is helping u in this thesis.....

Get ure solution in a rectangular channel. Do anothere grid finer than the first and look if the solution is the same.

Hi,
In fact velocity profile is not so sensitive to your mesh, not like the pressure drop.

Of course, you could use the hagen-poiseuille law to obtain the velocity profile.

thank u so much
I attach picture of my microchannel, I obtain the analytical solution for velocity profile in rectangular channel, velocity profile in inlets agreed with analytical solution but when two inlet join together, the velocity profile doesn't agree with theory, I don't know how should I do?

Hi
i didnt understood ure geometry...where are the inlets??
what is ure Re???

Anyway look what u must do:
monitor variables in a point inside the domain and see how it behave. See if this variable become constant or if it do a cyclic behavior.

Hi I am new in CFD and writing code for laminar channel flow in 2D. But I am not able to decide the boundary and intitial conditions like u,v and pressure. I have used Poissuile formula for u velocity. But hot to initialize v and pressure. Thanks

If you don't define any local friction factor UDF for a 2D rectangular channel, you are effectively simulating a flow between two infinite parallel plates.
In that case, 0.6m/s is correct (Umax = 3/2*Umean for flow between two infinite parallel plates).

I have also noticed that for finer meshes, it takes many iterations (in the order of few thousands) to reach the theoretical value. For example, for one of the cases I was studying, with about 5000 iterations for ~400x1000 (Ny x Nx) , my Vmax/Vmean was around 1.3. But when I ran it for about 20000+ iterations, it converged to 1.5.