[vetri]

Hi all,

I'm trying to compare the flow patterns of the early stage of development of the wake behind an impulsively started cylinder with the experimental results done by Bouard and Coutanceau 1980.

I'm not able to get the flow pattern at t*=1(as per the paper), even though I'm able to get the other flow patterns(but at different t*s).Could anyone please help in this regard??? I've attached the cfx simulation results.

Thanks in advance...

Regards,
vetri

[ghorrocks]

http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

[ojha.mayank485]

Can you tell me what are you exactly comparing your results to ??? I am also tying to do similar stuff. May be I can help if you explain your problem a bit in detail.


-Mayank

[vetri]

Hi Mayank,

Thanks for your interest in the problem... I'm trying to match the t* values and the flow pattern behind the cylinder.I could not match any of the t* values.

But I'm able to get three patterns out of four as per the thumbnails posted.

I'm calculating t(time step) values as follows.

t=(t*)D/v

where t-> time step input in cfx

D->diameter of the cylinder

v->velocity

I'm calculating velocity input from Reynolds Number.If you need I could post the other boundary conditions also.

Thanks again...

Regards,
vetri

[ojha.mayank485]

Hi,

I did analysis over a stationary and a moving cylinder for quite a long time now. From what I understand is that the formation of vortex depends upon the numerical instability build as time passes.

For example, If you have a single cylinder which is completely structured mesh, the vortex formation takes time, if you plot Cl vs time you realise it. On the other hand if you use a triangular (2-D) mesh, the vortex vormation takes place quicker because the mesh is not totally symmetrical about the horizontal. Also if your initial condition is something like Uinitial=c*y (c=some constant & y=domain ht.) the vortex shedding is much faster.

When it comes to unstable solutions like vortex shedding from cylinders always measure factors like Pr. distribution, lift coeff, drag coeff only once the cylinder has reached the quasi steady state.

I have never heard anyone comparing the time values from experiment to numerical simulations especially when it comes to vortex shedding/formation

Hope this helps.


-Mayank Ojha