Learning Forum

[singh33]

Dear all,

I am simulating transient flow past a smooth circular cylinder, I have made a structured  C-grid using ICEM CFD having the orthogonal quality of 0.7 and total mesh count 9.36 lac, the RE for the analysis is Laminar 20, 100 and 150 only 2D vortex induction, Simple algorithm is used for Second-Order discretization for the Pressure and Power law for momentum and second-order implicit for time. 

Reference values are Diameter for the area and length, depth is 1m and all calculations are from the inlet, the time step size is 0.02 sec.

 

The problem is I am getting an amplitude fo the oscillating C_l and C_d value of 22% overpredicted than the experimental results. I tried mesh independent as well but that too is not fixing the issue and the domain I have chosen is as per one research paper.

Please help

Thanks to all in advance.

[Amine Ben Hadj Ali]

Please Second Order Upwind for Momentum and Bounded for Temporal discretization. What about the convergence level of your results? What is the velocity magnitude here? What is the period of oscillation? Are you resolving that? The area of projected area has to be used for the coefficients calculations

[singh33]

Dear abenhadj,

Thanks for the reply I have followed your tutorials prior to my every simulation on the same, 

Now coming to the question you are asking, after 7-8lac mesh count the parameters did not change the parameters of interest are C_l , C_d , vertex average Velocity which encircles the cylinder.  

The velocity magnitude as per the Re 150 is 0.107364 m/s I haven't change the density or the viscosity of the Ansys materials.

Strouhal No is coming right Prior to Re 150 I was simulating the Re 100 and Strouhal number is matching with the experimental results, but the problem of the amplitude of the C-l and C_d occurs with that as well.

The experimental strouhal no. 0.183 and from that time period comes out to be 1.272 sec I am resolvin that with around 60-time steps giving time step size of 0.02 sec.

 

Thanks and Regards.

[Amine Ben Hadj Ali]

What about the reference area for drag and lift coefficients?

[Amine Ben Hadj Ali]

I do understand your point with vertex average

[singh33]

Since it is a cylinder the reference area is the Diameter X Length since the flow is 2D I have taken the area as equal to Diameter i.e Length as 1m 

[singh33]

Vertex average, since to check for the convergence I create points around the cylinder and plot the value velocity at those points, to create an additional parameter for convergence as mentioned earlier 

[singh33]

Dear abenhadj,

Did you find something relevant to help me out?

Any sort of help will be highly appreciable.

Thanks and regards