Validating HyperWorks CFD X 2019 by calculating Drag Coefficient of Ahmed Body

Fatih Kalay_20507

Hello all,

As I mentioned on the topic, I perform CFD analyzes on Ahmed Body,

In order to do that, I focused on 2 reference curves

In my excel file, you can understand my logic. The Excel file contains all the data which defines all details of the specific analyse except mesh type (Mesh type defines my element size-meshing method etc.)

I also added examples of my mesh structures below.

My main focus is to understand how should I use HyperWorks. When can I understand that, my analyzes give results close to real, how can I validate my results.

Thanks in advance

 

 

mesh type 10 (Surface mesh is 10 mm, Ref zone 1: 10 mm, Ref2: 25, Ref3: 50, Ref4: 100, Tunnel element size: 167 etc.)  Total of 4.5 million elements

mesh type 12 (Surface mesh is 5 mm, Ref zone 1: 5 mm, Ref2: 12.5, Ref3: 37, Ref4: 100, Tunnel element size: 167 etc.)   Total of 29 million elements

 

 

acupro

You may need to do additional mesh refinement studies - in the boundary layer, wake region, etc.  You may also need to try transient (along with a time increment study) versus steady-state to see how that affects the results.  And with transient, you may need to try standard SST along with DES based on SST - or maybe even S-A / DES based on S-A.  Not sure there's an exhaustive study on Ahmed body with AcuSolve.  What has been done in other publications on CFD for Ahmed body?

acupro

It will probably be necessary that the simulation match the experiment as closely as possible - supports/struts for the body, any 'larger' sensors or devices added for experiment data capturing, tunnel blockage effects, etc.  Likely one cannot capture everything exactly from the experiment, but at least the major details.

Fatih Kalay_20507

acupro_21778 said:

You may need to do additional mesh refinement studies - in the boundary layer, wake region, etc.  You may also need to try transient (along with a time increment study) versus steady-state to see how that affects the results.  And with transient, you may need to try standard SST along with DES based on SST - or maybe even S-A / DES based on S-A.  Not sure there's an exhaustive study on Ahmed body with AcuSolve.  What has been done in other publications on CFD for Ahmed body?

I added my mesh structures into main post. What do you think about the mesh refinement?

acupro

Fatih Kalay_20507 said:

I added my mesh structures into main post. What do you think about the mesh refinement?

It appears the images have not loaded correctly.

I believe this requires transient flow to match the experimental results.  One needs to capture the boundary layer and near-body region quite well - again, a sensitivity study.  One also needs good mesh resolution in the wake region behind the body.  One should probably do these studies at the highest Reynolds number of interest, since the results will be sensitive to the flow rate.