Zagreb, Vienna, OpenFoam vs ANSYS

I’ve had a busy and fun couple of weeks. After weeks of wrestling with OpenFOAM I finally got some results comparable to experiment and the commercial CFD software ANSYS. As usual, I’ve been traveling on weekends whenever possible. First the traveling. I was fortunate enough to have my friends Josh and Joe over from the UK, together with Mircea, we went to Zagreb for the weekend to do some sightseeing. Below is a picture of me with the statue of my favourite scientist and personal hero, Nikola Tesla.

 

2015-08-01 16.00.16

We were all very tired from the night before (we had been out to some clubs in Ljubljana) so before we went to the see the Zagreb nigthlife we decided to have a couple of hours sleep at the hostel. Five hours later we awoke and went to some pubs…

Here is a photo of the four of us in a Zagrab club called Peppermint (which I thoroughly recommend). Mircea seems to be trying out his facial expression for when he’s on the cover of Vogue magazine in a few years time.

 

2015-08-02 02.51.37

Josh and Joe headed back to the UK the day after. The next weekend Mircea and I went to Vienna. We were lucky enough to see the natural history museum and the palace. Behind the palace, there is a huge garden with a viewpoint on top of a hill. We climbed this hill and manage to take this stunning photo of Vienna.

 

2015-08-15 18.04.32

 

Here are some more photo’s of the city and the museum (we weren’t allowed to take photo’s of the palace).

2015-08-15 15.11.45 2015-08-15 15.23.26

2015-08-15 22.28.59

 

The most exciting thing about this trip was the dinosaur in the museum. Fortunately for us, it was extremely well behaved and waited patiently whilst I recorded the following video:

 

 

On the following Monday at work, we had the pleasure of seeing Prelog, the university cluster.

 

2015-08-14 15.49.37

Now on to some simulation results.

 

If you have read my PRACE bio, you may be aware that I am running simulations of wind flowing through/over a barrier in a wind tunnel. This was the subject of my mentor Marijo’s thesis in which he simulated the configuration using the commercial CFD software ANSYS as well as carrying out an experiment. ANSYS is expensive software (~25,000 euros for a 4 year license) whilst OpenFOAM is open source and hence free for anybody to use. Simulations were carried out for three different barrier configurations BC1, BC2, and BC3 with angles to the horizontal of 90, 60, and 30 degrees respectively. Below are the resulting graphs which show comparisons between the experimental data, the results from ANSYS, and the results from OpenFOAM.

 

forces

As you can see the OpenFOAM lift forces seem to be in better agreement with the experimental data than the ANSYS lift force. A nice result, especially when it could lead to a saving of 25,000 euros… Of course this could be a statistical anomaly, more testing must be done. Unfortunately the same cannot be said for the drag forces, ANSYS seems to provide better predictions. Notice the BC3 result for the drag force has a significant error. This may be due to the low quality of the mesh file used for this simulation, as OpenFOAM is known to be much more temperamental than ANSYS when it comes to mesh quality.

 

The software ParaView was used to visualise the results (again, open source software). Below is a slice of the 3D geometry, coloured according the wind velocity profile through the tunnel.

 

slice

Ignore the white thing in the middle it’s just an orientation axis. The point at which the velocity is at it’s lowest is the location of the barrier. Intuitively this makes sense as we expect wind to slow down when there is an obstacle blocking its path. Below is a plot of the streamlines (curves that are instantaneously tangent to the velocity vector of the wind flow). These are coloured by velocity.

 

streamlines

Again, the “colder looking” colours correspond to the barrier location.

Please follow and like us:
Posted in Blogs 2015 Tagged with: , , , ,

Leave a Reply

Your email address will not be published. Required fields are marked *

*