[sharonyue]

Dear OpenFOAMers,

After OpenFOAM was released in 2004 as a open-source free CFD code, it has been 13 years. By 13 years' effort, OpenFOAM has been the largest open-source CFD code all over the world AFAIK. Thanks to the pioneers (Henry Weller, Hrvoje Jasak, etc.), more and more models and algorithms were included in OpenFOAM. More and more organizations are developing and releasing their own code based on OpenFOAM.

However, as the largest open-source CFD code, I would not say OpenFOAM is the easiest code to use. Obviously OpenFOAM is that kind of code you can hardly master, or even hard to learn, if you know only a little bit about CFD theory. Easy to Use and Flexible Tailor-made sounds to be opposite in nature. This is further justified by the commercial software and open-source CFD code. The commercial codes try the best to make CFD as simple as possible (with GUI, automatic model selection, comprehensive documents, etc.) with expensive license fee. But you never expect OpenFOAM, or other open-source CFD code, to have a hand-by-hand tutorials.

Based on the contradiction discussed above, personally speaking, I see a distinction, or a barrier. Most of the industrial engineers prefer to use commercial codes, in which they can easily launch any CFD simulations and the algorithm is quite robust. I am not a subjective person, but I think every one here would agree that ANSYS Fluent, for example, is more robust than OpenFOAM (It should be stressed here that I am not advocater for ANSYS Fluent, but rather to admit a truth). Instead, in academia, most of the researchers prefer to use OpenFOAM, in which they can verify their algorithms and mathematical models.

The question arises. After 13 years' OpenFOAM developing, do you see the possibility OpenFOAM (or other free CFD code) will seize the whole academic and industrial market?


BTW, as a Chinese, I would say there is still a long way for a open-source CFD code to win both in industry and academia, at least in China.


Best,

[elvis]

Hello

i think OF has a big impact in industrie for example Volkswagen and it brands like Audi Seat etc. has adopted this software in many areas (hvac, acoustics etc.)

http://www.move-csc.de/paper/sae2009010333.pdf
http://www.move-csc.de/paper/aeroaco...erlin_2008.pdf

and you find lots of phd thesis (Hrv Jasak has not updated his collection for a long time )

http://powerlab.fsb.hr/ped/kturbo/OpenFOAM/docs/

PS. next years the OF workshop (OFW13) https://twitter.com/of_ws/status/890125194622357508 will be hosted in Shanghai do you know which university will be hosting

[student666]

Hi,

I think that big company with a strong R&D department could carry out the Openfoam project if it wants.

In small/mid sized companies (what I'm speaking about next) only few (maybe one) persons can carry out a CFD project and the limits you face when you deal with OpenFOAM, is the large amount of time you spend on mesh generation and case set up; and further, you don't have budget for some commercial tools.

Mesh generation, for example is not easy with free open-source tools, and when you start to think to buy a commercial mesher, you also think that a commercial package for CFD could satisfy your needs, but if you're using openfoam is also because (maybe) your company doesn't want to invest money and it wants "to evaluate"...

Case setup is sometimes very painful as for some dictionary that you can't understand, or as for your lack on knowledge on c++ programming that you don't have time to spend on learning c++ language; so commercial software with their customer care center helps you when you're stuck and tells you:"click here"..."open it there"...

But from a computational point of view; I'm sure that openfoam can challenge the best CFD package...at the least for the cases I deal with...

So "robust" as you say, by my point of view, it means only that some tasks are faster with a commercial software and for a "time to market", that it is always expired in the designing process: a commercial software is indeed "more robust and fast"...

Regards.

[sharonyue]

Originally Posted by elvis

Hello
do you know which university will be hosting

Shanghai Jiao Tong University. They did lots of work on OpenFOAM, especially naval applications (interface tracking, overset grid, etc.)
I am also planing a OpenFOAM conference or workshop in 2019. Lets wait and see.


Thank you all for the information. I hope OpenFOAM will overthrow the CFD status (leaded by commercial codes) someday. Although some large company is funding OpenFOAM for development. But for the smaller one, hard to say

The emerging clouding computing is limited by the network connection as well, even all the vendors embrace it.

Best,

[alberto]

What's the point of discussing "winning in both academia and industry"? It's not a war or a competition even. There are different needs and objectives, and trying to compete, imho, is a waste of time. Filling needs and solving problems is what actually counts.

On one aspect, I would, however, spend a few words.

Originally Posted by sharonyue

Most of the industrial engineers prefer to use commercial codes, in which they can easily launch any CFD simulations and the algorithm is quite robust. I am not a subjective person, but I think every one here would agree that ANSYS Fluent, for example, is more robust than OpenFOAM (It should be stressed here that I am not advocater for ANSYS Fluent, but rather to admit a truth). Instead, in academia, most of the researchers prefer to use OpenFOAM, in which they can verify their algorithms and mathematical models.

I think there are more reason than "robustness", "ease of use" and academics wanting to verify the algorithm in the choice between a commercial code and an open-source solution like OpenFOAM.

First, talking about industry is too generic. What industry, what size, what level of competency, and what are the motivations of the specific company to do CFD? The panorama of companies internally using CFD varies significantly, with one extreme of the range made of companies which value CFD as an essential asset of their design/development tool-chain, and others considering a little more than a generator of colorful pictures for their brochures (and insist on the myth of useful simulations that run "overnight"). Obviously all the companies tending to the second extreme of the range have no interest in using an open-source solution for CFD, and I would argue they hardly have the in-house competency to do so. For the rest, it becomes a question of needs, resource availability, time available to be invested. I would say that the more a company finds CFD useful, the higher the incentive to use an open-source solution is.

When it comes to academia, the question is slightly different.
First, we need to distinguish between academics who use CFD in their research as tool, exactly as a company. For this group, the same considerations made for industry apply. For researchers who develop CFD algorithms, models, numerical schemes, it makes no sense to use a commercial package to perform the development because of the lack of control and the impossibility of accessing to the implementation of the underlying numerical framework. In this view, and with the increasing pressure, especially for publicly funded research, to have open access, reproducible, verifiable research and publications, using an open-source solution will likely become an implicit requirement.

Best,

[hjasak]

Just moved and cleaned things. Please check out the new 8th Floor CFD web site:

www.fsb.hr/cfd

https://foam-extend.fsb.hr/

Hrv