Ok - This is a general science philosophy topic - so, lets connect it to CFD.

I've seen that one thing is a "built-in" thing in your brains throughout the education mediums from elementary to graduate schools. Almost all of you (I suppose most of you are graduate students) are claiming that "... technology is being advanced - SO, we are doing somethings correct ..."

Well... This is the main, biggest problem in their minds of researchers. This is why Einstein-like true scientific minds are very few in millions of researchers, so-called researchers. You're students; meaning still-learners. So, such illogical things are still being built-in your minds by brainless professors who are a big majority of scholars.

What you do not understand is that ".. technology will advance whichever direction you decide to go... this is unavoidable result.."

Let me be more specific by relating it to the CFD as this is a specific topic here.

Lets say you're undecided about using a mesh size in a certain highly nonlinear flow like a turbulence. Whats your first step is to look at the Re number in different characteristic scales. If it Re. nr. exceeds a certain value, you classify the flow as a turbulent flow. Right? Yes, so. Then, you chose the mesh size, etc according to the flow classification; mesh size depending on how much nonlinear / turbulent your flow is. You discrete the differantial equations according to this selected! classification. And, you iterate to solve series of equations. UP TO here, CFD brings nothing new to those partial differential equations you started with - meaning that you are NOT using an independent method. You do somethings, solve equations transformed from infinitesimal spaces to discrete spaces and you obtain somethings you say they are applicable and hence technology is being advanced (by you.)... WELL... Now, go back to mesh size determining stage and stay there without moving to any direction. Now, I'll tell you to go to the another direction different than you went before when you were doing your advances. Chose another mesh size completely different than you did before. Then, you do the same things; solve equations, etc etc. At the end, again, you'll see some advancements in the technology by chosing another mesh size. So, what does this mean? It means... whichever direction you go, you'll see some advancements in technology. So, this means.. technology advancement in the life is an unavoidable thing and it doesn't depend on you dear researchers...

Now, let me make this example more complicated, but more cleaers... (using the same mesh size example.)

Study a little more about those nonlinear partial differential dynamical systems before transforming them into a discretized space; i.e. into computer (linear) language. Study about their infinitesimal point symmetries. and see unexpected sudden bifurcations you'll later realize. This means turbulent scales are as small scales as infinitesimals. Then, how come, how dare you can use coarser scales (in time and space)? But.. your already using the coarsers already. you are ignoring "point turbulences" which will cause some unpredictable flow shapes quickly. If you remember from physics that even the motion of a single point can go into chaotic motion. Am I clear about what I'm trying to tell? Let me summarize... Even though an infinitesimal point motion can go into a chaotic, and more chaotic / turbulent motion, how come can you use a finite small mesh size to study about the details of turbulence by CFD which is based on original infinitesimal continous mathematical equations????

YES, you can use.. and can get some useful results. BUT as I said you'll get some other useful results whichever direction you go - and you'll face some other kinds of advancements in technology. So, WHICH advancements is correct??? your current advancements or advancements you have never experienced??

Run your CFD codes backward and see what other advancements you can get.

???????

I am not sure to get your point !!

Steve

So Nomad, what alternative do you offer?

Navier-Stokes equations are a model. If I follow you, I should not use Navier-Stokes!

1) If you really want to discuss something, don't rant. I never met a "brainless professor" in my entire life. Not everybody is John von Neumann, but that doesn't mean that you're doing bad research.

2) Just solving the Navier Stokes equations numerically is not science. It does not advance anything. First when you start drawing conclusions about what you saw, science begins. Everybody knows that the Navier Stokes equations are only a model. And will interpret the results based on that.

3) The quuestion how to close your turbulent equations, such that the computations can be carried out in an acceptable time, but that the turbulent effects on the fine scales which you are not computing anymore make your computations meaningless is a hot topic in the turbulence comunity IMHO.

Should be use Boltzmann instead? mmm... Disregard the continuum hypothesis ?, depends . Is not about solving N-S, but getting some statistics more or less right. Doesn't it?, Transition? we are never going to get it right. Universal turbulence model, probably impossible. Acceptable turbulence model in very small scales, probably acceptable. Never accept results with less than 3 grids.

Dear Nomad,

I have to admit, I read your post twice, and I don't understand what is your point. I admit, I'm more idiot then genius, but still.

There is one think which comes to my mind when reading your posts:

1) you are so ingenious ; 2) you've been educated by aliens; 3) you're trying to get off your irritation; 4) you want to disscuss thinks but you do not know how, and you listening is extremely selective;

If you want to disscuss things you should not write things like "...in your brains..." but "... in our brains...". It sounds much less offensive, so more people might feel OK to join the disscussion.

It is always valuable to question the theories and models so widely used for ages, that some are not aware of the limits of these theories. Some people spent their lives working with the N-S and turbulence definition as it is known now, if you want to disscuss with them you should use nice language. All of us have to take care of our ego - it is vital thing, you cannot live without one ).

I'd be happy if you could formulate your questions more clearly, for us, dummy humble PhDs.

matej

he sounds like an old lecturer i used to have.

Nomad??!!! What do you want? If you are coming with an alternative intellegent discovery, why don't you discuss it instead of your offensive attitude towards the scholars and professors we respect, otherwise please stop this shit.

Bring back JohnC, far more involving than this troll.

Yep.. what happened to JohnC? I still have collection of his posts printed in the office...

-- Jarmo

Hey Nomad, you are using a misnomer for this thread, you should have said "CFD doesn't give right results"!!But then I presume you don't understand the concept CFD at all. Your example of choosing a coarse mesh size and "point turbulences" shows it.

If you are well read in the area of CFD then perhaps you would know that to capture these infinitesimal scale turbulences DNS (Direct Numerical Simulation) can be used, where you use infinitesimally small meshes. But it happens to be immensely costly based on the CPU time. You can spend years, to get the "right results" for a simple flow. So CFD has alternatives (RANS, LES, DES etc....) where you concentrate on the broader aspect of the flow rather than the smaller insignificant ones!!! Everyone knows that CFD is an approximation, and thats the reason why so many people are working on improving it. It is a better option than taking a pen and trying to obtain an analytical solution to these NONLINEAR equations!!!

Talk to one of these "brainless professors" and he would be able to explain this to you. Or you can spend a lifetime proving what others already know....