CFD Online URL
[Sponsors]
Home > Forums

Small Rants, that look more like something that came out of dreams...

Register Blogs Members List Search Today's Posts Mark Forums Read

Rate this Entry

Small Rants, that look more like something that came out of dreams...

Posted September 2, 2011 at 20:32 by wyldckat

Well... I'm somewhat lazy, so instead of creating a personal blog elsewhere, I'll rant a bit more here at cfd-online.com/Forums, right here in my user blog


________________________________________
Rant #1 - Technically, it seems that Google's name is explained here:
Nonetheless, today I noticed something interesting when I wrote somewhere [1]:
Quote:
the good people at google
Mmm... "good people" <-> "google" ... "goodple" -> "google"! This would lead me to believe that their moto of doing no evil, could actually have been subliminally put into (or taken from) their minds when coming up with the name and/or motto.

[1] - here if you must know


________________________________________
Rant #2 - After (re)viewing a lot of the comments at the thread OpenFOAM acquired by SGi and reading Alberto's call to sanity and hope (I'm not quoting ), my mind did some math around the acronyms FUD (Fear, Uncertainty and Doubt) and CFD... and came up with: Constant Fear and Doubt. And oddly enough, it makes some sense!
Those who have dealt with Computational Fluid Dymanics long enough, learn the easy or hard way that results can't be fully trusted unless you:
  1. Understand the physics behind the simulation;
  2. Validate them (somehow) with real experiments;
  3. Have a good mesh to begin with;
  4. Define proper boundary conditions;
  5. Have the residuals lead to convergence;
  6. Use the correct models for your case at hand, or at least understand the limitations of the chosen model.
Even when you have all of these, if any single point is changed... all hell could break loose if your assumptions are wrong! Just one bad thing... just:
  1. Not understanding how the numbers of Reynolds, Courant, Fourier, Biot... can relate and/or have significant meaning in your simulation case, could lead to disaster!
  2. No validation, not even even for a simpler related case, means that your results are as meaningful as the online game "World of Warcraft"... you might be right about your assumptions, you might know how to fight, but without real life experience, it's all virtual.
  3. One wrong cell, point, face or even cell structure (polyhedra vs tetrahedra), could mean that your results might either be garbage, or simply only have an error of 0.1%.
  4. Boundary conditions... math can't get any more real than this, when it comes to simulations!
    With luck, setting the wrong boundary conditions only means that it will simply take longer to converge. With bad luck, no convergence occurs; or even worse, converges but has no real physical meaning!
    Example of the latter: I've had experiences with the good old OpenFOAM 1.5, where bad boundary conditions lead to residual convergence, even when the flow was piercing a hole in the mesh - where no hole should exist - therefore making all of the injected air go out in a blaze of glory through the hole that does not exist!
  5. Relaxing variables (I can't remember the correct term )... although I'm all for a stress-free work environment... how on this Earth is relaxing a variable going to make virtual be more real? Wait... the word fuzzy comes to mind...
  6. LES or RAS... compressible or incompressible... Lagrangian or Euler... VOF or Multiphase... k-Epsilon or k-omega... viscous or non-viscous... CFD or flipping burgers at MacDonalds!?
Conclusion: no wonder that people were so uncertain about the future of OpenFOAM, when OpenCFD was bought by SGI! CFD is already deep in their subconsciousnesses.


________________________________________
Small rants, to be continued... here: Small Rants, part 2
Posted in Rantings
Views 1154 Comments 0 Edit Tags Email Blog Entry
« Prev     Main     Next »
Total Comments 0

Comments

 

All times are GMT -4. The time now is 15:59.