[mathias]

Dear openlb12,rnrnIn this exmple, the distances are computed to the wall which means that the preprocessing tanks longer than in the other examples. You can also run this example in parallel. Since this part of the preperation work is parallelized, it will be much faster.rnWe will come with a new release soon. In it the whole preprocessing will be much faster since it it completly parallelized rnrnBest regardsrnMathias

[mathias]

Hi openlb12,rnrncheck out the following 3 papers:rnrn1/ 2009, Heuveline, V. & Krause, M.J. & Latt, J.: Towards a Hybrid Parallelization of Lattice Boltzmann Methods, article in Computers and Mathematics with Appliactions, 58, 1071-1080rn2/ 2010, Krause, M.J. & Gengenbach, T. and Heuveline, V.: Hybrid Parallel Simulations of Fluid Flows in Complex Geometries: Application to the Human Lungs, proceedings of Euro-Par 2010 – Parallel Processing Workshops: HeteroPar, HPPC, HiBB, CoreGrid, UCHPC, HPCF, PROPER, CCPI, VHPC, Ischia, Italy, August 31–September 3, 2010, Revised Selected Papers Springer-Verlag Berlin Heidelberg, Vol. 6586, 209-216rn3/ 2010, Krause, M.J.: Fluid Flow Simulation and Optimisation with Lattice Boltzmann Methods on High Performance Computers: Application to the Human Respiratory System, Dissertation, Karlsruhe Institute of Technology (KIT), http://digbib.ubka.uni-karlsruhe.de/volltexte/1000019768rnrnSince OpenLB is a open source project it depends free contributions of a few nice guys. You are welcome to help the project by contributing text in the framework of GPL2, which I am happy to insert in the next user guide.rnrnBest regardsrnMathias

[mathias]

Hi Hanieh,rnrnI could see your problem as well in our development code. It might be that there are gaps in the stl mesh. We are going to fix it for the next release. However, a big mesh size in general is not a problem.rnrnIf you see that error, a bounce back node is set instead so that the simulation will work.rnrnBest regardsrnMathias

[mathias]

Hi,rnrnwhat did you do for the corner nodes? I would set a combination of BB and Zou/He. Further, have a look in Sukops book there is some pseudo code or/and the OpenLB inplementation.rnrnBest regardsrnMathias

[mathias]

Hi Shayan,rnrnI think LBM is able to solve it but you may need a new collision operator. Which macroskopic equation is governing the problem? rnrnBest regardsrnMathias

[mathias]

Hi,rnrnHave a look at our example code “”thermal2d”” and “”thermal3d””.rnrnBest regardsrnMathias

[mathias]

See the quotes.rnMathias

[mathias]

Quote:

Quote from jieeshtker on January 7, 2014, 10:13rnHi folks,rnrnI am trying to set up a LB-simulation to simulate a pipe flow and I am experiencing several problems. I hope, you can help me in an understandable (not heavy formula using) way.rnrnFirst and most important question set:rnWhat should I do with any walls? I understand there are a lot of different approaches (bouncing back, inamuro, free slip, no slip, complex things, …) – but I fail programming any of these. As far as I understand, for a simple pipe flow I need a bouncing back condition.rnrnIs this still right? rn>>Yes, thats right. Have a look at our example code.rnrnIf so, all I have to do is to take the distribution functions pointing towards the wall (e.g. pointing North, Northeast, Northwest) and put them into the wall (South, Southeast, Southwest). [This ought to be free slip or what it is called.] Now, if I switch east and west I should have no-slip.rnrnIs this still right? rn>>Yes, thats right again. You may have a look in the standard literature, e.g. Sukop’s book “”Lattice Boltzmann Modelling””.rnrnAnd then, what do I do if an obstacle is to be simulated? There could be corners of which I have absolutely no clue what to do with them.rn>> You can also use bounce back or an interpolation sheme like those of Bouzidi. Have also a look at our 3d example “”cylinder3d”” where we use the later one.rnrnrnTo the not critical questions:rnrnIs there any way to set boundary conditions by defining macroscopic parameters (e.g. von Neumann, Dirichlet)? rn>> Yes, there are may scheme. You may have a look in the standard literature, e.g. Sukop’s book “”Lattice Boltzmann Modelling””. rnrnrnIs there a way, to realize different viscosities? I think, viscosity is only used in the collision, but can I just use different (and chaning) viscosities in a single cell (and keep conservation of mass)?rn>> You can set it different for any cell for the collision step. However, it may not make sence physically.rnrnMany thanks for your time and answers.

[mathias]

Hi peili_enhy-ro,rnrnYou can a/ install your own boost and replace the include in Makefile.inc from local to your new boost paths or b/ do as recommented change in “”./../src/external/boost/bits/predefs.h:20:3″” “”<bits/predefs.h>”” to “”<features.h>””.rnrnAfter that you must recompile the whole libarary by “”make cleanbuild””, “”make clean”” and “”make””.rnrnBest regardsrnMathias

[mathias]

Dear lixuhui365,rnrnSorry, we dont have. But have a look at our aorta example where the boundary conditions are set by analytical functions. You can define your own functions and initialize the geometry as well as the boundary conditions by that easily.rnrnBest regardsrnMathias

[mathias]

Dear jomidi,rnrnIf you install boost on cygwin and change the local path “” src/external/boost “” in Makefile.inc to where you installed boost on your machine, it should work.rnrnWe tested it before for Mac and cygwin on Windows rnrnBest regardsrnMathiasrn

[mathias]

Hi,rnrnIn general yes. Have a look at our examples, where is the ratio realized by different body forces. We have extented the model in a way where you need no body force and we will release it very soon.rnrnMathias

[mathias]

Hi,rnrnYou can use WinRAR on a windows system. I will send you an email with zip code.rnrnMathias

[mathias]

Hi,rnrn Please, indicate exactly what you tried and what did not work.rnrnMathias

[mathias]

Hi,rnrnYou need to unzip it before using. You might need to install a program to do that.rnrnBest regardsrnMathias

[Author]

Posts