hi; i am trying to code 3d rans code for solving turbulent flow around a cube as my term project. i have searched the web and i found no source about how to code or solving procedure. so i have some questions. 1)in some papers they write that they use backward differencing however i remember we are solving a elliptic equation so we have to make it central differencing for higher accuracy(if i am not wrong). is this for the case in the paper or this is a general method used in turbulence solving? 2)in the equations which variables will be treated as knowns.(i mean which values should be taken from previous solving step). or will all the terms be unknowns. i know that taking some variables from previous make quicker the code. 3)do i have take care of surface of the cube. does it needs some equations to solve for boundary. 4)for initial condition what value should i give

thanks for all

i do not know where you searched. If you searched the forum carefully you will see time to time people keep asking about sources for navier stokes. OpenFoam is one always get mentioned. just one or two days ago someone suggested inavier.

Yes,Inavier seems easier to control than OpenFoam. OpenFoam packs verything together and is a nice package, but too complicated to a non-CFD guy. its complexity aslo puts down many researchers who only want to treat it as a research platform and exploit their own physical model on base of it. Therefore, these researchers perfer to using a commerical package to go on with their work, not based on the open-source OpenFoam. Hope Openfoam's team can notic this and will make a grogress on this problem in the future. Inavier is good to a beginner and also capable of bring you to the real world of CFD.