Turbulent steady flow around a circular cylinder
 

Hello Fluent Users, I have some problems with computation of turbulent incompressible steady air flow around circular cylinder (2D and 3D) using k-epsilon model at Reynolds number from 1E+2 to 1E+5 (subcritical). My problem is in a pressure distribution on cylinder surface. In the front side of the cylinder everything is OK according to literature datas (and my own experiments). But problem is from about angle = 40 degrees to 180 (most of all). This pressure distribution on cylinder surface is typical for SUPERCRITICAL FLOWS but Re number of the flow indicates that it is subcritical! I read in literature that Re subcritical is below 1.86E+5 and supercritical is under 6.7E+5. In my experiments the pipe has diameter of 150mm and cylinder has diameter of 12mm (this is an impact flowmeter). At v=30m/s I obtained in the front of cylinder (in stagnation point) p stat = 616Pa (596Pa in Fluent) but at the back side of cylinder in experiment it was p stat = -872Pa (-155Pa in Fluent!OVER 400% ERROR!) My question is: How to force Fluent to make a proper calculations? My mesh on cylinder has intervals equals 0.27mm, I used inflow as velocity-inlet B.C., walls of pipeline and of cylinder as wall B.C. and outflow as outflow B.C., and convergence in case of 3D simulations on level of 0.0001, computation domain has about 2mln elements. Now, I'm trying to get correct answer in case of 2-D infinite turbulent steady incompressible flow around circular cylinder and when it'll be successful I'll try to solve my problem in 3-D case. Can anybody help me with my odd problem? Thanks anybody in advance for help.