Help with modelling turbulent flow through tube bundle
 

Hey guys,
I am just having trouble re-creating an experiment conducted a fair few years ago, so we can validate it. I am trying to minimize computation and not sure how to go about this.

The experiment consists of an array of staggered tubes with a flow going through them at 1.06 m/s. The data required is taken off the 5th set of tubes as flow is deemed periodic here. I have been suggested to make a singular elementary cell with periodic boundary conditions (1 inlet and 3 opening boundary conditions) on all four sides. However when I have tried this with the inlet condition the flow is not periodic as it would be in the 5th set of tubes....I see that maybe I will have to make a geometry of 3 x 5 cells and take the data off the centre cell on the 5th row? or will it be possible to simplify this?

The link below is the details / picture of the previous experiment. Any help would be appreciated as I need to get this done ASAP!

http://imgur.com/aiLPcS9

I simulated this case in COMSOL, first, you need to specify two boundaries as periodic, left and right, upper and lower boundary use symmetry. Also note that in the image at the bottom, the boundary is not periodic, left: solid is at bottom, right: solid at top, so you need to use a domain double in size in horizontal direction.