periodic boundaries - flow through a net

PK

I want to simulate the flow through a grid (net). But I am more interested by the flow through a "basic cell" of this net. That's why I have a cruciform as a basic cell.

The turb model is keps / solver pressure based / node cell ;

I have define a Control volume around this cruciform. The boundaries are defined as Periodic (opposite by pairs).When I run my simulation I got the flow back this cruciform : No Prob;

The problem is that when I run the case with symmetrical boundaries (instead of periodic /net) , and so I simulate in this later case a cruciform with infinite arms : I get exactly the same results : flow, Cd….;

It seems that in the first case (periodic boundaries ) the interactions with neighbored cells are NOT took in account .

How to setup correctly the periodic boundaries conditions. I can't specify the Mass flow and I don't know neither the pressure gradient. (But I thought that for Periodic Boundaries, that the flow which is going out of a face is automatically "injected" in the opposite face of the CV, but it is not the case)

Thanks for any kind of help