Initialization patch
 

Hello, I need to simulate a steady state release of hydrogen in a container with an open side opposite to the leak.
The geometry consists of a parallelepiped with a pressure outlet on one of the faces and wall boundary elsewhere.
Opposite to the pressure outlet there is a small cylinder with a mass flow inlet of hydrogen, so this is the only inlet of the domain.

When I first initialize the solution, I patch the domain in order to have only air inside the container, and I verify it with a contour.

As I start the simulation the air concentration drops to 0 and all I see in the domain is hydrogen.

Am I doing something wrong with the boundary conditions? I also need an air inlet?
Thank you