In the example file I have built a basic room where the room is a static object and in the room are a number of circles that I've set up to emit particles from. The idea is that the circles are going to represent openings in a pipe where water is flushed out. I'm running into a number of problems getting as expected.

1. I've reduced the gravity quite a bit and was expecting the particles to interact with the room geometry. However they just pass through it.
   
2. I've set up the POP Source's initial velocity to 1 in the z axis. For one side it works fine. But for the other side it's spraying in the opposite direction.
   

So, how can I get the particles to collide with the room geometry and how can I reverse the direction of the particles on the one side? Thank you!

Construction of collision geometry is very important in obtaining a successful simulation with Pyro, Particles and Fluid. So just clicking the make Deforming or Static collision button is not enough. You have to go in and inspect the result of the tool and adjust the resolution to make sure there are enough voxels to represent your collision geometry accurately.

A collision mesh must be “air tight” or it will not work. This means no open faces may be present. Your box is basically one open face on the top.

I put together a revised collision object using two boxes and a boolean subtract. This way the resulting mesh is closed and there is some padding around the walls in case a particle leaks through, which can happen.

Thank you for the help. A couple of questions.

Do you know how I can get the orientation of the mirrored emitter surfaces reversed so both sets of emitters are spraying toward the center of the room?

And, I didn't know the collider needed to be air tight but looking at your solution I'm confused about that. Isn't the result of the boolean operation being used as the collision source and if so, the top is still open meaning it's not air tight.

For reversing, I would just duplicate what you have, and then invert the copies normal or velocity. Or maybe just rotate it so it is facing the other direction.

The resulting boolean I presented is air-tight, even with the opening on the top. Review the mesh. No wall is 1 face thin. I have opened up the top but there thick walls all around to collide with. This is a bit different compared to just removing the top of a box, for instance.