gravity on or off?
 

I'm dealing with a big geometry (internal flow), in which the vertical distance from bottom to the top is around 50m. It's single phase and surrounded by walls.
My only concern is that since I'm doing a particle tracking on it, I need to apply the gravitational force on the particles for that purpose.

Should I apply Gravity during the fluid flow iterations or not?

The convergence with gravity-off is nice! with gravity-on the convergence is not easily achievable (still I can get something), but the streamlines look weird and there are many incomplete streamlines!

PS: I'm using realizable k-e with scalable wall function.

what do you solve exactly in FLUENT ? gravitional affect may be important for your model or not. That is relation to your expectation. No gravitonal effect affect Navier Stokes equation. That is why your model is easily convergenced (negligible gravity)

It seems you're modelling multiphase flow and your particles are moved by buoyancy forces. Am I right? If so you must activate gravity. Take in mind that buoyant flows are not easy to converge especially using steady-state solver, often they should be modeled as transient with appropriate time stepping. Also you should carefully specify reference density/temperature and BC.

It's a single-phase water flow through a huge internal system. and I'm trying to see the rate of sedimentation of solid particles in that. I was wondering that maybe I can solve the fluid first (with no gravity), and then activate the gravity and try the DPM-particle tracking.

That's the part I'm debating. It's not multi-phase. I mean, it is, but my solution approach is a single-phase fluid simulation and then tracking the particle's trajectory in the fluid.