I'd do all of the post processing in Fluent. DPM and CFD Post tend not to work so well; if you do want to try the particle tracks can be saved & exported via the File menu.
Looking at the particle tracks I suspect your problem is mesh related. Some of those tracks look more like particles bouncing around the top section than following the rotating flow: look at the apparent kinks in the trajectory plot.
You also now want to look at velocity vectors on planes, try with vectors "in plane" and see what the flow is doing.
In terms of DPM boundaries. What do you want to happen? Then look at the options and suggest what you think you should use.
You are navigating away from the AIS Discovery experience