My understanding is, the problem you (@Prashank) mentioned is not unusual. The number of incomplete particles will depend on how many steps with what step length factor/length scale Fluent is tracking your particles. If your particle is still trapped inside your domain (e.g. due to vortex formation) after the longest possible tracked length is reached, it will be declared incomplete. As your iteration progresses, there must be some fluctuations in the solution of the flowfield which will lead to fluctuations in the residence time of particles. Therefore, your number of incomplete particles will naturally vary as well, I suppose. 

I would suggest that, you try increasing tracking parameters (e.g. Max Number of Steps) under your DPM settings, in order to let fluent track your incomplete particles to the end. This way you can account for the definite fate for all your injected particles.

I hope that helps.