Particle self-gravity with radius module

[Ken Rice]

Hi,

Is anyone aware of any issues when including particle self-gravity while also using the particle_radius module?  I'm getting very different results when compared to simulations where I specify a stopping time, rather than using the particle size to determine the stopping time.

Thanks,

Ken

[Ken Rice]

Okay, I may found my issue, but it's raised a new one.  I'm using the particle_radius module and running simulations with a particle size distribution.  I'm then defining the dust-to-gas ratio (eps_dtog) which then sets the mass of the superparticles (mp_swarm) and the mass density of the superparticle (rhop_swarm).  Unless I'm mistaken, the latter is the mass density relative to the cell volume.  When setting it up this way, np_swarm (the number density of the superparticle) is zero.  However, the self-gravity routine needs the particle density which it seems to get from particle_map.f90 and uses np_swarm, which is zero, hence there is no particle self-gravity.  The obvious solution would seem to be to use rhop_swarm in particles_map.f90, but I'm wondering if I'm missing something obvious.  I don't want to go changing the code, if there is a better way to do this.

I will add, though, that the calculation for the backreaction of the particles on the gas seems to simply use rhop_swarm, so I did wonder if using np_swarm for the self-gravity would mean that it's not consistent with the back-reaction and that using rhop_swarm would be consistent with the calculation of the backreaction.  Also, I would like each superparticle to have the same mass, since that gives me the size distribution that I'd like to use.  Using np_swarm seems to imply that each superparticle has the same number of constituent particles, but not the same mass, so gives a different size distribution to one in which each superparticle has the same mass.  Again, maybe I'm missing something here.  

Hope the above makes sense, and I'd appreciate anyone else's thoughts on this. 

Thanks,

Ken