VOF-to-DPM solver questions

  • Last Post 29 April 2019
  • Topic Is Solved
bpal posted this 23 April 2019

To whom it may concern,


I've got some questions on the VOF-to-DPM solver transition. The 19.2 theory guide states the following:

"In order to avoid spurious momentum sources, a volume of gas with the same volume as the liquid lump is created in the VOF simulation to maintain the volume conservation. The mass source that is equivalent to the mass of the gas volume will affect the overall mass balance."

1 - Is there a reason that the paragraph discusses volume conservation instead of mass conservation?

2 - Does the last statement mean that mass conservation is not enforced at the time of the transition? 

3 - How is the number of particles per parcel calculated? 

Thanks for your time.




Order By: Standard | Newest | Votes
abenhadj posted this 24 April 2019

Lagrangian particle parcels do not have any volume displacement effect on any VOF phase. Therefore, in order to maintain volume conservation in the VOF solution, when liquid from the VOF solution is converted into one or more Lagrangian particle parcel(s), an equal volume of the selected gas phase is placed into the VOF solution as a replacement for the converted liquid. That is why volume conservation for Eulerian Phase and Mass Conservation for the discrete particles.


Regarding the number of particles per parcel: that is controlled by the overfill factor or Split any DPM Parcel that Exceeds the Cell Volume by Factor. Please refer to the documentation User's Guide here.

Best regards,


  • Liked by
  • bpal
bpal posted this 29 April 2019

Thank you for your timely reply. A few follow-up questions:

- The mass of the DPM phase is not taken into account in the continuum calculation. So if one were to check for mass balance in the domain, dpm mass flow and mixture mass flow would have to be measured separately at the outlets and summed up. Is this correct?

- After the VOF-DPM conversion happens, there will be a significant mass imbalance in the cells where it occured due to dense liquid being replaced with much less dense gas.  Is this correct?

- Regarding the "Exceeds the Cell Volume by Factor" parameter: My understanding is that this will determine how many parcels the ligament gets split into based on a volume limit w.r.t. the original cell size. However, no mention is made of how many droplets each parcel will contain and how this is determined. In other words, what determines the sizes of the droplets? ( Since parcels represent a group of droplets that have the same size and velocity) 

Thanks again! 



rwoolhou posted this 29 April 2019

Yes, but as it's transient you also need to check what's in the domain. 

Yes, sort of. Fluent should deal with that and adds (non droplet) phase into the cell to conserve volume. 

Please review in the 2019R1 documentation: we've been working on the model so it's a good idea to update anyway. If it's not covered we can't elaborate in any detail as this is a public forum. 

abenhadj posted this 29 April 2019

1/Yes you can do that

2/See my last post

3/If the lump fits into a coarsened cell times that factor then the lump is converted into 1 parcel with one particle in it. If it is larger then the lump is splitted to a number of parcels related to the number of cells covering the original lump. Mass conservation implies now that a parcel has less than one particle in it

Best regards,


  • Liked by
  • bpal