[abgoel363]

I am using fluent for the simulation of the reaction of supercritical water and metal salt solution in which metal oxide nanoparticles are forming.

How to decide the mesh size for the reactor of high pressure and high temperature (673K and 30 MPa)? should it be smaller than the particle size?

Thanks in advance for the help.

[vinerm]

The only good approach to model nano-particles is to modify the properties of the fluid itself and do a single-phase simulation. People have tried using Lagrangian particle tracking but it is useless to do it that way since the particles are too small to affect the flow of the continuous fluid. It is only its thermal properties that are slightly affected by the nano-particles.

[abgoel363]

Thank you very much for your reply.

Specifically, I am doing the simulation for continuous hydrothermal synthesis of metal oxide nanoparticles for ceria nanoparticles. (P=30 MPa, T= 673k).

I have already done the flow and thermal analysis by using realizable K-Epsilon model and reaction kinetics analysis by using volumetric species transport equation for the reaction in single phase.
Overall Reaction: 2Ce(NO3)3 + 4H2O --> 2CeO2(ceria)(aq.) + 6HNO3 + H2
CeO2 (aq.) --> CeO2 (solid)

My simulation consist of UDF for water properties at high pressure and high temperature and also for the rate of reaction as this reaction follows born type equation rather than Arrhenius rate.

Now I am interested to use population balance equation to get the particle size distribution of the ceria nanoparticle generated in reaction but this case is diverging whenever I use multiphase.

the reference for our work is

N. K. V. Nadimpalli, R. Bandyopadhyaya, and V. Runkana, A coupled CFD-PBM and thermodynamic analysis of continuous supercritical hydrothermal synthesis of nanoparticles, The Journal of Supercritical Fluids, vol. 136 (2018) 164–179.

-they have done the particle simulation for 2-dimensional reactor, but in our case, we have 3-dimensional reactor


Please give a comment on this how I can proceed.
Thank you very much in advance.

[vinerm]

This is not flow of nano-particles, rather its synthesis. You can do it using either PBM or Lagrangian but in both cases you have to write UDFs for nucleation, aggregation, coagulation, etc. Born type reactions rates are rather unstable but I doubt you have divergence because of that. PBM model requires proper setup and initialization. I'd suggest you to follow the guidelines for PBM setup and running as well as look for any tips the author might have given in the paper, if any. Or at least, follow their settings. 2D and 3D don't make a difference; it's just about an additional equation and geometric complexity. The physics remains the same.