[TallesC]

Hello, can anyone tell me if the Ansys CFX software uses the Particle-Source-In-Cell (PSI-C) schema for the Lagrangian particle tracking, I am simulating spray (liquid particle dispersion)?

Note: I did not find it in the manual!

[ghorrocks]

I do not know what the PSI-C schema is. The documentation describes the approach it uses - simply the force acting on the particles represented by a lagrangian particles are applied to the particle to gets its motion and onto the fluid at the control volume where the lagrangian particle is located. This approach could be described as a source term in the cell at the particle location, but whether it is the PSI-C method I have not idea. You are going to have to check the approach used and see if it is the same.

[TallesC]

hello glen, thanks for the answer, what I wanted to know if Figure 6.41 contained in the link below that uses psi-C describes the operation of cfx-solver to solve the problem:

https://books.google.com.br/books?id...figure&f=false

can you help?

[ghorrocks]

Google does not allow me to view that link. But I do not have time to do a detailed check of whether that model is the same as the CFX model. Can I suggest that the best person to check that is yourself.

If there is something you don't understand or unclear then feel free to post a question, but I am certainly not going to do all the work for you.

[TallesC]

my doubt was precisely in the flowchart, is that it states that the system is solved without drops only the gas phase, then is calculated: size, trajectory, temperature of the drops, soon after solves the equations with the gas phase (continues) with the terms If the result converges it stops, if it does not return to the calculation of the variables of the drops (size, temperature,..), I wanted to know if that is the same as the cfx solver calculates, especially if the first step is to solve the system without drops?

[ghorrocks]

If you use the default uncoupled volume fraction solver (this is volume fraction-NS coupling) then the first iteration solves based on whatever the volume fraction initial condition is. If that is drop VF=0 everywhere then it is effectively as you describe.

if you use the optional coupled VF solver then the first iteration solves a big coupled matrix of the momentum, pressure/mass and volume fraction equations all together. This is not the same as the segregated approach you described.

[TallesC]

Ok!, I inject particles in the domain to simulate spray leak of flammable material in the environment (air only) in steady state, in the software in global initialization is with automatic initialization for mass fraction of the components, so initially the software(CFX) must be with VF = 0, since no particulate was injected into the domain, right?

note: in fluid specific model ->air is constant and inflammable component is transport equation.

[ghorrocks]

On second thought, I am not totally sure the coupled VF solver works on lagrangian particle tracking. It is normally used in Eularian multiphase models and I am not sure it is implemented for Langrangian multiphase models. If you are interested in this I will leave it up to you to check this.