[Benoit85]

Hello everybody. I'm new to CFD. Please find screenshots in attachment to clarify my case.

I'm trying to run a simple axisymmetric flow case: 2 mass flow inlets (air), 1 pressure outlet, 1 axis, and some walls. Fluid: all air. Flow: inviscid. Solver: density based. In the future, I want to add and change some elements of the case, but for now I wanted to start simple. Apparrently I'm already doing something wrong.

The problem: at first, the solution seems to converge, but it happens very slowly (the total mass outflow approaches the sum of both mass inflows), but in the end it diverges.

When I choose pressure inlets instead of the mass flow inlets, the solution converges very quickly without any problem. Why is it so different with mass flow inlets? I need mass flow inlet BC because of the nature of the problem I want to simulate later on.

What am I doing wrong? Thanks!

[arnie333]

Hi Benoit85,

Are you sure that the inlet BCs you have modeled are realistic/representative of actual equipment/system ?
It looks to me like you will get significant reverse flow at the inlets and a very unstable system/analysis.

[Benoit85]

Originally Posted by arnie333

Hi Benoit85,

Are you sure that the inlet BCs you have modeled are realistic/representative of actual equipment/system ?
It looks to me like you will get significant reverse flow at the inlets and a very unstable system/analysis.

Thanks for your reply. For now, I'm using only air, but the system will be N2O inlet (the mass flow is known through experiments) and burning fuel grain (modeled as vaporized C2H4) of which the burning rate (and thus mass flow) is also known via experiments. That's why I want to impose a mass flow rate.

I just ran another simulation with Standard k-eps turbulence model and you can see the results in attachment. It converges way too slow, and actually, it still isn't really converged (see corner at outlet).