# What should be the boundary condition for this case ?

Member Posts: 1

Hello, I am trying to simulate this paper and I am stuck at this. I have tried piecewise polynomial function for specific heat and Ideal gas for density for air and for inlet condition I have used mass flow inlet and specified mass flow rate and temperature and for outlet pressure outlet. Still I am getting around 80K error. What should be the boundary conditions ? I have also tried applying gauge pressure with mass flow rate but solution diverges ? I have checked mesh quality.

Tagged:

• UKPosts: 8,855Forum Coordinator

I'd start by using the same material properties as the study.

• Posts: 28Member
edited December 2020

I think you have used correct boundary conditions. You need to find the error somewhere else.

Is your model 2D? The Swiss-roll recuperator is tubes with circular cross-sections, but if your model is a 2D and the pictures are showing your domain, note that the solver takes your model as a rectangular channel with theoretically infinite depth (perpendicular to the screen) so that gradients in the z-direction are zero.

• UKPosts: 8,855Forum Coordinator

It looks reasonable. What's the near wall & free stream mesh like?

• Posts: 6Member

In the paper they have specified that it is a 2D geometry so I have created a 2D geometry. The mesh looks like this. Also when I only provide mass flow rate the solution will converge but when I provide both mass flow rate and pressure Solution starts diverging.

• UKPosts: 8,855Forum Coordinator

Mesh looks OK.

You shouldn't need to add a pressure to the inlets as it's not supersonic. Why a mass and not a velocity boundary? Is heat & mass conserved?

• Posts: 6Member
edited December 2020

Because in the paper they have given mass flow rate as Inlet Condition. Yes, heat and mass conserved.

• UKPosts: 8,855Forum Coordinator
edited December 2020

OK, we tend to use velocity bc's over mass flow for incompressible flow. If mass & heat is conserved how well do cp etc compare with those used in the paper? Ie if you do a hand calc how far off on the properties would you be to account for 80K.

• Posts: 6Member

In paper they haven't specified Cp value but have only told us that specific heat is temperature dependent. So I am trying to match outlet temperatures.

My simulation Results :

Output Cold : 812 K

Outlet Hot : 615 K

Reported Value in paper :

Output Cold : 759.3 K

Outlet Hot : 673.4 K

Here is the link to the paper : https://www.sciencedirect.com/science/article/abs/pii/S1359431108002895

• UKPosts: 8,855Forum Coordinator

So, the solver is behaving and your problem now is to try and figure the material properties they've used. If it's temperature dependent you may struggle as you're trying to curve fit rather than find a single value.

• Posts: 6Member

I would like to know what could be the reasons behind this deviations and how could I replicate their results.

• UKPosts: 8,855Forum Coordinator

The deviation is between a case in which all the fluxes balance, and a case where we don't know the material properties and (probably) all of the fluxes balance?

• Posts: 6Member

So how can I overcome this problem ?