[Hamda]

Hi there,
I am supposed to model heat transfer through a packed bed of spheres. But there is some problems .
First, each sphere is a energy source (transient) and energy source is a function of time. So the Temp. of center of each sphere is different with the surface temperature. So how can I define it to CFX?
Second, I use CFX 14.5 and I don't know it is beneficial to switch to 19 V?
Many Thanks,

[ghorrocks]

Your question cannot be answered without knowing what you are trying to model, what you are hoping to learn from the model and why you are doing the model.

[Hamda]

Originally Posted by ghorrocks

Your question cannot be answered without knowing what you are trying to model, what you are hoping to learn from the model and why you are doing the model.

Thank you for your response.

The aim of my project is to simulate the temperature distribution on spheres (fuel elements) as well as fluid temperature (coolant).

Each sphere is an energy source that is a function of time (unsteady condition). So it is evident that there is a difference temperature between center and surface of the sphere.

Now My problem is how to define this to CFX? should I specify each sphere as a different zone and name that? How can I define energy source for each sphere? should I define just ONE energy source for all and then just specify Tc and Ts for each sphere?
Thanks

[ghorrocks]

You will have to explain in more detail than that.

What sort of heat source are they? Where does the power come from? Is the heat source uniform, in the centre or on the surface? How is the fluid affected by the heat? What is the fluid? What is the time scale of the simulation? What is the time scale you want to look at? Why are you looking at this anyway?

Until I have some idea of what you are doing and why you are doing it I cannot help.

[Hamda]

Originally Posted by ghorrocks

You will have to explain in more detail than that.

What sort of heat source are they? Where does the power come from? Is the heat source uniform, in the centre or on the surface? How is the fluid affected by the heat? What is the fluid? What is the time scale of the simulation? What is the time scale you want to look at? Why are you looking at this anyway?

Until I have some idea of what you are doing and why you are doing it I cannot help.

yes, you're right.

The heat source in the packed bed generated by nuclear chain reactions . Indeed each sphere contains thousands of Uo2 as shown in pic1. This packed bed is a slice of a reactor core.

for the steady-state condition, the energy source is assumed to be constant (w/m3), but for the unsteady state condition, it would be a function of time.
I should use the following equation :

Q =q A (ts^a - (t0+ts)^a)
where, q is reactor power (w/m3) in a normal condition (steady state) and t0 is the reactor operating time, ts is the reactor shutdown time, and A and a are the constants.
working fluid is helium and considered as a coolant. Helium coolant flows through the spheres and gets heat from them and as a result its temperature would increase along the packed bed.

well, time scale is variable. it can be 10s to 10^8 s and it is specified by defining ts. Also I can provide you with an article if you are interested.
Thanks

[ghorrocks]

As soon as you mention time scales from 10s to 10E8s then you can forget CFD to cover all that. The fluid time scale is going to be short, so you are going to be using time steps of the order or fractions of a second. This means that simulations covering 10E8s are not going to be practical. You are going to have to separate the time scales to cover that, which probably means assume a constant heat source and work out the fluid regime. This will give you a convective heat flux. Then model the long time scale heat source using a constant surface convective heat flux. If you want to get sophisticated you can model a few heat fluxes representing a few points over life and get a heat transfer coefficient versus time curve and apply that to the long time scale model.

For the fluid model you need to think about whether periodicity as you have modelled it applies, and if so how it applies.

To answer your initial question: implementing this is simple, and similar things are covered in the tutorial examples. Also any recent version of CFX will be able to do this. But, I think you have more fundamental issues with your approach which requires further thought.

[Hamda]

Originally Posted by ghorrocks

As soon as you mention time scales from 10s to 10E8s then you can forget CFD to cover all that. The fluid time scale is going to be short, so you are going to be using time steps of the order or fractions of a second. This means that simulations covering 10E8s are not going to be practical. You are going to have to separate the time scales to cover that, which probably means assume a constant heat source and work out the fluid regime. This will give you a convective heat flux. Then model the long time scale heat source using a constant surface convective heat flux. If you want to get sophisticated you can model a few heat fluxes representing a few points over life and get a heat transfer coefficient versus time curve and apply that to the long time scale model.

For the fluid model you need to think about whether periodicity as you have modelled it applies, and if so how it applies.

To answer your initial question: implementing this is simple, and similar things are covered in the tutorial examples. Also any recent version of CFX will be able to do this. But, I think you have more fundamental issues with your approach which requires further thought.

Sorry, I made a silly mistake
Time scale would be 0.01 s to 10s. It depends on ts (the reactor shutdown time). Now I think I have some ideas. I had a glance at ANSYS CFX Tutorials (14.5). However, I didn't find an example of a packed bed of spheres. It will help me a lot if I find such tutorial. Do you have any idea? I really appreciate it.

Thanks

Hamda

[ghorrocks]

If your maximum time is 10s then your timescales look OK and pure CFD should work.

There are no tutorial examples of packed bed of spheres. The tutorial examples I was talking about show you use of CEL expressions and implementing heat sources, but on very different geometries.