Hi,

I have been trying to simulate a double pipe heat exchanger ( 1" OD, 1/2" ID, 6' length) using CFX 5.4.1. I am getting promising results. Now I am trying to reduce the model size by simulating only one inch length ( out of 6 feet) of heat exchanger, by using PERIODIC boundary conditions. I came to know that PERIODIC boundary conditions can't be applied becoz, Pressure fileds are not periodic like the Flow fields do. Can anybody give me suggestions to reduce the model size?

Thanks,

Satya

Why can't the pressure field be periodic? i.e., a constant pressure drop. You can do this by setting a momentum source term to give the pressure drop you want. You might have to manually iterate a bit if you want to get a particular mass flow rate though.

Neale

Hi,

Obviously, you didn't understand the real meaning of PERIODIC conditions applied to variables such as Pressure, Velocity component, Temperature etc. In fact, PERIODIC condition for velocity component is just simplified as U|x+dx=U|x, however, the PERIODIC conditions for P or T is: P|x+dx-P|x=const and T|x+dx-T|x=const.

Wish you lucky.

Hi Neale & Jianwen,

Thanks for your clarifications and suggestions. In this case, I want to apply translational PERIODIC boundary conditions. If I model only 1 inch length of heat exchanger pipes, Do I need to specify OUTLET BC?( I guess not). When I modelled total length of exchanger, I specified OUTLET bc as "Relative Static Pressure =0". Neale suggested to set momentum source term to give required pressure drop. I dont know the pressure at outlet. I would like to predict the Outlet temp. and convetive heat transfer coeff. on walls. The BC's I can give are, Normal velocity and Static temperature at the INLET. Can you clear my confusion here?

Thanks, Satya

The problem with this is that eventhough momentum and mass may become periodic, the heat transfer (or temperature rise) is not really. Well, maybe in some section of your heat exchanger tubing the mass averaged temperature has a roughly constant streamwise gradient.

So, if you do what I suggested before you will simply end up with a uniform temperature distribution, i.e., the fluid eventually heats up to the values specified at the boundary, so that the temperature gradient simply goes to zero.

I suspect that what you want to model is the thermal boundary layer growth in an individual tube, which is a different problem that what you are considering.

Neale

Hi Satya,

In your case, in accordance to my understanding, is just a PERIODIC fluid flow and heat trasnfer problem. As what you mentioned, even you model only 1 inch length of heat exchanger pipes, you need to specify OUTLET BC ( Explicitly in fact!). What you did when you modelled total length of exchanger, you specified OUTLET bc as "Relative Static Pressure =0" and what Neale suggested to set momentum source term to give required pressure drop are both right. In fact, the pressure at outlet can't be settled and only the pressure drop can be settled in this problem. So is the conditions of temperature.

Jianwen

Hi,

what are the ball park values for the energy and momentum(U,V,W) soruce terms. I have been trying by guessing values randomly, but I can't get the flow rate and pressure drop I want. Can anybody threw me some values?

Thanks,

Satya

Hi Neale,

You are right. when I tried the simulation as you suggested, I have end up with uniform temp. distribution which I dont want. You said, we can model thermal/fluid boundary layer growth. How to do that? can you give some clue?

Thanks in advance.

Satya