CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)

 coolcrasher June 16, 2015 04:58

Simulate cooling of bread in ansys cfx

I wanted to simulate cooling process for a bread. Following is the test case:
I have bread initially at 95 deg C, and it is cooled by passing air at 25 deg c with some flow velocity.I have to estimate the required time to cool the bread and observe the flow pattern around the bread at each time step. ( transient analysis )

However it is an initial phase of my work. Later I have to cool the bread by mixture of air and water droplets. ( mulitphase flow )

Can you guide me how I can go forward with the initial situation and simulate the initial case in Ansys CFX?

 ghorrocks June 16, 2015 05:20

Do you need the temperature profile inside the bread?

Is the air cooling slow or high speed? In other words, is natural convection significant?

What do you have to observe on the flow pattern?

You can model this as a simple ODE with no need for CFD at all if you want. Just Newtons Law of cooling.

And most importantly..... won't cooling the bread with water droplets make it go soggy? Yuk!

 coolcrasher June 16, 2015 05:38

Do you need the temperature profile inside the bread?
Yes, I have kept bread in a box like structure which has an inlet and outlet. Air flows through inlet and flows out through outlet. At each time step I have to observe the temperature and flow velocity contour until my bread cools down.

Is the air cooling slow or high speed? In other words, is natural convection significant?
I guess it is a problem of forced convection. We send air through inlet with some velocity of 0.5 m/s to cool the bread.

What do you have to observe on the flow pattern?
velocity pattern of the air and tempurature profile in domain

You can model this as a simple ODE with no need for CFD at all if you want. Just Newtons Law of cooling.
However it is part of my work to simulate this case in CFX.

And most importantly..... won't cooling the bread with water droplets make it go soggy? Yuk!
It is not exactly air with water droplets, in industry the bread cooling is done by sending moist air so that the moisture in the bread doesnt go out during cooling which make bread dry. So they regulate the moisture in the air to prevent this from happening.

Mine is the conjugate heat transfer case but there is no heat source or sink. But just some initial conditions on bread. I have followed the Chapter 16: Heat transfer from a heating coil from ansys cfx with some modifications. ( I do not have the heat source as it has one ) Hence looking for options.

 ghorrocks June 16, 2015 06:09

The air cooling one is easy using CHT. Give the bread representative density, thermal conductivity and specific heat for material properties. Then give it an initial temperature.

If natural convection is not significant that this simulation probably does not need CHT. You can do the air simulation to get the heat transfer coefficient, then just apply those heat transfer coefficients to a solid only thermal model. This will be much faster and easier than the CHT simulation.

 coolcrasher June 18, 2015 09:39

I have not understood, natural convection is significant also.

Here is what I have done:

Inintially I have given the properties for bread as told and initialized its temperature with 95 deg c. And solved for heat transfer with thermal energy. However when I do this in around ten seconds the bread is getting cooled to 25 deg, which is not true. The problem was I was using steady type in analysis.

Now I have initially simulating for steady state with heat transfer as isothermal for bread for some iterations then I am using this solution as a initialisation for transient analysis changing the heat transfer in bread to thermal energy for transient case. So now in steady state the bread doesnt cool down ( it is as expected ) but once it comes to transient case the temp of bread remains as same as before without drop in temperature.( which is expected to drop exponentially)

Please guide me where I have went wrong.

 ghorrocks June 18, 2015 18:52

Of course you need to model this transient. Don't be fooled by the time step size reported in steady state analysis - it is not "real" seconds.