[YuguiYo]

Hi all,

Once again, i'm stuck in my thermal simulation...

I got one solid that send thermal wave (volume source pulsations). 6 thermal sensors are placed all around it.
The aim of this configuration is to calculate the time taken by the thermal wave to reach those sensors.

My problem is that I don't see that phase shift of the temperature wave... In fact all the sensors see the thermal wave in same time but with not the same temperature.

Does anybody know how can I fix that, or anybody have faced the same thing ?

Thanks for your time,

Hugo

[ghorrocks]

Please post an image of what you are modelling, an image of your mesh and an example out file.

[YuguiYo]

Hi Glenn,

Thanks for answering me !

I put all the file that you're asking.

To sum up :

I got a platinum resistance that create a thermal source. 2 sensors are placed at 15 µm at both sides of this resistance. All of them are placed on a 500 µm SiO2/Si wafer.

There is a water stream, with inlet velocity of 0,05 m/s.

The thermal signal created by the resistance diffuses in the water and the wafer. With the stream, the temperature received by the sensors are different.

Now I want to know how much time that signal has taken to go from the resistance till the sensor.

Thanks again for your continuous help Glenn !

Hugo

[ghorrocks]

Have you done a time step sensitivity check? You need to make sure that your time step is fine enough to capture the physics. Rerun with a finer time step and see if it makes a difference.

While you are at it check your convergence criteria and mesh density as well.

[YuguiYo]

Indeed you're right, when I do the simple calculation between velocity - time - distance, i found a minimum time step of 3.10^-4 s !

I'll try with that time step, but to simulate 0,2 s with 3.10^-4 s time step, it will be very long for the old computer i'm using ..

I've looked on internet for time step sensibility check, but couldn't find a good threat about it. Can you explain me how to check or do you have a link to a good explanation on how to do it ?

The convergence criteria is at 1e-4 (i'll try with 1e-5), and the mesh density is ok atleast and OK for all the domain solids.

I'll let you know with the new implements if i can see the time phase shift for my thermal signal.

Thanks !

Hugo

[ghorrocks]

I do not understand your minimum time step of 3e-4s. Where did that come from? Do you mean Courant number?

I think you should have a look at this thread which is discussing time step selection - http://www.cfd-online.com/Forums/cfx...-stepping.html

My recommendation is on post #3.

[YuguiYo]

My waterstream velocity is 0,05 m/s and the distance between is 15 µm, so basically in 15.10^-6 / 0,05 = 3.10^-4 s the thermal signal go from the transmitter to the receiver. Therefore, if I want to see the thermal time of flight, I have to put a minimum 3.10^-4s timestep (i've tried with 1.10^-4s)

Thanks for the link !

Hugo

[YuguiYo]

It's working with the 0,0001s timestep !

Thanks !