CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FloEFD, FloWorks & FloTHERM (http://www.cfd-online.com/Forums/floefd-floworks-flotherm/)
-   -   Problems using TEC in Floefd (http://www.cfd-online.com/Forums/floefd-floworks-flotherm/96737-problems-using-tec-floefd.html)

Nadav January 31, 2012 04:37

Problems using TEC in Floefd
 
Hi,
I need to run an analysis that includes a TEC that is responsible for a heat-cool cycle. The TEC initial condition is that the temperature difference between the cold and the hot side is 0 (both at ambient). It then has to heat a component to a designated temperature and after a curtain time - cool it back to ambient temperature again.
I seem to have an inconsistency between the analysis and some tests I ran: The TEC in the Floefd seems to heat up much faster and to significantly higher temperatures than it does when I measure the real TEC. It heats up out of its spec.
Could it be that the TEC in the Floefd is not modeled for this operation mode?
I would be grateful for any help.

Thanks.

Boris_M February 6, 2012 10:40

Hi Nadav,
I don't think this will work with the way the TECs are defined in FloEFD. They are basically just a bridge with a heat transfer between two parts but themselfe have no heat capacity or thermal resistance etc. Now heating them up transiently can result in much faster heating time as no transient effects of the TEC itselfe is known and for the limits, it should also work according to it's data. But I cannot really tell as I personally never used them. If there is any test case such as a benchmark case you should try with that but steady state first.
Also for heating up it works in general but then cooling down would not work as you define the hot surface and you would then have to switch it. Don't think it works just by using a negative current.

Regards,
Boris

Nadav February 7, 2012 02:50

Hi Boris,

Thanks for the reply.
Floefd indeed does not accept negative values for current in the TEC deifnition.
I also ran a test with the real TEC I intend to use and found that the temperature differences in the test are much smaller than in the Floefd. I figure from what you mentioned that it may be that the TEC is deifned in the Floefd as a perfect insulator while in reality it is not.

What I plan to do now is:
1. test the TEC to find it's heat resistance (by heating one side with an outer source and measuring temperature both sides until steady state)
2. assume the heat capacity (Ceramic material)
3. deifne a part in the Floefd that has the temperature profiles and resistance I measured and capcity I assumed and hopefuly get a close enogh model of the TEC to work with.

Does that make sence to you...?

Boris_M February 7, 2012 10:28

Yes, it's basically an engineering model where you have the delta between the two sides and the body/part in between (the TEC) is just functioning as a reference geometry to the features settings. As you specify the hot surface, the rest is done by the properties you defined. That way you avoid considering the complex physics in it but then of course these models have limitations just like fan curves. They only give you a flow and swirl but you don't have the details of what happens inside the fan as if you calculate it completely.

I don't think this will work, as you said the TEC is an insulator, just using the properties defined in the TEC feature. You will probably either not be able to define a material on it orit will ignore the material. And simply a material doesn't give you the specialities of a TEC to suck the heat against the gradient.

Boris

zdunol March 24, 2015 08:38

Hi,
I want to simulate performance of three stage TEC in flotherm, my strategy is too calculate temperatures and heat fluxes on all four surfaces and put together three TEC smartparts in flotherm, but with this approach I don't know what to do with each Imax DTmax and Umax. Do you know if I can just put three stage TEC (whos geometry is not a simple cuboid but more like a pyramid) into flotherm as one smartpart?

Thanks for any help ;)


All times are GMT -4. The time now is 11:01.