SolidWorks Simulation: Thermal 1

[Bingo123]

I am trying to run what should be a simple thermal analysis on Simulation, but it keeps giving me odd results.

I want to run an analysis where a flat, square, steel plate (representing a panel of a container). On one side (surface A) there is one fluid (fluid A) that is at ambient temperature. On the other side (surface B) is a fluid (fluid B) that begins at ambient temperature also, however the temperature decreases with time before settling at 4C.

I want to be able to plot the change in temperature of surface A with time, as a result of the temperature difference between the two fluids, which are separated by the steel plate.

I have attempted this by applying a convection BC on side A of the plate, an initial ambient temperature condition on side A of the plate and a transient temperature condition on side B of the plate, reflecting the change of fluid B temperature with time (which is known). However, this has yielded bizarre answers (Temperatures in the range of thousands of Kelvin). I have checked and all dimensions/ input values are correct.

Please could someone advise as to the best approach to solving this/ the correct BCs to apply?

Thanks!

 

[corus]

Mixed up the heat transfer coefficient (generally in thousands) and the ambient temperature?

Another cause may be beacuse you are running a thermal transient and the solution is unstable, causing odd results. You can spot this if the temperatures appear more reasonable later in the transient.

Tara

 

[Bingo123]

Thank you!

I have just had a look, I don't think this is what the problem is unfortunately.

The plate seems to be being heated by the BCs I have applied, although what I want to model is the effects due to conduction through the plate.

 

[corus]

As a check on your model, your calculated temperatures must be somewhere between A and B unless your initital temperatures were outside that range, or you have internal heat generation. If not, then the problem must lie with the boundary conditions or the software. If you'd made a mistake in the thermal properties then you might get the wrong answer but still within the limits I've said above. Hence it must be the boundary conditions that are wrong. It would be extremely doubtful that the software would give such bizarre results for a tried and tested load case.

Tara