In a thermal simulation in SolidWorks Proffesional 2011, I have 237.6 Watts of power going in (heat capture.png). This is 6.4*36 + 2*3 + .4*3 Watts. I have repeatedly verified this. But in the heat power summary, 489 Watts are listed (heat capture 2.png). What is the mistake I am making here?
I happen to be working ona thermal model and mine is as expected. You may want to check the sttings on your specified input powers. There is an option for total or per item. You issue may lie there or it might be something else.
Bill,
I understand your skepticism. But you can see in the attached photos that the power input is more or less as I described it.
I also suppressed all of those power inputs and reduced them to one 6W input on one surface, so that there was practically no chance to make a mistake.This doesn't get rid of the problem. Here is the output with one 6W input.
From Solidworks Help Files, located at Simulation > Analysis Background > Thermal Analysis > Heat Power/Energy PropertyManager
: "If heat power is applied to a part with multiple bodies, the software adds the positive (flowing through the body) and negative (flowing out of the body) heat power for each body. The total heat power flowing in and out of all bodies is listed under the Entire Model."
So, if this is true, then a single body with a 6 Watt input will have a 6 Watt output. The sum of those is 12 Watts, and your problem balances. But, I actually think the value for "Power In' is the sum of your 6 Watts into 1 component of your assembly and that same 6 Watts moving into another component of your assembly. I think the same is true for the 'Power Out' value.
I have a model with multiple bodies, and about 50 Watts total input and output. The values for 'Power In' and 'Power Out' are each about 200 Watts. I thought about this, and realized that 50 W goes into several PC boards, then 50 W goes into several page frames, then 50 Watts goes into several housing panels, then 50 W goes into the baseplate. Summing those up results in about 200 Watts.
If I add what goes out: 50 W out of PC boards, 50 W out of page frames, 50 W out of housing panels and 50 W out of baseplate, that also adds up to about 200 W. My 'Net Power' value is very low, similar to the one in your example above.
Thsi could be tested by doing an serial conduction analysis on 2 bodies, then repeat with 3 bodies, then repeat with more bodies. If the logic works, then adding a body would add the same amount of heat each time.
So, is this a good approach for Simulation? Do they need to make it clearer on the screen, instead of requiring us to dive through the help files to find a rationale?
not sure what the question is here.
if you select the components that create heat, all the faces you'll get power in
if you select the faces where the heat leaves, you'll get the power out
in mike's example the software is reporting correctly. power in = power out. i'm not sure why the calculated power doesn't match his expectations. without the files, its hard to investigate. but we've done a bunch of these balances for customers on tech support and in our consulting and they've always come back right.
This is an old one. But the problem was that the total power in in the jpg did not match the total power into the model. The problem turned out to be that power moving from one body to another was counted as out and in again. So the value reported is not the total power into the model and the total power out. It is the sum of the total power into each body and the sum of the total power out of each body. These are different values and that was confusing, because all of the poweer is double, triple, quadruple counted..
sounds like double counting. would be interesting to look at a simple example to see if that is just the case or if there is a workflow or software limitation/issue here.
great example of double counting is reaction forces. if you select all the components that you want, your reaction forces will be correct. if you choose them individually, and add them, you'll get double counting at the edges. but inside the interface, this shouldn't exist.
The easiest example to understand would be to nest spherical shells and put 1W in the middle. You will see 1W x the number of shells in and out.
hi mike, if you have example, i can check it out. not sure i'm visualizing it properly.
