You don't have to use any goals at all. Flow sim will use it's own internal convergence criteria to decide when the solution is finished. The whole business of goals is to add (optional) convergence criteria. For the last year or so, I've stopped adding goals to my simulations. Keep it simple!
Does the simulation run okay if you delete all goals?
Joe from solidworks said that he just uses cut plots.. that's probably what you use too.
I'm interested in heat flux which will be a per area measurement.. I'd also like to verify that the surface heat source is dissipating 5W.
When I look at the goal plot then it shows only 3W. This makes me believe that the cutplot will also be inaccurate.
The simulation always runs ok whether I have goals or not.
In that case, use Surface Parameters under Results. You can select each of the inner walls and get the heat flux for each. If you are adding 5 watts, then you should get out 5 watts, once temperatures have stabilized.
I changed the simulation to have a surface source to keep things simple. I made the surface source 1W.
Looking at the heat dissipation paramaters and I think they should add up to 1W, but they don't.
It seems like the total energy shown is .5W from the pictures. I just wonder where the rest of the heat is going? Or some way I should see 1W of energy somewhere.
This was my concern with the surface goals I was setting--that the surface goals weren't valid or weren't converging.
The material for the surface heat source is the 'tutorial component package.'
The heat sink material is aluminum. and the other side to close the cold plate is some epoxy with low thermal conductivity.
The heat source is touching the coldplate.
The ambient is 80 degrees C.
I found out what was going wrong. I needed to increase the local mesh settings. To get the result to within 3% I had to set the channel, gap, and wall thickness requirement to very small..like 1mm.
I found that once an object touches another object then I need to mesh around that area. I haven't found out specifically what values I should set, but I know that this is how to solve the problem.
The point is that if your results aren't what you expect then you should try setting your mesh around your fine featured object very high. Then let your simulation run for a long time. Then check your results. If they are what you expect then you can set a more course mesh to make your simulation run faster.
I read a post that one guy sets the local mesh by taking off automatic settings. Then he raises all the refinement options to max then he simulates. That seems pretty good too.
Hope this helps someone.