For this application we would recommend using Particle Studies. This allows injection of liquid droplets into an air stream, and prediction of where they will condense or accumulate We have a basic tutorial here which uses an oil catch can / separator, but the physics should be pretty similar: kzfaq.info/get/bejne/gM5_isp0va3eYXU.html We'll also add the idea for the drift eliminators to our video topic list! Thanks!

Yes, Flow Simulation calculates the convection coefficients. To extract this value I would recommend setting up a surface goal on the fins to track "Heat Transfer Coefficient" parameter. It could also be extracted from surface parameters in the results.

Yes! Did you see the recent video on setting up an internal analysis? kzfaq.info/get/bejne/brhmY9xqvp-YaGQ.html We will be uploading a video on the details of meshing in Flow Simulation in the next couple weeks.

Hi Angus, you can do this by creating a Cut Plot and choosing option such as Temperature (solid) or Temperature. You would also need to adjust model display transparency, or hide the solid part so that you can see the cut plot

This example uses the built-in "Air" ideal gas model. As long as Gravity and Pressure Potential (gravity) is enabled in the General Settings then the buoyancy caused by density change of the fluid is automatically accounted for, this is what forms the natural convection currents and is responsible for all the air velocity visible in the velocity cut plot, as there is no other fluid flow defined in the project.

Yes that's correct, if you input some velocity as shown at 3:26 it would represent a forced convection case (such as a wind or breeze flowing over the unit). Note that some testing standards accommodate a condition similar to a mild breeze even for "natural convection" problems, and this can have a large influence on the result. Typically leaving all the velocities at zero and relying strictly on the natural convection will be the most conservative estimate. Including radiative heat transfer can also be an important factor for natural convection problems.

The heat source will always input heat so long as the sign is positive. If you want to extract heat, such as to represent some cooler or cold plate, you can enter a negative heat power in the heat source.

@@hawkridgesystems Thanks for this. Would it be possible to make a video explaining how you would find the temperature difference if a fluid (inlet and outlet), using equation goals, with a heat source applied? More so on what general settings to click on in wizard. I've found that choosing the heat transfer rate there and also applying the heat source does something weird. On the other hand, clicking adiabatic and then applying heat source works, but this seems contradictory. I'm also presuming the meshing on the fluid and solid is done automatically.

@@orrinbishop3326 it sounds like maybe you are referring to an internal analysis (such as a pipe / duct). The adiabatic condition you're describing sounds like the Default Outer Wall Condition. The Default Outer Wall Conditions will control heat loss through the "exterior" of the internal analysis - it will apply to any exterior walls that are not affected by a boundary condition. Typically we would leave that setting at adiabatic unless there is some known heat transfer loss rate to the outside environment. When the default outer wall condition is set to adiabatic, all the heat transfer will occur internally within the computational domain - heat may flow out of an outlet boundary condition for instance but no heat will be lost through the side walls. Setting it to adiabatic is equivalent to treating any exterior wall that does not have a boundary condition as insulated. It is a good idea for a future video and we will add it to our list!

"Volume" sources will be distributed equally in terms of volume. So if the two CPUs are identical (same volume) then it would be equivalent to do either method. If you are not sure if they are identical geometry then it is best to define the sources separately.

There was some video magic there - a timelapse of 1.5x speed for the first analysis and 10x speed for the second analysis. In reality the simulations took up to a few minutes to solve.

The heat transfer should occur automatically between coincident faces. Possibly the geometry has a small gap in it or an interference. In this case you could either correct the geometry issue, or use something like a thermal joint condition to span the gap. Another possibility is the solid material definition is incorrect - if one body is defined as a perfect Insulator then you would see no conduction through it

You can do this as a "Volume Source" selecting the third option. This would be the recommended method if you expect the source to absorb or add heat However take caution that when assigning a temperature, the result will always be this temperature. The unknown will be the amount of heat power absorbed/rejected There is also the option for wall temperature in Boundary Condition when specifying Real Wall