CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   Electronics Cooling (

Russell King July 3, 1998 11:00

Electronics Cooling
Does anyone have any issues with respect to electronic thermal management that they would like to discuss?

I can start the ball rolling with;

In what areas do current CFD technology fail to meet the requirements of the design engineer?

Q. Pat Diao July 4, 1998 21:22

Re: Electronics Cooling
I think the current CFD tools are quite enough to handle electronic package cooling problems. Because compared to processes with phase changes, electronic cooling is a relative simple process.

Like to hear more from all of you.

Russell King July 6, 1998 06:51

Re: Electronics Cooling
On the face of it, electronics cooling sounds as if it should be a relatively simple application for cfd to handle. However, as power ratings go up and feature size comes down there is a mounting problem for thermal management engineers and heatsink suppliers alike. So much so that in just a few years time a new technology will have to be developed to cope with dissipating these larger heat fluxes.

In recent times the technology has advanced to produce devices such as heat pipes. Heat pipes contain a liquid that evaporates over the hot zone and condeses in the cool zone. Thus, touching on the multiphase fluid issue and other examples also exist such as transfering heat into a wax that melts when hot and solidifies when cold.

Currently the electronics thermal designer may have to deal with many application areas in order to meet his needs. Such as conduction, natural-convection, mixed-convection, forced-convection, internal-radiation, solar-radiation, laminar flow, turbulent flow, swirling flow, single-phase, multi-phase, steady-state-flow and transient-flow and on many physical scales form the micro scales of the integrated circuits to the macro scales of a pcb encosure or even the external flows around these enclosures.

There are many commercially available packages that can cope with the general fluid flow and heat transfer senarios but the one I would recomend is IcePak from Fluent as it can handle all the physics (including multiphase), it uses unstructured, automatic meshing and it can run in parallel (not that I am at all biased).

PS If you have any issues with predicting casting flows I would be happy to discuss them with you.

Jonas Larsson July 6, 1998 07:31

Re: Electronics Cooling
I'm a complete novice in all aspects of electronics cooling, but one question comes to my mind:

Is it possible to use CFD to predict the noise generated by a cooling-arrangement?

This is the cooling system aspect that at least I, as the end user of for example a computer, notices most. I assume that noise generation must be fairly complex to predict (closely coupled to turbulence etc.). Perhaps it's too complex to even try to predict. Can IcePak predict noise? Is anyone even trying to predict this kind of noise?

Russell King July 6, 1998 09:44

Re: Electronics Cooling
The prdiction of noise is not something that can be modelled with IcePak but there are specialist noise prediction programs available. IcePak does not model the detail of the individual fan blades, it simply uses the manufacturers fan-curve to automatically adjust the flow rate of the fan to the local pressure conditions, so there are no rotationally based transients imparted on the flow.

However, if there are flow instabilities within the domain, such as vortex shedding, then IcePak will be able to predict the strength and frequency of the vortex streat. There is a nice animation of this on the fluent home page

Robin July 6, 1998 12:28

Re: Electronics Cooling
The idea of modeling heat pipes using a CFD code capable of handling multi-phase flow is indicative of the academic and time consuming approach inherent in general purpose CFD mind-sets.

CFD codes such as Flotherm from Flomerics provides an engineer friendly electronic thermal analysis software package that fullfils its role by reducing time to market in the design and manufacturing area of the industry.

Heat pipe modeling is easily acheived with a knowledge of it's thermal performance in terms of an effective thermal conductivity vs. power graph. Not everything is as complicated and convoluted as general purpose CFD vendors would lead us to believe.

Russell King July 8, 1998 08:24

Re: Electronics Cooling
The mention of multiphase flow in electronics enclosures appeared in a previous posting to illustrate that there is more to electronics cooling than meets the eye.

I am sure that manufacturers of heat pipes would be very interested in the internal flow and performance of their devices. Whereas, opperators would only be interested in the macro performance of a heat pipe.

There will always be companies that are interested in the detailed physics and workings of devices and others that are not; it is reassuring to know that they at least have the option with IcePak.

As regards the ease-of-use question that arose, IcePak is a fully graphical package with a "drag-and-drop", object orientated interface designed explicitly for use in electronics enclosures. Even the unstructured meshing is invisible and automatic; what could be simpler?

Rob September 8, 1998 11:48

Re: Electronics Cooling (ICEPAK or FLOWTHERM)
How would be compare Fluent ICEPAK with FLOWTHERM. I'm currently using Flowtherm and the interface seems a little primitive. Since there are not may commerical software in the electronics cooling field, is Flowtherm the only option ? Dont know if any other software could/will challenge the capabilities of Flowtherm. CFD users in Electronics Cooling field, pl. throw some light to this question.

Russell King September 9, 1998 06:31

Re: Electronics Cooling (ICEPAK or FLOWTHERM)
Icepak has been specifically designed to address the needs of the electronics community and has many features that make it a world class product.

There is a demo on the Icepak website;

that may be of interest.

Robin Bornoff September 11, 1998 02:08

Re: Electronics Cooling (ICEPAK or FLOWTHERM)
Flotherm V1.4 was indeed a simple but effective package with what could be regarded as an interface that was a little primitive. Flotherm V2.0 is radically different in its UI with a wealth of CAD/library/Hierachal assembly functionality.

Remember, Flomerics is dedicated to the electronics cooling market. The software is only one piece of the overall approach we have. It is backed up with educational services, experienced support teams, www resources and advanced integration capabilities so that Flotherm can fit easily into current industrial design environments.

...competition is always healthy though...

Rob September 16, 1998 14:33

Re: Electronics Cooling (ICEPAK or FLOWTHERM)
Can Flotherm v.2 handle non-rectangular shapes? What about ICEPAK? Does it have any restrictions in modeling shapes ?

Russell King September 17, 1998 11:39

ICEPAK - Flexible Unstructured Meshing
Unlike other cartesian based packages, the underlying Icepak solver is based on Fluent's unstructured software which has no real restriction on geometry type or orientation. This is because Icepak can handle various cell types, (ie tetrahedrals, pentahedrals and hexahedral or even a mixture), thus, accurately capturing the geometry.

See the image of the heat sink on

Not only that, but using solution based adaption it is very easy to automatically refine the mesh locally where the gradients of temperature (or any variable) are high. This leads to maximum solution accuracy with the minimum grid size.

KwangWon Suh September 23, 1998 01:42

Re: Electronics Cooling (ICEPAK or FLOWTHERM)
CFD-ACE have good capabilites for Electronics cooling such as conjugate heat transfer through multi-media, Discrete Ordinate methods in radiation and you may be able to simulate a thermal stress using CFD-STRESS to be released commercially soon using strongly coupled method. If you need more information, please visit ,

Drazen Beatovic September 25, 1998 22:47

Re: Electronics Cooling (ICEPAK or FLOWTHERM)
You may want to check

They have package for electronic cooling that is integrated in I-DEAS by SDRC.

All times are GMT -4. The time now is 11:23.