[zdunol]

Hello good people,
I would be greatful if anybody could calrify to me

what are the differences between those 3?

I used Flotherm and FloEFD so I know the basics like the way Flotherm captures geometry, or that I can do rotation in FloEFD and all the stuff from the tutorials. It looks to me that FlothermXT is an advanced version of Flotherm, but it would be surprising if Mentor would still work on Flotherm in case Flotherm XT was just better, so I am wondering if you could name any cases where flotherm can do better than Flotherm XT.
The same thing applies to FloEFD vs Flotherm (XT), is there something Flotherm XT can do that FloEFD (with i.e. electronics/led extension) cannot?

I could only think of interfaces between EDA software and Mentor - it looks like FloEFD wouldn't work so seamlessly with Altium, or Xpedition, am I right? Oh yah, and meshing is different in Flotherm, it is quicker I got used to it
So, why would anybody buy FLEXX license, which gives you Flotherm and Flotherm XT whereas Flotherm XT is just better?

[Boris_M]

Hello Pawel,

No Problem:

FloTHERM and FloEFD are two completely different products with different meshing technologies and different applications and capabilities in the ones they overlap.
FloTHERM is specifically for electronics cooling and also uses Cartesian mesh but has to create stair-steps for round geometry whereas FloEFD is a general purpose CFD tool that has some capabilities in electronics cooling but not as sophisticated as FloTHERM and uses a Cartesian mesh with a special meshing technology similar to the Immersed Boundary method to resolve the complex shaped surfaces without stair-steps as it is integrated in the CAD system and needs to treat the complex CAD model in detail.

Now FloTHERM XT is a hybrid of both tools with the meshing and fundamental solver technology of FloEFD but tuned to the application of electronics cooling with the capabilities and features of FloTHERM. It is still evolving as the current version is 2.3 so basically about 2 years old and more and more new features will go into XT, features that FloTHERM has but XT not yet but also features that FloTHERM does not have but FloEFD has.

So FloTHERM XT can capture complex geometry better than FloTHERM but will not be able to calculate cavitation in liquids or combustion and many other physics that FloEFD can calculate as it is not required in electronics cooling which FloTHERM XT is designed for.
With the capabilities of FloTHERM XT it is able to better plug into a Mechatronic environment where MCAD geometry is getting complex and ECAD (EDA) information is of course important for good electronics cooling simulations.

FloTHERM is very fast and ideal for everything more square shaped like big server racks or anything boxy. FloTHERM XT is handling complex geometry better, so everything where design in the MCAD geometry comes more into focus such as a nice shaped Xbox housing etc. It is simply easier to work with such geometries then.

FloEFD is ideal for those who don't want to leave their Catia, ProE/Creo or Siemens NX environment and don't need detailed electronics cooling but are happy with a heat source on a block as FloEFD does not have detailed electronics modeling such as Delphi models, it only goes up to 2-Resistor models as probably only a few MCAD designer really do e-cooling simulations in high detailed electronics modeling.

Yes, Meshing in FloTHERM is there from the second you place a component because it uses the boundaries of that component to create the mesh and since it are only Cartesian cells in their simpliest form they are described in an instant compared to the partial cell technology that contains either solid and fluid regions or multiple solid regions of different properties. They have to be generated by Boolean operations which are done during the meshing operation and is therefore a separate step in the model setup before the solver runs.

Let's consider a traditional computer vendor/manufacturer that has boxy desktops but also nicely shaped notebooks or even WLAN routers with more complex housings like Dell with their 19" rack units and projectors or tablets and desktops as well as other electronic equipment or HP with similar products. FloTHERM is ideal and super fast for racks and boxy desktops but has more issues with design objects.
Just as an example, which one would you rather use with FloEFD meshing technology that FloTHERM XT has or FloTHERM meshing technology from your experience in these options:

http://www.bestbuy.com/site/sony-3d-...&skuId=6359019

http://www.bestbuy.com/site/viewsoni...&skuId=8848612

Or this example:
http://www.bestbuy.com/site/netgear-...&skuId=8998506

http://www.bestbuy.com/site/belkin-n...&skuId=2089307

Or consider curved TVs vs. regular TVs

Or these two desktops:
http://www.bestbuy.com/site/hp-pavil...&skuId=3311063

http://www.bestbuy.com/site/hp-pavil...&skuId=7117105

I hope this helps,
Boris

[zdunol]

Thank you very much for the answer Boris, yet I'd like to know more
-Is it really impossible to incorporate Delphi models into FloEFD? It looks like a great disadvantage to me, since 2-resistor models might be inaccurate.
2-FloEFD has this extensions like LED cooling package, or electronics coling package, or advanced modelling package. It looks like LED package covers radiation phenomena pretty well, which is a plus in aerospace industry wchich the company I work for is doing, so do you know if Flotherm XT or Flotherm is capable of modelling radiation like FloEFD (LED extension), wouldn't FloEFD electronics cooling extension cover delphi models, heat pipes, fans, TECs, heatsinks and all the smartparts we have in Flotherm?
3- FloEFD can also do phase changes (evaporiation, condensation, freezing?) can Flotherm XT do that?
4- I'm not sure if I would need to consider cavitation, rather heat transfer due to boiling. I guess FloeFD incorporates heat of evaporization but I am not sure if the model it uses allows doing sophisticated stuff or it is just there to check if cavitation/boiling occurs, what do you think?
5- can i create a heat exchanger in Flotherm XT? It requires velocity/flow rate inlet/outlet boundaries, good modelling of fluid flow, Flotherm has this fixed flow smartpart, can I use it as an inlet boundary? :P

Thanks,
Zdunol

[Boris_M]

Hi Zdunol,

1. No, of course it is not impossible but FloEFD is not planned to become a FloTHERM competition, there is a cut at some point where you will have to go to FloTHERM to do more detailed e-cooling simulations.
It is like with every product, Porsche also builds Panamera and Cayenne and both are sports cars and have 4 doors still there is a difference. HP builds Workstations and weaker Desktops because there is a difference in the user requirements.

2. Yes, the LED Module but it is not just for LEDs but also for lighting in general and is used in the automotive lighting industry extensively because of the Monte Carlo radiation model and the condensation model. It also has the E-Cooling Module with some features such as heat pipe, a simple PCB model, the 2-R model and joule heating etc. The Advanced Module has combustion, hypersonic flow and some other features which only few users need with some more advanced physics requirements.
The most advanced radiation model (Monte Carlo) is only in FloEFD as it is extremely important for the lighting industry or in applications where the focusing of radiation either by a reflector or a lens is important. FloTHERM and FloTHERM XT have radiation models but in electronics cooling there usually is no absorption in semi-transparent solids or focusing to a hotspot or wavelength dependency like the Monte Carlo is capable of. FloTHERM XT has the DTRM model and FloTHERM either the same or similar one like the Surface-to-Surface, I'm not sure and would have to look it up. These models are enough for regular heat radiation as it appears in electronics cooling applications.
Not all of what you mentioned. FloEFD's E-Cooling Module has 2-R models, heat pipes, fans curves (do exist in the basic version but there is an additional database in the E-Cooling Module with fan curves), TECs (also in the basic version but again an extended database in the module), there is a black box model for a heat sink with fan as a kind of smart part but no specific heat sink model or smart part where geometry is created automatically or something like that. FloEFD is based on Geometry in the CAD system whereas FloTHERM is based a lot on Smart Parts you can simply place somewhere by some basic definitions in the feature menu.

3. No, FloTHERM XT cannot do that.

4. The condensation/evaporation model is a film model so you can define a film thickness as an initial condition in a transient simulation and the evaporation energy is considered in the simulation but it is a film model so the film thickness is to be much smaller than a cell size. The cavitation can handle isothermal but also for water specifically the full temperature range as the properties for water are well known over the temperature compared to some industry liquids such as oils or fuels. There is also the real gas "Water" which partially can work in that range but all of them do not distinguish the different boiling methods which have a significant influence on the heat transfer. If you have any more detailed task definition I can give more information if that falls within the range of capabilities of FloEFD. FloTHERM and FloTHERM XT cannot do that at all.

5. I'm not 100% sure but it should basically be possible as liquid cooling of power electronics is of course an important factor and an air side should be possible too. I just cannot tell about the specific boundary condition types.
But if heat exchangers get more detailed with louvers etc. a special porous media type in FloEFD is way more better suited than resolving such structures with cells in the whole model. This is used for example in automotive oil-water cooler or oil-air or water-air cooler like for intercooler of charged air etc.

I hope this helps,
Boris

PS: I wasn't quite accurate about the FloTHERM stair-step meshing as it uses a special approach for the cells FloEFD calls partial cells, where the cell is partially fluid and partially solid. Just to be correct here :-)

[zdunol]

Thank you very much Boris, it really looks like you know a lot about Mentor's products.

I really need to take a look at Flotherm XT because detailed ICs modelling is a must for us. Do you happen to know if Flotherm XT covers variation of material absorptivity with respect to wave length, or material temperature, or maybe even angle of incident (im not sure if the third one is physically correct or if I will ever need to include it in simulation - it's been a while since my last physics class )

It's hard to choose a proper software since we need it to cover robust CFD solver (for doing heat exchangers), EDA features and radiation since we will do space stuff. I guess I will have to do a compromise somewhere
It would be great if Flotherm XT could handle flows like FloEFD does.


Btw do you happen to know if any of theese can incorporate radiative heat transfer from and into fluid?

Regards,
Zdunol

[Boris_M]

FloEFD is capable of handling such radiation cases in detail but lacks the Delphi model for chips. FloTHERM can do the Delphi but does not take into account the wavelength dependency.

What do you need the wavelength dependency in the absorption for? Do you have glass (semi-transparent) solids in your assembly?

What exactly do you want to do with the radiative heat transfer from and into fluid?
And what fluid do you want to consider?

Boris

[zdunol]

Yes, there will be some lenses, filters and windows and it would be great if we could use one software to do electronics, conjugate heat transfer, phase changes and radiation phenomena, and do it fast with concurrent approach
But I guess FloEFD's LED extension is there for thermal analysis and it won't be possible to check if for example our housing will suffice in protection against gamma rays right?

I was just wondering if radiation from fluid is somehow included in CFD in general. It just
occured to me that all the solvers concern subdividing solids into surfaces but I have never seen anybody mentioning fluids (liquids). I guess it would be important in space cooling applications, where there is no convection outside the assembly.

Lack of delphi models in FloEFD seems to exclude it from consideration unfortunately

Btw, as a customer I would love to be able to easily access more technical documents describing each product in general. Materials Mentor distributes is impressive I have to admit - great marketing, but remember that CFD does not stand for Colors For Directors (only)

Thank you very much for the help Boris.

[Boris_M]

Since gamma rays pass through opaque solids they are not really considered in FloEFD as radiation in opaque solids are only interacting with the solids surface (reflection or emission). Transmission and absorption are handled as a volume property and only for absorptive (semi-transparent) solids. So I doubt it will help with gamma radiation.

Yes, the reason is often because it depends especially in gasses on the pressure of the gas and especially in combustion cases it is of interest and here you also have tiny sooth particles which make out a large amount of the radiation emitted in the gas.

The only way to go around the delphi model is with a detailed modeling of the chip and applying the material properties but that requires of course detailed knowledge of the chip and its materials.

There is at least a detailed document that comes with FloEFD that describes all the features and physics of FloEFD such as the radiation models etc. There are also some white papers on certain capabilities online and worst case, always contact the support if you have a more detailed question about a capability that you need and let them help you with the answer if FloEFD or one of the other products is capable of it.

Boris

[CFDfan]

Boris, thank you for your explanation.

My company is in the business of Electronics cooling of Power electronics. Some of our large power enclosures have a lot of parts and complex shapes and we use SW for their mechanical design. We've been using SWFS for more than 7 years and are happy with its integration with SW, user friendliness, stability and reasonable accuracy.

My questions are as follows:
1. What are the features present in FloEFD but not in SWFS (related/needed in Electronics cooling)?
2. I understood that Flowtherm XT uses the meshing algorithm and the solver of FloEFD. In that case what advantage would XT bring to electronics cooling people who already have SWFS.

[Boris_M]

Hi CFDfan,

there is no difference between SWFS and FloEFD with regards to electronics cooling at the moment.
Yes, FloTHERM XT is based on the meshing and solver of FloEFD but uses the electronics cooling know-how of FloTHERM.
FloTHERM is the electronics cooling market leader and has waaaaay more electronics cooling capabilities than FloEFD has. It is specifically built to the electronics cooling applications whereas FloEFD is a general purpose CFD code and has some but nearly as many functionalities for electronics cooling as FloTHERM and FloTHERM XT has.
The good thing with FloTHERM XT is you can use the SW models directly in FloTHERM XT as is is based on the SW modeler and therefore all parametric data of the SW model is available and the geometry can be adjusted just like in SW.
In FloTHERM XT you can use EDA import in which you will get not only the PCB data from a simple block to a detailed PCB with single traces from ODB++ data (detailed Copper nets) as well as automatically getting all the EDA information about heat generation rate of the imported ICs and other components in that comes from the EDA tool. So if you import a PCB you can filter components by size as well as dissipated power and also select which detail you want to have them, as a block, 2R or more detailed like a DELPHI model etc. from FloTHERM Pack (www.flothermpack.com) where you can generate detail chip models or also use them from most major vendors. You can also capture power modes of components
And FloTHERM XT has just like FloTHERM a smart part library so if you need a heat sink you can basically drag and drop it out of a library with just a few definitions about the size and fin height etc. The same goes for many other components.

So if you are working strongly in the electronics cooling application FloTHERM or FloTHERM XT would be the best solution from Mentor Graphics. If you like the meshing and the direct use of your SW models then FloTHERM XT is the strongest electronics cooling solution you can get that even uses the FloEFD/SWFS meshing and solver technology.
If you are interested then you can always have a try of it by contacting one of the Mentor offices or resellers, depending on where you are based. It is like test driving a car. You don't have to buy it but if it helps to do you job more accurate and better then it is worth a try.

Boris

[CFDfan]

Thank you Boris for your explanations.

I talked about XT among my colleagues and it turned out one of them has evaluated it for about a month.
He confirmed its suitability for electronic cooling application , the smart part library and the tight integration with Solidworks (SW), which was great. The things he didn't like was that:
1. XT couldn't read native SW files created by later versions of SW (than the one XT was integrated with).
2. The meshing in XT he said was somewhat different than in SWFS. He run the model he had with the 3 different meshing levels in XT and got quite different results. This was not so prominent in SWFS, where the difference in the results when running (the same) model with mesh resolution factor of 3 and say 8 was just a couple of %. To get the same results in XT he had to run his model with the finest level of meshing there, but then the number of cells was about 30% higher than in SWFS' model and the simulation run slower.
In summary he had mixed feelings about XT - very good idea but not refined yet.

[Boris_M]

Hi CFDfan,

Yes, XT is always a little behind the SW version as SW releases it first and the next release of XT usually gets the next SW version as well.
The same is with FloEFD and SW. FloEFD releases new capabilities and they are not or not all of them in SWFS at the same time.

The meshing technology to be precise is the same but the way to mesh the models is different as it is tuned to the way FloTHERM meshes and that is based on smart parts. So each smart part will have a local mesh setting available and of course the way to specify the mesh by the look of the GUI is different. It simply has a different menu but the mesh technology is the same.

It depends on when he tested it. There has been some changes since the very first release. As in every product, the first release is the first release and never the best as the next is always better. It is always a convergence case. The older a product grows the more optimized it is.
As for the mesh settings I cannot tell what he did. Maybe he was used to the SWFS meshing and did it the same way in XT and here it simple has to be done differently. In such case it is best to talk with the engineers of the reseller or Mentor Graphics directly to get their input.
I've seen users of traditional CFD tools start using FloEFD and they tried to create meshes as they know it from the other tools. You had to teach them that this is not necessary as the technology is different. It's like learning to drive stick if you are used to automatic :-) (or the other way around in terms of automatic meshing)
Often you have to learn things new in a new product and forget about the old way of doing it. It will always frustrate you because you'll say "But I'm used to do it this way." You don't call your new girlfriend by the name of the old one because you are used to using that name for your girlfriends, right :-)

I mean you can certainly try it on your own to see how it is and make your own opinion of it.

Boris

[CFDfan]

Boris a continuation of this interesting topic.
I read a promotional material about the SmartCels in Floefd that allow for good accuracy with less cells and shorter calculation time. This sounds like a very good idea and my question is:


1. Starting from which version of Floefd will/are these "smartcells" be available?

2. Will the SWFS module in Solidworks 2017 be supporting the smartcells and if not are there plans to introduce them in SWFS as well and when?

[Cengizhan Arslan]

Hi CFDfan,

It's a marketing decision, they just changed the name

[Boris_M]

Hi CFDfan,

The SmartCells are the combining technology of the formerly called partial cells and the solver technology that makes them work with FloEFD’s meshing technology, i.e. working without body fitted meshes and still get that good results.
The white paper you read eaxplains the technology a little better than the meshing and turbulence paper you can find on the Mentor Graphics website.

But it is nice to see that this technology seems to be so exciting that you want to have it in SWFS as well :-)
However, you have been using it already all the time, this just shows how goodthe technology is :-)

Boris

[Boris_M]

Hi Cengizhan,

Not quite true, it was a technical decision to give the "partial cells" a technical terminology that is better understood than its previous name and with that the need for a better explanation of the technology was required. But since it is not just the mesh but also the solver technology that make it work (the combination of both), this had to be shown and so the name SmartCells was given that describes the combined technology. Partial Cells were only the cells at the fluid/solid boundary called.

Regards,
Boris

[Cengizhan Arslan]

Hi Boris,

What i was trying to say is that instead of smart cell, it could be intelligent or clever or wise or "geometry and solution adaptive, helping the solver with boundary layer treatment 'two-scales wall functions model' with just mesh number at boundary layer and whether it could be laminar-turbulence-transient flow; newtonian non-newtonian fluid or not etc.., you don't have to worry about the type of it because FloEFD only uses cartesian based, includes boundary layer" mesh. Well, maybe the last one is a little bit longer

But of course you are right, it's very different from other meshing technology like cut cell although they like similar to each other and like you said not just meshing, it's combined technology with FloEFD solver. Maybe solver needs a special name either

Regards,
Cengizhan

[Boris_M]

Hi Cengizhan,

Yes, exactly the long term was the reason for a good short name that basically summarizes the technology.
And SmartCell includes the solver technology that applies for those cells that make them work. The rest of the solver to the far field is not much different to other solvers, just the way to handle the boundary layer and the type of mesh the way it is and provide good results is what is the big difference and that is part of the SmartCell naming.

Happy holidays to everyone,
Boris

[DB99]

Don't know if this thread is still monitored due to being started in 2016. I use both XT and EFD and have used FloTHERM "classic" for many years.
FloTHERM XT vs FloEFd. ECAD import in XT automatically produces PCB Smart Part with three level options of detail. FloEFD does not. Changing level of PCB detail in EFD is a manual operation. XT has more bells and whistles for specialty electronic functions. Though on that note I would say that EFD is more receptive to improving product features where XT appears to be more stagnant in this area.
There are many idfference too many to go into open ended. FloTHERM classic does not import CAD geometry (tries to but way insufficient). XT does.
History......Classic and EFD separate products. EFD and classic (Flomerics Inc.) bought by Mentor Graphics. XT is porting of Classic functionality to EFD solver and geometry processor with new XT GUI. EFD accepts CAD data and meshes it completely differently than FT Classic by use of advanced partial cell technology. EFD has Electronic Module which is subset of XT and Classic funtionality. EFD was originally a flow solver and has evolved over time. They are currently still separate branches of now Siemens whom bought Mentor Graphics. Over time I suspect they will merge but until EFD raises to the level of XT and Classic FT it cannot replace them in terms of electronic modeling functionality. EFD has many nice features but also many productivity shortcomings when it comes to Electronics modeling.
Hope that helps if you're still listening.

[Boris_M]

Hi DB,

Yes, this is partially correct.
FloEFD has also the EDA Bridge Module which is about the same as in FloTHERM with which you can import also three levels of detail from the EDA data such as ODB++ files.
But you are correct, FloEFD doesn't create smart parts. It's aligned with the CAD modeling approach as it is designed for CAD users originally. So a feature-based modeling approach.

Where in electronics you usually pick and place existing components and therefore a Smart Part approach with a good library is key, in CAD modeling most of the models are totally new and maybe use standard parts such as screws and washers etc. but those are in libraries of the CAD system already. So the smart part approach is not the typical approach for thermal simulation.

In terms of capabilities, it purely depends if you need specific features or not. Often it is a lot of user preference as well in the way to handle the simulation setup. Some might need a specific capability others are not, some might like the handling of the software, others might not.
Therefore I suggest always to try the tools as it needs to fit your preference as well as needs to provide the capabilities you require to solve your problem.

Yes, FloEFD does not have a pure focus on electronics cooling as it is a multi-purpose CFD tool and therefore has many features that need enhancements. In V18 we got new radiation capabilities, new electronics cooling capabilities, brand new battery cooling models etc. So a wide spectrum of enhancements whereas application or industry-specific software like FloTHERM can focus on electronics cooling focused features purely and have therefore more capabilities to these applications.
FloEFD will get more electronics cooling capabilities as well of course but it won't be the focus such as with FloTHERM. In V18, for example, we released also the calibration capabilities that uses the T3Ster measurements to build better thermal models of semiconductors which was already available in FloTHERM for 2 years or so. And the same the other way around. FloTHERM XT received sliding mesh after it was in FloEFD for 1-2 years already. Often the features have to be validated in the field first as well as the GUI adopted as they are not the same and some features will probably never come to the other tool as they are simply not needed or never requested by the users.

Regards,
Boris