I’ve done the section print and the results are very close to what I got in RFL (I’m interested in 3 surfaces)
What I don’t get is that the section print is negative, which indicates that the surface is gaining heat according to your description of the current convention? But it should be losing heat.
I’m also getting this error, but it’s displaying the results just fine.
I would choose fewer fasteners because each fastening point involves construction work, which translates into costs. Each fastening point, for example, to the ceiling, creates a thermal bridge and temperature drop at the point of penetration of the thermal insulation. Without fasteners, complex 2D and 3D calculations wouldn’t be necessary. Simply calculate the total thermal resistance R and the U-value. If possible, stainless steel is used instead of aluminum. This is obvious knowledge, the basics of building physics and economics. Reality is more difficult, which is why calculations like yours are sometimes performed. And due to lack of funds, the cheapest aluminum solutions have to be used. The standard “EN ISO 10211:2017 Thermal bridges in building construction - heat flows and surface temperatures - detailed calculations” helps in making design decisions and standardizes calculation principles. I’m glad you’re doing this, as there aren’t many such calculations. Thank you for your patience and I apologize for any translation errors. English is not my native language.
See again with the uninformed assumptions. This isn’t as simple as picking a colour for a new car you’re about to buy. Each supported part has a load, and it needs a minimum number of fasteners. The design comes from a manufacturer of thermally-isolated cladding support systems. And if they could get away with fewer fasteners they would because it would make their product perform better and also cheaper.
I’d really like to know your profession/background at this point.
The last two sentences are incorrect. Fasteners represent a point thermal bridge (Chi value). But there are also linear thermal bridges (Psi value). Both are assessed using 2D/3D FEM (if you want to be accurate). Even without cladding support system, FEM also helps with an accurate wall U-value calculation if you any penetrating element or an exposed slab. In this case, the FEM-calculated U-value can be up to twice that calculated manually (I’ve tested this).
Things aren’t that simple and ‘obvious’. If you want to reduce thermal bridges, then stainless steel is better if both designs are identical. But if you’ve already provided a thermal ‘break’, then you don’t need the low conductivity of stainless steel, which comes at a premium cost. You can get a Passivhaus-cetified window that’s got an aluminium frame. How? It’s got a thermal break.
You need to change your mentality when it comes to making uninformed comments especially if these comments turn out to be incorrect. And you also need to stop presenting things as ‘obvious’ when in reality things are a little more complex
Yeah, the CalculiX User’s Manual states:
positive = flux leaving the element through the surface defined by the parameter SURFACE
https://www.dhondt.de/ccx_2.23.pdf#subsection.7.115
I made a simple test - cube with 10 degC temperature on one side and 100 degC on the other side. I got positive surface flux for the side with lower temperature and negative for the side with higher temperature.
Then I made the same test in Abaqus and also got positive flux for the side with lower temperature and negative for the side with higher temperature.
I’ve also tried it with two convective fluxes (the same coefficient, different sink temperatures: 10 and 100 degC like before). And got equivalent results (surface with 100 degC has negative flux).
This is expected - section print in PrePoMax doesn’t support heat flux yet, so it can’t read these values from the .dat file. I’ve tried implementing it and failed exactly on this part related to .dat file parsing.
I just wrote to you about 2D and 3D thermal bridges. Much smaller models are made for this purpose, in accordance with EN ISO 10211.
And often in 2D or using some other simplifications:
There is some dedicated software for that.
So what’s the explanation? is this a bug? Heat flows from hot to cold – surface with lower temperature (gaining heat) should be negative and vice versa.
I don’t really get what we’re arguing over, here. I doubt anyone’s heard of 10211 before my post on validation, and all of a sudden it’s treated like a fundamental law of physics. It’s a guide, and I’ve already explained why I did the large section and not a smaller one.
Again, unless you’ve done both 2D and 3D, you’re not really in a position to judge. That post is 6 years old. I’ve done both 2D and 3D and there can be significant differences in heat flow. Once you’re proficient in FreeCAD/meshing, you really aren’t going to save an awful lot of time doing 2D, especially with the potential errors that come with it. CAD capabilities in THERM are awful, for instance. I can’t even begin to imagine how I’d model something like this in THERM
The video is about SimScale – so what? I’ve seen that video before I found PrePoMax. As we’ve seen in the validation post that I initiated, PrePoMax is ‘valid’ for thermal bridge calculations. Why do I need to learn another FEM software when both are equally valid? And one is for free and the other is $5,000 a year.
Section print bases the sign convention on the flux leaving or entering the elements underlying the selected surface. Since heat goes from hot to cold, it will enter the cube through the surface with higher temperature and go towards the surface with lower temperature, leaving the domain there. So the signs I got in CalculiX and Abaqus make sense.
That’s not what I meant. I just shared some examples (doesn’t matter what software they use) to support the discussion (mostly initiated/continued by MisiaKu) about modeling assumptions and simplifications for thermal bridge analyses. And to show the approaches used in practice regardless of the software. Also for other users reading this thread and not familiar with such studies.
That graph really explains the concept behind the convention quite well – I just have to think about it differently. Thanks @FEAnalyst!
Sure, at least this is how I understand it based on the documentation. I made the test to confirm this and to see if it really works as stated (CalculiX manual is sometimes outdated and such conventions may change, especially if they differ from other conventions in the solver).
In this forum, apart from the EN ISO 10211 standard, there is no validation according to EN 10077-2. There is also no EN ISO 13788. To check the marking convention, it is enough to create a simple model of cubes in which heat flows in one direction:
Anyone who does a lot of these calculations probably knows this. Just as the PSI coefficient can be positive or negative.
THERM is widely used in the construction industry to verify the thermal transmittance of building envelopes and determine the minimum temperature. It is officially recognized by the National Fenestration Rating Council for performing simulations required for energy certification of windows in the US.
You are showing the Query tool which averages/sums the field output. The discussion about sign convention was related to *Section print (and it’s already resolved).
Regarding CalculiX validation against various standards, I would suggest CalculiX forum or GitHub for that since we are mostly validating the solver rather than the pre-/postprocessor. There is one thread for that already: Validation against ISO 10211-2017 - Examples/Tutorials - CalculiX (official versions are on www.calculix.de, the official GitHub repository is at https://github.com/Dhondtguido/CalculiX).
