Because topic “Measure load-bearing capacity in SU” went way off topic and became such a mess, …
(9 out of the 11 replies in the old thread were off-topic,) … here’s the same but all on topic posts …
A full 9 out of the 11 replies in the old thread were off-topic !
The previous thread was such a mess, that the Moderator declined to spend any time fixing it.
The OP appears to be about building structural analysis into SU. I see several problems with that despite it being potentially a very useful feature.
Firstly, SU seems to have been developed as lean software but with the ability to add “fat” (functionality) through extensions. So it seems improbable that something as specialized as structural analysis would be part of the main program. But even as an extension, SU would have to have certain things built in, such as an “understanding” of materials and their mass.
Secondly, structural analysis is highly complex and needs an understanding not only of loads but of load paths, bending capability, formulae, etc. It’s just about possible to imagine an extension that dealt with, say, simple reactions for a lintel or floor beam, but the OP’s structure looks as if it would test even the most experienced engineer!
A third problem might be universality. I expect American Building Codes differ from the equivalents in other countries. I am not a structural engineer but I can imagine that things like safety factors may differ from country to country and from structure type to structure type.
Maybe this is a challenge rather like designing driverless cars. You can get so far and everything can look encouraging until one day, mysteriously, the car kills someone and no one knows why.
Thanks. You might as well check the original entry here: Measure load-bearing capacity in SU - #14 by endlessfix since there are more answers and scenarios there in the original posting.
One would think, however, that some sort of “Materials” A.I. would be able to calculate load variables for a given span in 1G environments. That is, if a beam made of X (concrete/steel, whatever) is stretched over Y distance, what is the Z weight it can support before bending? I am simplifying greatly, but only in degree, I think, and that’s what powerful computers are for.
Having a powerfull computer is not sufficient. Combined with the rigth software, calculating (existing) or traditional building spans, loads and bearings have just come to the same level of the imagination of engineers and carpenters, eg. if you are experienced, you won’t be needing a powerfull computer at all. With new techniques, there are also new things no one can imagine how the loads and bearings have to be calculated for one material, let alone all the different combination of materials. It is why there is special FEA software. There are educated engineers that program for a living, They might not have the most powerfull computers, but the right software and, they know how to use it!
I think the degree of simplification here is pretty extreme. Simplified load tables for steel ibeam spans are available online. But for even relatively simple composit structures or frames, it’s garbage in garbage out unless you are accurately accounting for the many many variables of building in the real world. Stress analysis is tricky business on buildings.
I’m not trying to be negative about your project, go for it by all means! Just suggesting that looking to prove theoretical “bulidabiliy” at this stage of the development might be difficult, or at the least require some high level expert input.
Where this idea might have legs is at initial concept stage. Being able to know the rough size of beam (say) for a given span would be useful even if it has to be properly calculated and “proved” later on. Even so, some knowledge of structural engineering would be needed that the software cannot easily encompass.
Take one of the simplest task an engineer might face: a beam over an opening. First you need to assess what kind of loads are being supported. Is it universally distributed or are there any point loads? Are the ends fixed or simply pinned? Is the beam just taking the weight of the triangle of wall above it or are there floor or roof loads imposed as well? And so on. Knowing the span, the material, and the profile of the beam, is only the starting point.
So how could it work? Well, if you are able to input certain load types and a rough idea of loading, you might be offered a selection of steel beams of different sections from which to choose. To be useful, it would probably have to err on the cautious side so that when proper calculations are done, the likelihood is that you will be able to use a slightly smaller/lighter beam.
What I was hoping for - in my utopian wish list - is that a certain type of beam X could be given some initial characteristics and then, no matter what length or width it is scaled to later, would keep those material characteristics on a per inch (or less) basis. My columns are 4’ X 6’. If I, say, specify them to be some choice of industry grade reinforced concrete, the characteristics of that kind of column should remain constant regardless of size throughout the structure. Then, other beams and columns laid on top and around should produce stress in calculatable ways, perhaps showing up as red Xs where it is too much for the span.
This is not totally off-the-wall, pardon the pun. There is an Engineering Toolbox extension for SU:
Oh…hang on. Here is something I hadn’t seen before:
These 6 links at the bottom actually go out to a sponsored site for structural engineering! Maybe that’s where to start. Thanks for spurring me on to look closer at this. This may be the answer.
I am NOT a structural engineer, nor do I claim to be. However, I did spend a lot of time in the last 25 years working closely with structural engineers. I can say that what is proposed here is a big deal. Designing a single member spanning a single distance is something that can be automated, pretty easily. This, however, is a group of members bending out of plane and working together to oppose forces. This is quite a bit of work for a super smart brain.
My Engineer partner John McClenahan and I decided early on not to bend the beams, but to just attach them to connecting joints at slight angles to make the overall arch. Bending beams is very difficult and costly, and would have made it harder to create individual floor layouts; it’s hard enough with slanting beams.
I reached out to the Structuretech Engineering people.
