Industrial trusses

[oldkettle]

Just curious 

 

Here is Heathrow terminal 5 roof structure. These triangular trusses are nothing new. I wonder whether they can handle a higher load in this position or when 2 tubes are at the bottom, one at the top - does anybody know a good calculator for this type? In which case - why are these not used in residential buildings? I would happily swap 300mm 8.2m beam for one of these. 

 

https://as1.ftcdn.net/v2/jpg/04/42/41/30/1000_F_442413078_THZGOCx6OOJlTlti0NAhcdL1MY5fGna7.jpg

[George]

Double chord at the top because the chord in compression is liable to buckle, whereas a single large tension chord won't buckle. 

 

These are massive and as you reduce the size the relative cost of fabrication increases. A truss is used because the additional fabrication cost is worth it for the reduced material and weight saving (and some architectural value too). You wouldn't change a 300mm deep beam for a truss because the fabrication cost would many multiples of the cost of the raw steel.

 

Many years ago when steel/iron was costly and labour was dirt cheap, you sometimes found small trusses. But aside from prefabricated things like metal web joists and roof trusses you won't find fabricated steel trusses at a domestic scale.

[Alan Ambrose]

These are space frame trusses - you do see smaller ones being used e.g. for lighting rigs.

 

Maybe the extra cost is justified in that case by the truss being lighter and thus easier to erect for temporary installations.

 

e.g.

 

https://www.ltt-versand.de/fr/structures-alu/naxpro-truss-black-label/fd-24-noir/24829/naxpro-truss-fd-24-longueur-300-cm-noir

[Gus Potter]

@George, elegant and practical explanation.

 

As an aside a lecture of mine made his own Glulam beams, think they were curved. He was a keen woodworker and would get big hardwood logs off a Tree Surgeon. He a had one of these big chain saws that ran on  a track and would saw them into planks, season them all and so on. I remember he wanted to do a contemporary type oak frame with these curved beams and had got a frightening quote.

 

I would give it a go myself if I had the tooling, space, time and importantly have a use for them.

 

I think there are about three ways off the top of my head of doing it.

 

1/ Set up a jig and dry fit everything using a long furniture clamps. Pilot drill and counter sink each plank. Then mix up lots of structural glue and put some overalls an a hat on. Lay down the first plank and hold in place. Next secret screw each plank to the next using plenty glue. The screws provide the compression to the glued surfaces. Scrape of the worst of the overspill, clean yourself up and wait. Once all has taken up get the belt sander / electric plane out and start finishing it off.

 

2/ Dry fit as above but pilot drill each plank as you go for an M12 / M16 rod say spaced at 400 - 600 centres at least the overall final depth of the beam. Here I think you need to use an oversize pilot hole to cope with the curvature , that you won't get right first time.

 

Start fitting it all together. Maybe do 4 planks one day then another set the next day and use the rods and big washers to clamp them.

 

3/ Go for the more industrial look where you use a steel plate top and bottom to act like a big long washer and try and turn the beam into a steel / timber composite section. You can get standard stock flat plates at 6.0 - 6.1 m so for a long span beam you would need to couple the plates if going for long lengths.

 

12 hours ago, George said:

you won't find fabricated steel trusses at a domestic scale.

Could be a job @Onoff would tackle as he has the cad and fabrication skills? Technically things need to be CE marked for structural steels but as this a bespoke item then there could be room for manoeuvre to deal with the paperwork.

 

 

 

[joe90]

I made wooden springs for a camper conversion bed years ago, each was 3” wide 3mm ply strips x5 glued together and clamped to a curved former, worked brilliantly. I recently did a factory tour with my son to the Morgan factory where they still use the same former and clamps over 60 years old to make the rear wheel arch (I can highly recommend visiting the Morgan factory).

[Onoff]

10 hours ago, Gus Potter said:

Could be a job @Onoff would tackle as he has the cad and fabrication skills? Technically things need to be CE marked for structural steels but as this a bespoke item then there could be room for manoeuvre to deal with the paperwork.

 

I've stacks of 25mm box section, circa 3.5m lengths. Did toy with the idea of making up welded lattice beams for garden building roofs. I could make it but the structural calcs aren't my thing  

[oldkettle]

On 18/12/2023 at 08:45, George said:

These are massive and as you reduce the size the relative cost of fabrication increases. A truss is used because the additional fabrication cost is worth it for the reduced material and weight saving (and some architectural value too). You wouldn't change a 300mm deep beam for a truss because the fabrication cost would many multiples of the cost of the raw steel.

 

Many years ago when steel/iron was costly and labour was dirt cheap, you sometimes found small trusses. But aside from prefabricated things like metal web joists and roof trusses you won't find fabricated steel trusses at a domestic scale.

 

Thank you. Below is a very naive take, please correct my mistakes if you have time.

I only consider a straight one, not curved, and one beam would contain 3 pretty standard long tubes plus a few dozen connectors which should be pretty standard as well. I don't know whether there is a way to assemble the whole thing using bolts (say, if connections are top/bottom of the tubes) or they have to be welded, in either case doesn't sound terribly time consuming (hence expensive).

The benefits are 

Easy to handle - no lifting equipment required. 

Easy to deal with services (the same as web joists).

Can be left exposed and would look OK as opposed to a steel beam. 

I would consider paying extra for these benefits - obviously, depending on the difference. 

 

I saw a large personal pool where the roof was supported by cables - I guess they also protect against the wall spreading. Looks great. 

[George]

I don't know whether there is a way to assemble the whole thing using bolts

You could certainly have that designed but it wouldn't be a straightforward design and as you get smaller the influence of bolt holes removing material becomes more significant. 

I charge maybe £300 for a beam design... because I don't do it very often I'd probably charge £900 for a truss! Unless I could push almost everything onto the fabricator.

 

or they have to be welded, in either case doesn't sound terribly time consuming

You need to consider than a standard beam size required very little fabrication time and labour is the key cost. So with a standard beam you are mostly paying for materials. A truss only has a moderate reduction in material cost but many more times more fabrication. Each member will need cutting and prepping, lots more welding, flipping it over, weld the other side etc. 

 

Easy to handle - no lifting equipment required. 

A beam lifter is maybe £50 for a day. It's a minor thing. If it's prefabricated then no benefit.

 

Easy to deal with services (the same as web joists).

True but this doesn't really save any significant money. 

 

Can be left exposed and would look OK as opposed to a steel beam. 

A nice beam doesn't look too bad. A bolted truss would look a bit shoddy. A nicely welded truss could look good but a bit of a dust trap! 

Also need to consider fire proofing. 

 

 

It's not outside the realms of possibility. I'd say probably between x3 - x8 the cost for a standard beam. 

[oldkettle]

44 minutes ago, George said:

Unless I could push almost everything onto the fabricator.

This is it: I would expect this to be a standard detail. "Here is your 10m long 400mm truss supporting usual loads, if you need more it's X to calculate. I mean no disrespect to the profession at all because there are plenty of complex scenarios but this seems to be something AI should be able to handle easily. Yesterday. 

 

49 minutes ago, George said:

You could certainly have that designed but it wouldn't be a straightforward design and as you get smaller the influence of bolt holes removing material becomes more significant. 

I just tried to check whether rivets were an option as well 🤣 Again just curious where the balance is wrt manufacturing costs. 

 

1 hour ago, George said:

A beam lifter is maybe £50 for a day. It's a minor thing. If it's prefabricated then no benefit.

Yeah, I was rather thinking about loading / unloading and moving 1t 9m beam 50m up the garden around the house. I know it is doable but wouldn't it be nice to have something much lighter 🙂

 

1 hour ago, George said:

Easy to deal with services (the same as web joists).

True but this doesn't really save any significant money. 

Well, it saves having to drill through or losing space in a false ceiling. For example, right now I am not even sure how to get MVHR ducting from one side of the house to the other. Like in manufacturing, ultimately it saves time. 

 

1 hour ago, George said:

I'd say probably between x3 - x8 the cost for a standard beam. 

Wow, that's quite a lot more expensive then, 1K for 305mm I-beam vs 3K+ at best. No surprise few would suggest / consider this option.

[George]

Currently available AI are language models and don't seem to be able to reliably handle mathematics.

 

There's an ongoing thread on an engineering forum I'm on where we keep testing Bard, ChatGPT etc and it produces credible looking but mathematically incorrect answers to structural engineering (or complex maths/physics) questions. Entertainly on Bard you can switch between drafts and it sometimes give different answers in each. I don't fear for my job yet! 

 

There are standard designs of course and I wouldn't be surprised if someone out there has tried to market a truss beam replacement. But I've only ever worked on large trusses. I did see a footbridge made up of space frame trusses but that was in an awkward location so could be brought in and bolted together, making it worthwhile. And a footbridge is a relatively lightweight use case with repeatable applications.

 

Structural steelwork rivets are obsolete in the UK. They're outside the scope of modern structural codes so wouldn't be acceptable to building control without a lot of effort.

 

Drilling a hole in a steel beam, even if it required reinforcement, would be a lot cheaper than a truss. You should be able to get a 75mm hole (for radial ductwork) through a 300mm deep beam but would need local strengthening. If it's a main duct then could use rectangular ducting to get through . Even if you didn't have a fully coordinated design you could get a few designed holes put in.

 

 

 

 

 

 

 

Edited December 20, 2023 by George

[oldkettle]

13 hours ago, George said:

Currently available AI are language models and don't seem to be able to reliably handle mathematics.

 

They are getting better at applying standard formulae and especially when using "tools", such as various calculation engines. I played with chatgpt to see what size glulam would do the job while SE was working on his I-beam calcs and with some supervision - stopping it from forgetting inputs etc. - I got numbers closely resembling load tables with supporting formulae confirming the deflection. Again, this is some way away from being usable. On the other hand, while searching for options yesterday I came across a picture of an existing specialised calculator for this specific beam type so the problem is clearly resolved for standard use cases. 

 

13 hours ago, George said:

Drilling a hole in a steel beam, even if it required reinforcement, would be a lot cheaper than a truss. You should be able to get a 75mm hole (for radial ductwork) through a 300mm deep beam but would need local strengthening

 

Yeah, I know it's doable, but does require a commitment to specific locations of the ducts, something I am clearly not good at 🙂

[SteamyTea]

On 18/12/2023 at 21:22, Gus Potter said:

 

I would give it a go myself if I had the tooling, space, time and importantly have a use for them

Just do it for a laugh. What I do.

 

 

 

[SteamyTea]

1 minute ago, oldkettle said:

while searching for options yesterday I came across a picture of an existing specialised calculator for this specific beam type

Care to share the link?

[oldkettle]

3 hours ago, SteamyTea said:

Care to share the link?

 

I used Incognito view, sorry, can't look at the history. But I did visit https://www.steelconstruction.info/Modelling_and_analysis#Trusses_and_lattice_girders which refers to an application called Trimble (if I am not mistaken). 

Also once I knew the names for these trusses like Warren or Pratt it was easy to find calculators, i.e. https://skyciv.com/free-truss-calculator/

[George]

Oh we're all over the modelling software. I used a cousin of Fastrack for most of my day to day design work. The calculation of forces in an isolated structural system such as a truss isn't the hard part any more (although it can be time consuming and is now rarely done by hand)... what the engineer needs to know is what forces to apply to the model, in what combination and what the output means and whether it is acceptable or not. As yet you can't show a AI a house and get it to reliably work that out. It probably could one day but I won't hold my breath. 

[Gus Potter]

23 hours ago, George said:

There's an ongoing thread on an engineering forum I'm on where we keep testing Bard, ChatGPT etc and it produces credible looking but mathematically incorrect answers

Hi George.

 

Can you let me know any more about this?

 

For all.. when you get into truss design it becomes a very complex animal. Even more complex when you start to use cold formed cee sections. The nature of trusses often suggests long spans. If you have a truss with bolted  connections.. these connections can often slip as they tend to work on a 2.0mm oversized holes and in shear rather than tension and shear on say a standard steel portal frame. This can lead to unpredictable behavoir. Once we enter this world as SE's we start to worry about things collapsing with no warning. Buckling of the frame, not just the individual bits but .. rafters / column interaction as a whole becomes a big issue. We call this in plane buckling where the whole structural frame buckles and falls down suddenly.

 

Now this may seem like me being a bit geeky.. but if you are  converting a portal frame barn shed in England, I think you call this class Q.. you do want your SE to understand what in plane buckling means?  Ask them if they know about it. If they don't then point them to the SCI guidance and start to wonder if they are competant to undertake this kind of design.

 

I have various bits of software to model structures. Some is main stream off the shelf that all SE's use. Some is more bespoke that I use when analysing complex problems. I have some pretty high end stuff for FE modelling, and cold formed steel design, don't often use it these days as my current Client base can't afford to pay for the analysis. It can take at least a week to set up a good FE model that you can present for other Engineers to verify. But always I do some hand calculations to make sure any computer model is on the ball park.

 

You NEVER trust what the computer says and you NEVER trust anyone else's computer model, always check by hand to make sure the forces / stresses and material resistances are within the range you expect. A good SE / Architect / QS etc can look at things and say.. that looks odd.. if it does not look right it probably is not.. only a human can do that... smell the fishy stuff.

 

Now AI is miles off being able to replicate what an SE does, Architect's etc are safe too. SE's, Architects spend many years learning how to make things work.. but a big part of the job is being creative and design things that are buildable.. AI just can't do that as it will always be behind the human curve. It won't know how to make things buildable unless we tell it how.. for every job and every different set of circumstances.  Even if we could then market forces.. prices of materials, labour availabilty and local cost. The complexity of programming for a relatively small market I think will make it not viable finacially .. it ain't going to happen any time soon where AI takes over our jobs, mind you it it can make invoicing folk easier.. quite happy with that.

 

 

 

 

 

 

 

 

 

Edited December 21, 2023 by Gus Potter

[Gus Potter]

On 20/12/2023 at 00:45, George said:

There are standard designs of course and I wouldn't be surprised if someone out there has tried to market a truss beam replacement. But I've only ever worked on large trusses. I did see a footbridge made up of space frame trusses but that was in an awkward location so could be brought in and bolted together, making it worthwhile. And a footbridge is a relatively lightweight use case with repeatable applications.

Chewing the fat at my end. I like your train thought George.

 

The big problem with big bolted space frame trusses is the bolt slip as the bolt ends up mostly working in shear and the second order effects that they generate. If you weld them then that problem partly goes away. But the cost rockets.

 

Round about Glasgow they have loads of old bolted trusses, got one of these on my books at the moment.

 

I'm trying to tempt you George into PM'ing me on your research and in return I hope to reciprocate. That is my offer, no point in mucking about.

 

Gus.

 

 

 

 

 

 

 

Edited December 21, 2023 by Gus Potter