Permissible Stress or Limit State Design?
Hopefully this blog won’t be as boring as the title suggests?….
Even though it was a very long time ago….. I remember being at university and learning about permissible stress design (it obviously wasn’t a Thursday morning). At the end of the lecture a wise old professor informed us all that we’d never actually design anything this way, that it was pretty much obsolete because everything was moving well and truly towards the future…. and the limit state approach. Indeed, when I worked in industry prior to joining the military I found this prediction to be true, and the idea was reinforced during Phase 1!
Imagine my surprise when I arrived in the US and found that, rather than being a bastion of enterprise and advancement, the US actually embraces and encourages the two design philosophies to be used in parallel! Over here they are known as ‘Allowable Stress Design’ (permissible stress), and ‘Load Resistance Factor Design’ (limit state). When I say ‘encourages’, this isn’t quite fair. What I mean is that the American Institute for Steel Construction (AISC) publishes both methods side by side in its design manuals, allowing designers to pick which method they prefer. Interestingly, concrete did make the jump years ago, and no longer practices ASD.
So I’m currently designing a glorified giant dog kennel and training facility for some fierce secret service working dogs. I’m working to ASD for the steel and masonry, which is nice because it makes things simpler and I don’t have to worry about any complicated plastic analysis or behavior etc…. it’s also handy because all of the joists come from the manufacturer designed and specified using ASD. Nice and simple. Until it comes to designing the footings, which of course being concrete mean that I had to alter my loads using a crude conversion factor of 1.4 to ‘upgrade’ them from ASD service loads to LRFD design loads! But even then the fun isn’t over; the soil bearing capacity (4000psf John, very stiff!) isn’t factored, so I need to multiply this by an additional ‘resistance factor’ so I can work in LRFD, making sure I don’t accidentally divide by the ASD global ‘safety factor’ instead, because that would be a disaster! Who needs a sinking building and (expensive) squashed mutts!?
My Design: Permissible Stress or Limit State? Why choose, have both!
Apparently things are improving! Until recently live load and dead load factor of safeties for steel and concrete construction using LRFD where different! 1.6 & 1.2 for steel and 1.7 & 1.4 for concrete; imagine having to swap between the two sets whilst working with the same philosophy on the same design! You all thought Euro-codes were frustrating, you’ve got it easy! Also, (and back to the point) the AISC are apparently considering removing ASD from their manuals which will be a step in the right direction! However I’m not holding my breath; it takes a long time to buy a stamp over here, imagine how long it might take to alter something as ingrained and controversial as this.
This all seems like madness and an accident waiting to happen… and it nearly it! During my time on the JOC project over at East Campus it was discovered that a series of giant steel trusses had been sub-contracted out to a ‘specialist’ by the designer. When they were finally delivered to site it was discovered that they, together with all the connections had been designed using ASD rather than LRFD loads. On checking the truss turned out to be fine (phew, almost a very expensive error for someone), but all the connections were under-designed by approximately 40%! These trusses were very nearly installed, and it was only scrutiny of the drawings at the beginning of the week of installation that prevented fundamentally unsafe construction! Whilst it is considered bad practice to mix design philosophies, in my experience it seems to happen an awful lot. Particularly on simpler, small projects that need to be turned out quickly, and by older-generation engineers who grew up with ASD and see no reason to change.
I was wondering if anyone else had had any experience of permissible stress design during their attachments? Surely if there’s a slow backwards nation that can stand with America on this one it’s Australia?! Or is it really only America that lacks the will to embrace change?
Holdfast Training Services Trial Blog 
No Kukie, having spoken to my boss they don’t use permissable stress in Oz.
I’ve not seen this to the same extent. There is some limited mixing of BS and EC but both limit state although with different factors which can be annoying. Not really an issue, just need to be aware of it.
Some of the initial brainstorming and scoping is done by the old big dogs in permissible stress terms as its what they know.
Would some of the initial member selection tables be to BS as well?
On your truss issue i’d recommend you see my steel connections blog and then check out NSSSBC which covers the information that should have been communicated to the ‘specialists’. Sounds like a coordination error.
Well this is a huge issue
The truss problem you refer to happened while Brad Southall was there and I had him do a TMR on it. So maybe we start there. I saw the letters that went between contractor and steelwork sub contractor Permissible’ loads were quoted on the design drawing s yet the contract said all was LRFD so the loads were undercooked ( roughly by an action FoS) What I ended up pointing out was that, since it was a roof, the main risk was the variability of the variable action and therefore a higher adjustment was needed than the 40% they were proposing.
WHy?……….answers on a postcard but the answer is in your message.
There was a fundamental dumbness about what was being done as well as the exposure of the problem of co-existence that you highlight
What it always comes down to is a basic understanding of why limit state codes exist.
There are 2 BIG reasons ( and lots of little ones)
BIG 1: If we take an engineering problem to a limit state it don’t half make the analysis easy – think about LS reinforced concrete stress blocks )plastic) as opposed to the stress block that might exist on the way to the LS – clear?
BIG 2 There are three risks involved in verification calculation 1= we don’t know the actions 2= we don’t really know the material strength 3= we don’t really know how reliable the engineering model for the limit state is( stuff like how straight is a straight column ( see NSS) and does the Perry Robinson formulae properly assess its behaviour. Now, I’ve never really bottomed how we juggle with (3) but we do the other two by partial FoSs – clear?
What LS was meant to liberate is the management of risk- for example suppose we absolutely KNOW the actions on a problem . I find this is so in temporary works more often than not THEN why would we apply reliability factors that imply the structure ( which is temporary) may be receiving a load spectrum that might be associated with varying use over a long design life? That’s a bit rhetorical but what you will find is that temporary works designs slap all sort s of Fos on ( what are usually permissible stress) approaches- ie FoS on in built FoS – and say ‘ well anything could happen of site…errr… isn’t it that you actually have some supervision then?
This issue infects the teaching here. The CoWs are taught structures from PAM10. It is old, old limit state. Nothing wrong with that. But if a narrative were to exist that in circumstances where risk was an external threat how could a design be adapted to minimize exposure; then you’d down the partial; factors to a min and say that you’d responded to the external threat THAT WAS WHAT THE LS CODES WERE FOR. ( you may recall the GDR are go through just before ExCoffer- or you may not !
So to geo…..you may be getting tired now but stick with it
What I hate about BS EN is that the ideas of LS are just sinking in complexity. IF you look at pile design I can barely explain how the FoS’s are handling risk for the designer. IF I look at Australian pile design it is gorgeous…the FoSs use are attached to real risks and you can manipulate them by selecting in or out risks in design and execution…they are not hidden to all but those ‘in the know’
And finally to 4000psf…. nothing new or old to me this is what we used to call 2T/sq foot or about 200kPa
Stern face on
1 This is a strength NOT, NOT DO YOU HEAR ME? a stiffness!
2 This should like a permissible stress and that often stands in for the fact that the problem might be settlement but we’ve not a Scooby as to what the soil stiffness is – SO we take a (net) permissible stress at about one third of the ULS ( bearing capacity failure – a shear strength failure) and say ( to ourselves, as they cart us away for talking to ourselves) that , irrespective of stiffness , this will keep the associated strain ( settlement) low
3 So it looks to me like you’ve ‘got it up around your adam’s ‘ Just to confirm q(al)l of 2T/sqft would be a decent enough sand . To compare the ULS you’ve multiply this by 3 and add the overburden stress at foundation level and compare with the base stress form the factored actions in BS EN world you also take unfactored actions ( bar 1.3 on the variable) and compare to a bearing capacity of a weakened soil – so I’d do this to a qult ( from above )divided by about 1.25
That’s all folks
Sorry PAM 10 should say ‘old,old permissible stress’
Kukie
I am looking at an existing historical structure (an old dry dock) and trying to work out the original design logic and using a bit of permissible stress.
That’s it really – not using permissible stress to design a structure but trying to fathom how the existing structure works when “modern” design methodologies calculate the thing should fail…
In my particular case I find myself questioning Whether limit state really is a step forward…