Residential Portal Frame Deflection Limit

[Arun238]

Hi there,

Just trying to understand the deflection limit for residential portal frames designed in New Zealand. There is a 8mm or H/300 limit specified for specific bracing elements as per the P21 test paper by Branz.

Referring to 1170.0 serviceability limits, The in plane limit for plaster or gypsum walls in Height / 300 against SLS wind loads with a parameter specified as mid-height deflection.

Does this mean that the SLS seismic limit is 8mm or H/300 at knee and SLS wind limit is Height / 300 measured at mid height for a portal frame.

Thanks in advance.

Arun

 

[Greenalleycat]

That is a confusing criteria that we aren't 100% sure on in the office
Our interpretation is that the 'H/300' they refer to is the mid-height H i.e. for a 2.4m stud wall it's (2.4/2)/300 = 4mm at mid height
I assume the reason is that walls tend to have pretty linear deflection profiles but frames don't - that limit is trying to enforce some implicit consideration for deflection (in)compatibility between walls and frames

As you can see in the super scientific sketch below, peak curvature is through the bottom area up to midheight ish, so I think the limit is trying to stop you jamming too much incompatible deflection in that bottom area

This would be equally applicable for both EQ & wind SLS - the deflection profiles are fundamentally the same
In practice though, I don't know anyone who actually does this check - I've always seen checks for H/300 at the knee or whatever
This is probably not technically right but 99.9% of portals are designed as pin-based and neglect any base fixity so I think it's one of those things that "just works" when designed to standard practice

In general, there is greater indsutry awareness and reluctance for using portals and GIB walls in line due to a) deflection incompatibility and b) ductility incompatibility
GIB is very stiff at low displacements (<2-4mm), very ductile above that, and has pretty good damping
Portals are never that stiff (neglecting a fixed base at least), never really ductile, and have very low damping - but they are very linear in behaviour
There is some industry guidance on this in ENZ's paper and in Angelia Liu's research paper SR337 for BRANZ

https://d39d3mj7qio96p.cloudfront.net/media/documents/SR337_Design_guidance_bracing.pdf

 

[Tomfh]

Our interpretation is that the 'H/300' they refer to is the mid-height H i.e. for a 2.4m stud wall it's (2.4/2)/300 = 4mm at mid height
I assume the reason is that walls tend to have pretty linear deflection profiles but frames don't - that limit is trying to enforce some implicit consideration for deflection (in)compatibility between walls and frames

Isn’t it just this? Ie equivalent to limiting a beam deflection to L/300?

 

[Greenalleycat]

I assume the ?/300 criteria is consistent with a beam, yes - it's a fudge that limits deflection to a level to prevent damage to the lining
Though noting that a beam L/300 is not the same as that's out of plane bending in the plasterboard sheet, whereas this criteria is for in-plane deflection of the sheet/portal

The confusing part is that the in-plane deflection limit is specified to be mid-height rather than at the knee, which is how they're actually designed in practice (in my experience)
As I said, I assume that's because the critical deflection is actually at the mid-height (roughly) due to the non-linear deflection profile
If you use deflection at the knee then you underestimate the curvature through the critical region

See crude / realistic portal below - deflection at the midheight is ~60% of the knee

 

[Tomfh]

Oh you’re talking here about a wall running across the page, eg a wall built within the portal frame?

Interesting point that it might due to deflection non linearity within a racking frame.

 

[Greenalleycat]

Yea this clause/criteria is for in-plane deflection, so a wall within or beside frame

 

[Arun238]

Hi all,

Thank you for sharing your thoughts.

There is another paper ( Link below )

https://repo.nzsee.org.nz/bitstream/handle/nzsee/2591/Liu.pdf?sequence=1&isAllowed=y

The second paragraph of section 5 says the height as different. The first few lines referring 1170 says the sls limit is H/300 about 8mm ( considering a 2.4m wall ), which means the height (H) is 2.4m but the limit H/300 is checked against the deflection at midheight. This is how it comes down to 8mm as limit. In the image you shared 2.7mm < 8mm.

The last few lines says the in plane deflection limit for plasterboard walls at SLS shall be 8.7mm which again is close to H/300 - But at knee as per section 4.2 which sats top plate deflected 8.7mm at 0.72% drift.

But the paper itself says about seismic loading. So, is it applicable only for EQ SLS ? This is where the confusion came wrt to mid height limit ?

Thanks again

 

[Greenalleycat]

That paper you've linked is a bit ambiguous - it references 8.7mm interstorey deflection (which is the knee height, not mid height) controlling damage to the plasterboard to an acceptable level
This is consistent with using an H/300 limit for a typical 2.4m stud wall
Using H/300 at the mid height would be unconservative as (using my 60% estimation from the earlier model) would give you about 13mm at the knee - significantly more than the 8.7mm they said.

Honestly, my theory is that the table in 1170 has a typo
I think the deflection they refer to is INTENDED to be the in-plane deflection at the knee, checked against H/300 where H is the knee height
Instead, it says H/300 checked at mid height, which seems a bit odd - I reckon they copy-pasted the out-of-plane check and forgot to change the text...
Every paper I've read about this topic has dealt with knee deflection and every engineer I know uses knee deflection - no one is checking in-plane deflections at mid height

RE your last question about EQ SLS:
That paragraph is talking about SLS, not ULS, as it is referencing a damage control state (limiting deflection to keep plasterboard damage to an acceptable level)
ULS is discussed below Table 2 where it recommends a ULS drift of 1% - this is because a typical residential portal frame design would assume some tributary of load (say in a garage) with bracing elsewhere provided by GIB
If the portal frame drifts excessively then you can end up pumping more load into GIB elements on the same line, a deflection incompatibility issue that means your overall bracing design may not be valid

Note their key statement in paragraph 2 of section 5: "These limits are specified to protect plasterboard damage and do not necessarily reflect the bracing function of the walls."
This goes back to my previous comments/links about potential differential stiffnesses, damping, and deflection profiles between GIB walls and portals in the same line
An SLS deflection of H/300 = 8mm in your portal seems OK...until you see the hysteresis of GIB and realise that it is fucked by that stage and is massively reduced in stiffness

 

[Tomfh]

Honestly, my theory is that the table in 1170 has a typo
I think the deflection they refer to is INTENDED to be the in-plane deflection at the knee, checked against H/300 where H is the knee height

So, the non linearity thing is a red herring? Or does that still apply somewhere.

 

[Greenalleycat]

Non-linearity in which sense?
The picture I posted with the mid-height deflection from my portal frame model was me attempting to rationalise why Code would specify a mid-height deflection for in-plane deflection (rather than knee height)
I don't know if that is the reason, but I can't think of why else you would use a mid-height value

If you mean non-linearity in terms of plasterboard performance, that is very much a real and important effect, just one that isn't considered by many engineers

 

[Arun238]

Hi all,

Thank you for your explanation. I also agree with you that it could be a typo. Funny thing is its been referenced in these papers. For eg the one I shared. It says midspan.

Another thought is, the load case/combination mentioned in 1170 is only for wind sls. Not for seismic. So, does the paper is trying to put some thoughts on a seismic sls limit as knee deflection of 8mm / H/300 ?

And Wind SLS remains what said in 1170 ?

 

[Greenalleycat]

1170 doesn't include every load combination you should check - in fact, it kinda half asses it IMO
H/? limits are generally more useful than absolute limits as they obviously reflect that taller elements can deform more (in absolute terms) than shorter elements before the same damage is encountered
Serviceability combinations work backwards from the effect that you're trying to control, then they put numbers to it

SLS lateral movements are generally dealing with effects such as:
- occupant comfort (not wanting to hear or see visible movement)
- preventing excessive movement that could damage linings etc
- preventing movement that could damage windows (they have some tolerance inside the glazing units but will bind if pushed far enough)

Why would SLS wind deflections be different than earthquake SLS?

 

[Tomfh]

Non-linearity in which sense?

The non linear in plane deflection of the frame, ie the supposed reason you previously mentioned for the code prescribing “mid height”

I’m not sure if you’re saying it still applies, or if you’re saying it’s a typo.

 

[Greenalleycat]

The code wording says H/300 at mid-height for in plane deflection
I don't know why it uses mid-height as no one that I know, in practice, uses mid-height - everyone checks to H/300 at the knee
My thing about the non-linearity of the deflection profile was me trying to rationalise why Code would suggest something that doesn't make sense to me
The alternative explanation is that someone made a typo that was never picked up on

However, our code serviceability limits are only recommendations: you can not follow them and be OK, or you can follow them and something isn't satisfactory and you get sued even if you technically met code - it's fun
So in practice, no one uses the as-written limit

 

[Tomfh]

I tend to agree that it’s more likely a typo. It seems a bit pedantic to consider the knee deflection as too imprecise a reference point, for what are quite approximate guidelines to begin with.

 

[Arun238]

@greenalleycat & @Tomfh

Thanks for your input on this. Looks like everyone is on the same page that nobody knows why 1170 says midspan. I think its best to go with the H/300 for knee till the code revises.

Thanks again.

 

[human909]

However, our code serviceability limits are only recommendations: you can not follow them and be OK, or you can follow them and something isn't satisfactory and you get sued even if you technically met code - it's fun
So in practice, no one uses the as-written limit

Ah the magic and the traps of AS/NZ codes.

When I was more green I followed the recommendations of AS4100 of beam being 1/250. I was later chastised for this. What deflection criteria are most people using?

I generally have now run 1/350 with 1G + 0.6Q for steel industrial equipment floors which I plenty of. This seem a little conservative in my opinion especially when typical live load is 0. I've never had any issues with this approach, though I do wonder if I'm too conservative.

For residential I've been advised 1/350 or 12mm for soffits. That 12mm end up pushing me into some pretty large beams for open place construction. But I'm hesitate to step away from what has worked regarding servicabilility.

Any comments?

 

[Greenalleycat]

The vagueness is real!

I also tend to make up my criteria depending on the day and what sizes I'm getting
My main criteria is typically G + 0.4Q with k2 = 2.0 (long term creep for timber) to L/300 but I look pretty closely at absolute deflections and limit to 10-12mm for lots of uses
I like to look at the differences between the G & Q component too - I generally assume that the G component is being built out
Depends how critical the element is too - if it's a beam supporting a other beams then I'll either model it in 3D or limit to something like L/500-L/600

 

[Tomfh]

What’s the basis for 12mm? That’s a fairly specific number.

 

[human909]

I look pretty closely at absolute deflections and limit to 10-12mm for lots of uses

Good to know I'm not completely off base. I should have mentioned that my 12mm deflection is total long term deflection.

What’s the basis for 12mm? That’s a fairly specific number.

Excellent question. Unfortunately I don't have a good basis apart from that is what I've been told for higher quality residential construction. It could be simply half an inch. I've been told 9mm for window lintel which is ~1/3 of an inch.

I'd happily be told that is a ludicrous and arbitrary figure by anybody on this forum. (My step into residential design has been fast-tracked, I have alot to learn.)

My broader experience is structural steel in an industrial setting and is extensive. My general assessment criteria post construction has been my own foot-fall feedback or vibrating machinery. For large span floors this doesn't tell you much but for smaller areas and cantilevers your own foot-fall and bounce testing will give an indication of suitable user serviceability.

To date. I've never had issues from clients. But I operate on a continuous improvement basis and now that this topic has been raised I'm interest in what others do. (Both in AU/NZ and elsewhere.)