Am I in the right neighbourhood with these RSJ profile dimensions?

[DRM]

Hello,

 

I'm in the design phase of my retirement home. 

 

Briefly, this will be a single storey home on a reinforced concrete slab foundation. The house footprint is approximately 20m x 16m with a central 7.7m x 6.5m atrium glazed on 3 sides and topped by a glazed roof. Perimeter walls are structural 24cm AAC with external insulation, VRB and render. Roof slab is 20cm 2% reinforced concrete extending from the perimeter up to the 4 sides of the atrium (meaning the roof slab has an area of approximately 270 m2, volume of 54m3, and weight of approximately 135 tonnes).

 

The roof slab is supported above a suspended gypsum board ceiling by a framework of steel beams that are welded to one another and bolted to the perimeter walls into a reinforced concrete sub-roof transition zone poured on top of the AAC. Roof slab will be poured out to the perimeter of the building and will be clad (along with the sub-roof transition zone) with more efficient insulation than the insulation cladding of the AAC so as to provide a uniform u-value and flat vertical facade. AAC block will be laid above the roof slab for an additional 60cm for the parapet. The steel beams will be supported by rectangular hollow steel columns welded to base plates that are bolted to the foundation slab. 

 

My question concerns the steel beams and columns for my design. I'd appreciate some feedback of my initial sense of the size of beams and columns I will need. The beam assembly consists of two long spans running down the sides of the long axis of the atrium that are joined by two shorter spans traversing along the shorter sides of the atrium. To simplify transport and installation, each of the two 20m long spans will be composed of 3 shorter beams welded to one another to make up the 20m length. The two shorter beams traversing the 6.5m sides of the atrium will be welded to the 20m spans. All beam intersections (those where the 6.5m beams meet perpendicularly with the 20m spans, as well as those to join the 3 beams into a single 20m beam), are located above the 4 corners of the atrium and are supported by 4 RHS columns with top plates providing substrate for welded connections of the beams.

 

Apart from the roof/parapet construction weight other components of static loading will be a fairly extensive photovoltaic array and solar thermal water. There is not a requirement for heavy snow loads etc. 

 

Based on my research I am designing with the following sizes in mind: IPE 300 for all beams and 200x400mm RHS sections for the four columns. I realise I will need to engage an architect and SE eventually but before I make this contact I want to have a good sense of the general dimensions I will likely need now since this has knock on effects for other elements of the design. Do these specifications sound like I am generally in the right neighbourhood in terms of overall dimensions for my beams and columns?

 

Appreciate any advice from fellow builders here.

[markc]

SE first, wind loads, live loads, connections, torsional loads have more effect than the mass of a structure

Edited December 7, 2023 by markc

[saveasteading]

In a way it makes sense to have a stab at the beam depth to see if the geometry of the house works at all.

So here is my advice. A 20m spanning  beam is impractical, and would be very deep. Perhaps over 1m.

More likely a complete portal frame of 2 columns and a beam. 

 

Try that. Then revert to getting professionals in sooner rather than later.

 

Why do you need a 20m clear space?

 

 

[kandgmitchell]

In my minds eye I see a horizontal "picture frame" 20m x 6.5m with a post on each corner. I can't however see how each splice in the 20m beams will also be located over the corners of the atrium. This will need scoping by an engineer because there is so much here that can't be done easily on a domestic scale. No snow load? is this in Dubai? 

[GavH]

Took me a little time to get my head around what you are proposing.

 

The max wide of the building span is 20m, however due to the 4 columns providing the Atrium corners you only had to deal with a maximum unsupported span of 7.7m.

You will also have to consider how the beam will produce substantial point loading on your external walls and may required additional columns at the outer wall ends of each beams too.

 

This is more that some structural Calc's you need an engineer to be involved on it from day one.

 

 

[Conor]

SE, mainly because your design could be completely wrong and you'll be wasting a lot of time and energy. I'm an civil/ environmental engineer but not structural, and my preconceptions and design ideas were blown out of the water by what the actual SE came up with. And an SE isn't going to want on site welding.

[DRM]

Thanks for the fast responses. I didn't describe the design as well as I could have - although GavH seems to have pretty much got it. It's essentially 4 volumes distributed around the atrium. Picture is worth a thousand words so here are a few quick sketchup captures that hopefully give a better sense of the design.

 

 

[DRM]

...and some captures showing a few of the structural details. As I say - I know absolutely that I need to get an SE involved but wanted to get a sense first of whether my IPE 300 and RHS 200x400 thoughts are in the general neighbourhood of what the design would require.

[DRM]

...and a  final detail giving the column height with top plate.

 

[markc]

Is this a self build home or a design project? When designing a home it’s al about the rooms, spaces, light, livability etc. not the structure detail.

[DRM]

Self build for as much of it as is reasonable. I'll of course bring in specialized trades when I need to - e.g. the steel framing etc. 

 

I've certainly thought through the other issues you mention but at this stage I want a sense of whether I am working with a viable starting point in terms of the building envelope. If this is basically in about the right place in terms of sizing, then I've got a quite detailed set of sketchup plans laying out the living areas of the house. If I am overlooking something, or should be incorporating significantly different structural support dimensions I'd like to know about that early on. And again, I will of course be engaging an architect and SE etc. but just want to have a sense of this now.

[Mr Punter]

You are wasting your energy focusing on section sizes.  Get a design and layout that works for you and will be acceptable to the planners.  Get a good structural engineer to design the foundations and structure.   Make sure you plan for decent airtightness and insulation.

[saveasteading]

The beams will be about 400mm deep. To minimise intrusion to the room, the joists can be inset.

 

What will the atrium walls ( and external walls) be made of?

 

It's good to see you enjoying the development of the design. I'd say  work out the room layout and then get an SE.

 

 

[ETC]

Try it without steel. The Romans did.

[saveasteading]

2 hours ago, ETC said:

The Romans did.

A nice stone arch? 

[ETC]

18 minutes ago, saveasteading said:

A nice stone arch? 

No - atria.

[ETC]

[George]

Interesting design

 

If you want a rough estimate of a beam depth use the ratio L/24 (L is the span, same unit for the beam depth, rounded up to.ansection designation - this is very rough and ready but is the only structural calculation I'll let an architect do).

 

RC roof slab will be very expensive because falsework mobilisation costs will add up. Hollow core or just precast planks with a structural topping, or just composite decking, would be better. 

[DRM]

Thanks everyone.

 

Re Roman inspired stonework that's not really the aesthetic I am going for here. I'm a fan of the Japanese aesthetic and this atrium is inspired by the central courtyards and atria you see in Japanese homes. This is a departure all the same, as it will be sided on it's north south and west sides by full width sliding glazing between the columns. The east wall will be solid as that gives on to bedrooms. The house entrance is on the north facade (top of these drawings), and a good percentage of the south facade is glazed so on entering the home one should be able to see through the atrium out into the back.

 

George thank you for your L/24 rule of thumb - do I use the entire combined span or only the longest spans between columns as my L? Also take your point about the RC roof slab - AAC slabs are available up to 6m in length and those would be significantly lighter as well. I have given some thought to that but want some spans that are just a bit over the 6m...your line of thinking re hollow core and prescast sound like a much more appropriate direction.

[ETC]

Get an engineer!

[saveasteading]

15 hours ago, George said:

estimate of a beam depth use the ratio L/24 (L is the span, same unit for the beam depth, rounded up 

@George My SE boss taught  me to use L/18 as the first stab. That is safer in ensuring enough space is allowed for an economic beam.

This was in the days of number crunching, so being close with the first assumption saved a lot of iterating.

Perhaps Euro sizes can be shallower, but if anything they are skinnier. Please tell me if 1/18 is no longer correct and I'll change my general advice.

Edited December 10, 2023 by saveasteading

[saveasteading]

@DRM  

 

Just to avoid you disastering this project:

1. To support a sliding door, the beam will be much deeper.

2. There's a lot more to this than compiling a list of materials.

3. That's enough from me. It's  time for you to engage an Engineer or it will go horribly wrong.

4. Good luck though.

 

 

[George]

13 hours ago, saveasteading said:

@George My SE boss taught  me to use L/18 as the first stab. That is safer in ensuring enough space is allowed for an economic beam.

This was in the days of number crunching, so being close with the first assumption saved a lot of iterating.

Perhaps Euro sizes can be shallower, but if anything they are skinnier. Please tell me if 1/18 is no longer correct and I'll change my general advice.

Anywhere between L/18 to L/24 would be fine for a first stab. 

 

It really is just as an estimate to avoid an architect leaving 100mm void for a 5m opening. If supporting external walls etc then L/18 is the better end but there's always the get out of heavier weight or UCs. 

[Gus Potter]

Hi all.

 

Eveyone get a like here.

On 07/12/2023 at 06:33, DRM said:

I'm in the design phase of my retirement home.

DRM, I like the simplicity of your model and saw your post popping up a few days ago.

 

I've put a few numbers to everyone's suggestions regarding beam depths as a conservative first guess and made a few conservative assumptions (please forgive the typos / grammer as I'm off duty):

 

1/ You are probably going to put a lot of glass surrounding the courtyard bit and maybe a fair bit along the outside walls. All this means that we need to control deflections so we don't cause glazing problems and if we control the deflections usually the steels and connections will also be fine when we check that it won't fall down or blow over in the wind.

 

2/ Slab thickness. Concrete slabs are heavy things. Let's run with your slab thickness of 200mm for now. As others have said there are other solutions that may be more economic, provide more service space etc. In the round we will end up with a similar roof weight or a bit less. Budget on the really simple stupid for now and then we refine. It gives the Architect (maybe you) more room to be creative if they know the "worst case" at the conceptual stage. 

 

3/ I'm also going to treat all your roof slabs as spanning in one direction rather than two. One direction span (not the pop group) means that you could use a structural B reinforcing mesh. Main bars in the span direction are thicker, the transverse bars control cracking of the slab and a few other things. We can make slabs thinner later if we make them span in two directions.. but we need to balance the cost / knock on effects if we do this... there is no free lunch in this self build world.

 

4/ I'm going to treat all the beams as simple supported and see what the later results are. When a beam is continuous over say a central column this support can attract a lot more load due to the beam continuity and that can impact on what happens at the bottom so you end up putting more load down the internal columns (can be up to as much as 40%) and foundation pads.  Let's go simple stupid for now and see what happens. anyway very long beams are a hassle as you may need special transport.. and that often costs a lot..and they are heavy..  Always look at the cost in the round of steels.. too heavy.. too long and the cost to get them on site and handle can out weigh any clever SE savings. Also pick steels that everyone stocks off the shelf. 

 

5/ I'm going to rationalise all the steel sizes so we can later see if we can get the best out of the stock lengths easily available. It also simplifies the connection design.. many other things that will drive the build cost down. Be aware that when you have steel beams with thinner webs and flanges it can really drive up the fabrication cost and these beams are not a stock item. Don't try and be clever at this stage loking for the lightest weight of steel as it will trip you up later.

 

6/ There are load of other things but let's do some sums. I'll work from the top roof covering down.

 

Remember I'm sizing conceptually based on deflections, not strength design. I'm going to split up the loads into two components / calculations:

 

The first are how much the beams etc will bend by when say it snows or folk are working on the roof and you introduce say pv panels.... call this the imposed or live load. 

 

The second is to see just how much the beams will bend by when all the self weight of the structure and the live load is applied.

 

Lets look at some loadings you could have on the roof if you say live in England / Northern Ireland and benign parts of Scotland, Wales. I'm not going to touch on wind loads (that push the building sideways) as that is for another day.

 

For all I'm going to work in kilo Newtons (kN) as units.  1.0 kN ~ 100kg to give you a flavour of what things weigh.

 

Live loads on the roof prescribed by the design codes:

 

Say typical snow load in England 0.45 kN/m^2 ( kilo Newtons per square metre of roof when looking on plan, straight down)

 

Roof access load for maintenance 0.6 kN/m^2, in the design codes.

 

The access load is greater than the snow load so I'm not going to consider snow load any more, particularly as the roof is flat and I'm assuming that there are not any higher buildings nearby that could dump snow on the roof.

 

Pv panels.. say you have panels but want to put them on a frame to get the aspect right for best preformance 0.25 kN/m^2

 

Total live load =  0.6 + 0.25 = just say 0.75 kN/m^2 Live load ~ 75 kg/ m^2. Incedentally this number correlates to some of the old roof design codes. I use this when assessing old roofs for pv panels.. it's a funny world!

 

Dead loads on the beams:

 

Say a PTFE roof covering with some ply on PIR insulation. But we need to create some kind of slope in the roof to let the water drain. I'll play safe here and allow for a concrete screed to create the fall, say a cement screed or something that averages 50mm thick (say max 5400mm span with a 1:80 fall = 5400 / 80 = ~ 70mm of average screed required once you put the bells/ extras on. Average screed density + insulation  + waterproof covering = 22 kN/m ^2 thus 0.07m * 22 = 1.54 kN/m^2 say 1.5 kN/m^2 as a ball park figure.

 

Take the concrete slab as 200mm thick as per the sketch up model. The weight of this is the density of a lightly reinforced concrete (I'm avoiding a thin slab with lots of expensive/ complex rebar)* the thickness = 24 kN/m^3 * 0.2m thick  = 4.8 kN/m^2. Note: I have not spent much time mucking about with the other loads as the concrete weight really is the dominant component.

 

Now we need to put in a ceiling and some services on the underside. Allow say 25 kg/ m^2 ~ 0.25 kN/m^2

 

Add up the total  loads: 1.5 + 4.8 + 0.25 = 6.55 kN/m^2 say 6.6 kN/m^2 ~ about 650 kg/m^2.. Now that is quite a lot if you compare that to say a timber flat roof with say posi joists which will clock in at a lot less.. but I have a light weight warm roof on part of my house and you can hear the birds trotting about.. and I'm pretty deaf to boot!

 

Ok now we have some loads on the roof split into two components.. the live load and the dead.

 

I'm going to take your longest spanning beam as 5460mm and that it has to support a roof slab width of 6445mm.

 

First calculate the uniformly distributed load on the longest spanning beam.

 

Live load only = 0.75 kN/m^2 * 6.445m /2  = 2.42 kN/m = w1

 

Total load on beam = 6.6 * 6.445 /2 = 21.27 kN/m =w2

 

Now let's calcultate the amount of deflection we can expect:

 

Using the formula (assuming all the load is uniformly distributed along the beam) Deflection = 5*w*L^4 / 384 E *I.

 

Lets try a universal UK beam section of 305 (deep) x 165 (flange width) UB 46 kg/m which is a good mid range section, off the shelf and can head off problems (Architectural input) when you come to connection design and the beam twisting / buckling.. for another day.

 

w = the load

L = the length of the beam

E = is what we call the Young's modulus,  the elasticity of the steel

I = the second moment of area  = a geometric property of the beam

 

For the live load the anticipated deflection mid span will be about 5* 2.42 * 6445^4 / / 384 * 205000 * 9900 * 10^4 = 2.67mm that is not a lot but!

 

For the total load the overall deflection mid span will be about 5* 21.27 * 6445^4 / / 384 * 205000 * 9900 * 10^4 = 23.5mm

 

The 23.5 mm deflection is the key!

 

That is quite a lot for the longest spanning beam to bend,  nearly an inch in old money! But here we get a bit more thoughtfull. I know that once we start to stiffen things up and get them all interacting I can reduce the overall defection a good bit. I think I could get this down to under 15mm once I start to value Engineer. The main thing for me is the glazing and how you handle the deflections.. and all that feeds back into the Architectural side.. the detailing, thermal bridging etc.. it gets reallty interesting here.

 

Lets go back to what @George and @saveasteading were advising!  Span / 18  and span  / 24

 

@GeorgeMaximum beam span  / 24 = 6554 / 24 = 273mm beam depth

 

@saveasteading Maximum beam span  / 18 = 6554 / 18 = 364mm beam depth

 

You can see I have picked a UK beam with a slightly wider flange which is smack bang in the middle. In summary George and SaveaSteading are two highly experienced Engineers.. it should come as no surprise that they know their stuff.. all I have done is put some numbers to their experience and picked a beam with a wider flange!

 

My span to depth ratio is 6554/ 305 = 21.4!  I'm not shitting you all,  I've not manipulated the figures.. I just thought I'll write write this post, do some calcs and if at the end I disagree with the other posters I'll say so and why.. but no need at my end to disagree.

 

@DRM ask this.. now you have comment from three Engineers, an Architect @ETC, professional Contractors, self builders that have been thee and worn the tee shirt and more... time for you to make a donation to BH even if it's a tenner..  it will be good for your sole.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Edited December 13, 2023 by Gus Potter

[DRM]

Dear all,

 

My apologies for the long delay - I am on what is meant to be one of those "once in a lifetime" vacations with my wife in Australia. So lots of strange and wonderful sights, friendly people, very unusual animals and even though we have been "on the grid" and connected, I've made a conscious effort to be offline. 

 

Thank you so much for the combined inputs - and particularly Gus for your very thorough discussion which was very helpful and informative. Looking back over my posts I had neglected to add that this home will be built in northern Portugal - hence my citing Euro beam profiles - and the SE and architect I will need will both obviously be Portuguese. I suppose part of my desire to check in on some basic issues here was to get a sense of how viable my design is as a starting point amongst native English speakers prior to discussing with professionals whose first language will be Portuguese.

 

One thing this brief thread has also made me realise is that it seems one can lay roofing panels in perpendicular directions. If this is correct, it seems to me I could indeed use AAC roofing panels for the entire roof which would reduce weight, provide some thermal insulation, and keep me in a material with an overall reduced environmental impact. As I mentioned these are available in lengths up to 6m (6.75m in the UK) and so if I can orient the roof panels in roof section to the north and south of the atrium between the two long beams in a north south orientation and the roof panels down both long sides of the building in an east west orientation then I would be within the spans that are available to use AAC roof panels.

 

Thank you once again to all for your inputs - I really do appreciate the combined wisdom of so many - and as Gus points out so many who are professionals in the domains that are important to me.

 

Best wishes,

 

DRM