Structural Engineer Not Providing Connection Detail

[Gus Potter]

Good points from all. It's worth asking the question whether the connection design is included or not.

 

For domestic applications don't forget that although the beam sizes may be fairly small, the connections are also small so you can develop pretty high stress etc in these. Typically stress is calculated by Force / Area ... so small force / small area leads to the same/ similar stress concentration as ~ Big force / Big area.

 

A good number of frame analyses packages (this is a bit of software that allows you to join all the beams etc together and quickly calculate the forces / deflection etc ) have a bolt on connection design module. This module allows you to design the connections and export the connection detail to another cad package or just print it out along with the calculations. The main thing is you need to know enough / have experience to make sure that what the software is printing out is not rubbish!

 

A good thing here is that you can play about and tweek a connection  to say fit in with the Architectural detail, or you may be forming a big slapping in a wall for a set of bifolds. Here you may have some torsional forces (twisting) in a connection which you can't avoid and the look up tables don't work for this. The software does the donkey work and you can then modify the calculations to suit.

 

I find that there are a good few fabricators, partcularly smaller ones who don't have their own in house SE's (or are just too busy) to spend time connection designing on a small job. If you can hand it to them on a plate where they can just pass it to the shop floor you can make savings that offset (sometimes more) the extra SE cost. For domestic stuff a local SE will say know the local fabricators, how they like to work and play to their strengths and this can bring the cost in the round down.

 

For the unwary.. connections don't always play fair! For example you may think that if you have a detail that shows a 8mm thick plate and you swap this for a 20mm thick plate (thinking it must be better) you have lying about this can suddenly make the connection much stiffer as the 20mm plate is less flexible than the 8mm plate, now you can overstress the welds and welds are often not as ductile (stretchy) as say the steel in the beam.. so you can get a sudden failure.

 

An easy way of looking at this is if you take a ruler and support each end on a small bit of copper pipe, load it in the middle,  the ruler bends and the ends rotate on the pipe, call this a pinned connection. All the pipe needs to do is carry half the vertical load you apply to the ruler.

 

Now if you fix each end of the ruler to the pipe (which could actually be a column) it will bend less in the middle but this fixity at then ends introduces a "rotational" force called a moment as well. Now you have a moment connection. Here the top surface of the ruler is in tension, the bottom in compression. You can now see that the end connection has to deal with an extra force that can be quite substancial.

 

On a lastish point, I have found that the easier you make it for a small builder fabricator on particularly domestic work the better in terms of cost. Yes, the SE may be a bit more expensive but the savings can be substancial. I always like to see the options explained in terms of the brief and this allows for an informed decision.

 

 

 

 

 

[SteveMack]

Sorry for the slow reply guys, manic week in work.  It seems as though I've opened some interesting discussion though so I'll add some more if it helps the debate..

 

I've attached 2 files, 1 is the layout of the steel and the second is the list of steelwork to be used and forces for which the connections need to have designs.  I have an 80 page calcs document too in support of the design.

 

Some factors; house is 2 storey brick and block cavity with a loft conversion which is designed to put all the roof weight onto the gable block wall via steels in loft, so there is increased loading on the gable ends and thus can't use 2 lighter bolted together beams due to uneven load on brick/block skins.  Beam 1 is supporting ~80% of the gable wall length of the house.  Beam 2 takes out the corner of the house.

We live next to open fields and regularly get hit with 60mph+ winds with potential for a lot more if it's a bad storm, beams 1 & 2 are the walls facing the very open aspect.

Edited January 28, 2021 by SteveMack

[SteveMack]

I mis stated that RHS section too sorry, it's 300mm deep!

[SteveMack]

I've also realised there's no dimensions to give a size to it.  Beam 1 is ~6.1M, Beam 2 is ~4M.  Beams 3&4 were split from 1 single beam to reduce the weight a little on that beam span hence column 3 being present.  Beam 3&4 are quite deep because they need to accomodate a support for an outer brickwork soldier course (The bottom plate has been missed off, he's aware of this, he has agreed to amend further on).  Extension is single story.

[Gus Potter]

Hi Steve.

 

This is a bit of a puzzle as there is not much info, but enough to spark a bit of geeky interest!

 

Based on what you have posted so far re section size, forces and so on I can sort of see how something like this may work with the wind bracing and so on, but you are going to need a good bit of space to form some of the (assuming bolted) connections, you need to look carefully; if there are window details, a limit on the floor depth and so on. This could be a problem?

 

What you have posted so far looks like the maximum forces and maybe the fabricator needs more info. Take column 2 supporting beams 2 & 3. Simplistically these beams have a bending force at the ends.. the moments - kNm. Beam 2 moment loads the column about it's "strong axis" and beam 3 loads the column about it's weaker axis. If you combine these forces with the torsional moment plus the vertical load (V) then the column strength capacity looks debatable to say the least.

 

What the SE may have done is to model beam 3 with a pinned connection at column 2 so it does not transfer a moment (axial load ~V) but at the other end (at column 3) modelled a fixed connection and provided this set of more onerous forces rather than the different forces at each end. This may partly explain the fabricator's response.

 

The connection at column 2 in particular is relatively complex and will take a good bit of thought to account for all the combinations of forces.  Looking at this I think it's likely more than a ten minute job to design a connection that will be as small as you can get it and thus maybe fit in to the space you have available? You could end up with some pretty big bolts, thick end plates on the box section, stiffeners etc, nearly get it to work then find you have a bolt clash or the sequencing of the works stuffs you in terms of the order you connect in the various beams.

 

This could be a communication issue but I would just ask the SE roughly what sort of connection and size and form they have in mind. Also ask if they can do the heavily loaded connections with "ordinary" bolts which a local builder can manage ok.  Or does the SE think some kind of tension control (HSFG) bolt may be required which requires more specialist knowledge and tools to install. Make it too complex and most small builders will back off or add a lot to the price.

 

I dabble a bit with connection design so would be interested to see the solution and how it all fits with your structure.

Edited January 29, 2021 by Gus Potter
geeky typo

[SteveMack]

Morning @Gus Potter I've just messaged the full calcs etc over to you for some light lockdown reading

[George]

On 29/01/2021 at 19:31, Gus Potter said:

...

 

The connection at column 2 in particular is relatively complex and will take a good bit of thought to account for all the combinations of forces.  Looking at this I think it's likely more than a ten minute job to design a connection that will be as small as you can get it and thus maybe fit in to the space you have available? You could end up with some pretty big bolts, thick end plates on the box section, stiffeners etc, nearly get it to work then find you have a bolt clash or the sequencing of the works stuffs you in terms of the order you connect in the various beams.

 

...

 

 

This is the crux of the matter. I think I would have gone for two connected columns, each orientated for each portal frame - or as you say, not have a moment connected from B3 to C2 and rely on the C3/ C4 portal frame in the up/down direction.

 

There's something of an ongoing debate in the structural engineering world about whether steel frames in extensions can rely on the stability of the existing building (if enough walls remain) or whether they need to be independently stable. I have tended to use the existing building where possible...

Edited February 3, 2021 by George

[Gus Potter]

7 hours ago, George said:

 

 

This is the crux of the matter. I think I would have gone for two connected columns, each orientated for each portal frame - or as you say, not have a moment connected from B3 to C2 and rely on the C3/ C4 portal frame in the up/down direction.

 

There's something of an ongoing debate in the structural engineering world about whether steel frames in extensions can rely on the stability of the existing building (if enough walls remain) or whether they need to be independently stable. I have tended to use the existing building where possible...

Good points from George showing how you can make a small change that could have a big impact on the cost and buildability.

 

George also mentions the debate about how a steel frame interacts with an existing (assume masonry walls for simplicity) structure. I'll have a stab at this.. to all, please feel free to comment.

 

If you go back a bit, before the days of reasonably priced bifolds etc you often added a modest set of French Doors that maybe were a bit wider than say an existing window. Here, you still had plenty masonry each side to keep the building stable horizontally when the wind blew. Often you could just put in a concrete lintel or a modest steel beam just to hold up say; the wall on a second storey, floor, roof above.. you only had to deal with the vertical loads.

 

But now we are putting in much larger openings and are left with much less brickwork to stop the building from moving from side to side. When you do this you can end up with masonry each side of a large opening that looks more like a column than a wall. One problem that starts to crop up is that a column is much easier to overturn (push over) than a longer length of wall. Also, you have a shorter length of mortar bed to resist the sideways (shear) forces and a few other bits and pieces.

 

@George "I have tended to use the existing building where possible..." The key bit in Georges good comment for me is "..where possible."

 

If the remaining walls won't do the job then often the SE is faced with a dilemma.

 

Can you (SE) say.. well the remaining walls will take a bit of the sideways load and I''ll introduce a steel portal / box frame to take the rest and keep the steel sizes etc down or do I assume that the remaining walls don't make any contribution and design the steel frame to take all the load.

 

Theoretically you can assume the former but a brick wall (and what is attaches to) behaves in a different way from the steel. Steel is more elastic than brick and it (steel) does not suddenly crack. An easy way of explaining this is to imagine that you connect together a plank of wood and the same size plank of steel with the strongest and stiffest glue you can imagine. Put your composite plank over a river and walk over it. (how you do that I don't know ) Both the steel and the wood will bend by the same amount as they are connected together. But steel is less elastic (stiffer) than the wood (called the modulus of elasticity, Youngs Modulus) so the steel will take more of the load than the wood as the wood is more "stretchy".

 

This principle underpins the design of a flitch beam... it's very clever really but simple when you think about it.

 

Suddenly the analysis becomes very complex and thus expensive design wise and even then the structural modelling of this is still only a model.

 

Most SE's for practical purposes on small domestic applications (where the remaining walls are not enough) will design the steel to take all the vertical and sideways loads and limit the amount it moves by, so they don't crack the masonry excessively.

 

This may seem like the SE's are being lazy.. but many know that while you could undertake a more refined analysis you will get tripped up when it comes to designing the connection between the steel and the masonry.. so back to square one. There is also the issue about how confident you are as to whether the builder will actually do what they are supposed to do.

 

For small domestic stuff where supervision may be not as robust the simpler you make it the better, not just in terms of cost but also in terms of safety.

 

To get the best value for this type of work you maybe want look at this in the round. Do you for example need to live in the house while work is going on, how hard / complex will it be to prop and so on. Will the size of the steel connections start to interfere with your window / thermal details ect. How much of a ceiling height can you live with vs downstand of a beam. I often find that the cost of the actual steel is not the main issue, it is the labour and complexity of the fixing of the steel to the existing building and the cost  / labour involved in the temporary works.

 

Sometimes you can skin the cat in another way where it's more economic to take the wall down above , take the load of the propping, put your steel in and rebuild the wall after... any takers?

 

 

 

 

 

 

 

 

[SteveMack]

We've reached the point now of appointing a new structural engineer.  I've had quotes back for £500 to design the connections from fabricators due to the quantity and type and going off other discussions with 2 new engineers a redesign of the structural element will not only include this connection detail but also save thousands in steel and foundation costs.

 

Anyone know what this will do to our building regs approval?  Do I just get the drawings/calcs redone and send off again for checking or will they charge us again for this?

 

Are there any rules I have to follow when I terminate current engineer?  He's fully paid and nothing outstanding there.

[George]

No rules, just courtesy to tell them. The new Engineer might wish to contact them as well.

 

You'll have to ask BC but they may charge. To them it's just a design change, they don't really 'check' calculations.... they just check that you have had them done!

Edited February 13, 2021 by George

[SteveMack]

2 hours ago, George said:

No rules, just courtesy to tell them. The new Engineer might wish to contact them as well.

 

You'll have to ask BC but they may charge. To them it's just a design change, they don't really 'check' calculations.... they just check that you have had them done!

 

New engineer has just said that, he will give old engineer a call for courtesy and check for any objections to taking the project over. 

 

Sounds good anyway.  I assume the architect is going to moan for having to redo drawings too?  The principle layout isn't changing but the section detail probably will.

 

I'll be glad when I get a shovel in ground on this!