Joining Steel Beams

mickeyg14

Looking for some help/advice.

My builder has inserted a steel beam to support a previously removed chimney breast (well 2 beams with end plates and bolts connecting them).  The connection was designed by a structural engineer and I gave the builder the drawing.  

It turns out the builder has not followed the structural engineers plans.  This called for 6 x m16 bolts.  The builder has used 4 x m20 bolts and says this is plenty. The full beam length is 3.28m.  We've gone back to the structrural engineer to see if this can be rectified (I realise this is not his problem).

Whilst I'm nervously waiting for a response does anyone know roughly whether using 4 m20 bolts is going to be weaker than 6xm16 bolts.  I'm trying to remain calm and find a solution.  Very close to telling the builder he has to start again and comply with the original structural engineer's diagram but slightly concerned he''ll just consider it not worth it.

grumbler

3*16^2=768
2*20^2=800
However, it depends on how the bolts are loaded, and this isn't clear from you post.

LunaLater

An M20 has 1.56 times the cross-sectional area of an M16, so you have more total area holding the beams together, but if the failure mode was the threads being stripped then I don’t know how that compares.

This site has some data that suggest M20 is 60% stronger, so, again 4 M20 is stronger than 4 M16.

https://www.engineeringtoolbox.com/amp/metric-bolts-minimum-ultimate-tensile-proof-loads-d_2026.html

Section62

mickeyg14 said:

Looking for some help/advice.

My builder has inserted a steel beam to support a previously removed chimney breast (well 2 beams with end plates and bolts connecting them).  The connection was designed by a structural engineer and I gave the builder the drawing.  

It turns out the builder has not followed the structural engineers plans.  This called for 6 x m16 bolts.  The builder has used 4 x m20 bolts and says this is plenty. The full beam length is 3.28m.  We've gone back to the structrural engineer to see if this can be rectified (I realise this is not his problem).

Whilst I'm nervously waiting for a response does anyone know roughly whether using 4 m20 bolts is going to be weaker than 6xm16 bolts.  I'm trying to remain calm and find a solution.  Very close to telling the builder he has to start again and comply with the original structural engineer's diagram but slightly concerned he''ll just consider it not worth it.

Wait to see what the structural engineer says.  The design of structural connections involves a lot more than just the total CSA of the bolts used.

For example, you have to consider how the holes drilled in the end plates will affect their strength.  Larger diameter holes have a greater strength reduction effect than smaller diameter bolts.  The spacing between bolt holes, and the amount of material between the hole and the edge(s) of the plates can also be significant factors.  The positions of the bolts/holes relative to the axes of the beam are also important - for example a beam loaded from above may need the connection designed to have additional strength below the neutral axis to cope with (a greater) tensile load, whereas above the neutral axis the compressive load can be transferred directly between the end plates themselves.

Connections need to be designed and fabricated carefully.  Connections used in unusual or complicated circumstances need extra care.  Having a connection between two halves of a beam is sufficiently unusual that a builder shouldn't be winging it in any circumstances.  Only your structural engineer can give you the correct answer here - and that could involve a 'fix' which doesn't require the beam to be removed and completely redone, so don't panic yet.

Section62

LunaLater said:

An M20 has 1.56 times the cross-sectional area of an M16, so you have more total area holding the beams together, but if the failure mode was the threads being stripped then I don’t know how that compares.

This site has some data that suggest M20 is 60% stronger, so, again 4 M20 is stronger than 4 M16.

....that might apply to the bolts (although it does depend on what grade have been used compared to what were specified), but it doesn't necessarily apply to the connection as a whole.

mickeyg14

Thanks for the responses.  I will hold out hope.  

The beam was split into two to get it up the stairs and in the room.  The configuration of the bolts was precise in the structural engineers drawings with four of the bolts at or below the centre line and two above.  

The chimney breast it's supporting was removed to attic level by previous owner.  The stack serves/is supported by 3 other chimneys so everything is probably overkill but better to be safe than sorry.

stuart45

An end plate splice is normally joined with 6 bolts on beams that size. The positions are due to where the splice is in tension. 
Yo can only wait for the reply from the SE, but he might not be too happy.

Section62

mickeyg14 said:

The beam was split into two to get it up the stairs and in the room.  The configuration of the bolts was precise in the structural engineers drawings with four of the bolts at or below the centre line and two above. 

Who did the fabrication?  Normally steel fabricators would know the importance of following the SE's instructions to the letter... it would be unusual for someone who was used to doing structural steel fabrication to take it upon themselves to change the number, size and positions of bolts.

In a situation like this I'd be wondering whether the builder has got an unqualified mate to do the job, in which case you also have to wonder whether the welding was done correctly.

Is the size of the endplates correct? The photo isn't clear enough to see.  But for example it is important that the plate is oversized (i.e. 6mm extra top and bottom) so the welds on the 'outside' of the top and bottom flanges can be formed correctly.

LunaLater

Section62 said:

LunaLater said:

An M20 has 1.56 times the cross-sectional area of an M16, so you have more total area holding the beams together, but if the failure mode was the threads being stripped then I don’t know how that compares.

This site has some data that suggest M20 is 60% stronger, so, again 4 M20 is stronger than 4 M16.

....that might apply to the bolts (although it does depend on what grade have been used compared to what were specified), but it doesn't necessarily apply to the connection as a whole.

Yes, I’m aware of that; they’ll need to wait for the report.

grumbler

mickeyg14 said:

According to my calculations, in this design two bottom rows of bolts take about 85% of the load (35%+50%). Not sure what's critical here - shear stress in the thread or tensile stress in the bolt body, but I think this doesn't matter. What matters, is that one 20mm bolt replaces approx 1.7 16mm bolts that is very close to 1.56 figure posted above, but still slightly bigger. So, I think the SE engineer's approval is really needed.

TELLIT01

I've just looked at a larger version of the photo and the join seems to be supported on a couple of bricks placed on timber units!!!