What's this steel bearing plate going to do under this steel I beam?

[TimSmall]

I'm doing some work on my house at the moment, and have had a design produced by a structural engineer for a steel I-beam to span a newly made (4 meter) opening in a brick wall. They've given me a design (see the attached picture for my interpretation) for the bearing of the I-beam which I can't get my head around (I've phoned and checked that's what they definitely intended, and they've confirmed that it i...). I have no engineering, but I did study Physics and Maths+Mechanics to English "A-Level"...

The steel I beam has a 12mm steel "spreader" plate underneath it (or rather under the 150mm piece of it which bears on the wall), which is not welded to the I beam. As far as I understand it, the job of the spreader plate is to distribute the load which is imposed by the steel beam over the underlying masonry so as not to put it under excessive compressive force, but I can't see how that spreader plate is going to do this. Moreover, won't the masonry pier itself end up with 47.3N acting on a line only 75mm from it's edge (maybe closer to the edge if the beam deflects a bit)?

If both the wall and the I beam ran the other way (i.e. at right angles to the span of the I beam when viewed on plan), then I can see that the spreader plate would work just fine (and I've seen such designs both on paper and built). Similarly, if the I beam extended the extra 180mm over the top of the spreader plate, then I assume there would be no problem there either (but then the plate itself would be redundant)?

I'm hoping that my doubts are unfounded, and the design is fine... Any comments?

 

[afreiden]

You are assuming that the "spreader" plate it is there in order to "distribute" the bearing load.

Perhaps this is not the case?

Instead, maybe the "spreader" plate will be bolted down and is placed there in order to resist the tension in the row of bricks just under the I-beam. The engineer doesn't want that edge brick to come off, for example. This is plain old unreinforced brick, right?

Image search for corbel design and you see similar ideas.

 

[TimSmall]

The design calls for no bolting or welding - from the bottom up there is, unreinforced on this a mortar bed, then the steel plate (just bedded on the mortar, no other connections to the masonry), then the I beam sits on that.

The engineer refers to it as a "spreader plate", stating:

"All padstones to be 330x100x12mm thick spreader plate. All beams to have 150mm end bearing".

The sketch included in the design then shows the arrangement which I've drawn above (I've checked through this with her on the phone, but I haven't yet sent through the graphic which I did above).

I did a check using some beam design software "eurobeam", and that software includes a portion which designs a concrete padstone or steel spreader plate to go under the end-bearing of the steel I beam. Using that module of the software - a 330x100 horizontal padstone or plate was sufficient, but 150x100 was insufficient to distribute load onto the masonry without exceeding the maximum allowable compression in the underlying masonry. This software takes into account brick/block + mortar strength and defects etc.

So I'm reasonably sure that the intention behind the spreader plate is to "distribute" the load over a larger portion of the masonry, so-as not to overload it - but I don't see how it can do that if it's not welded or bolted to the I beam (I've checked, and there is no weld or bolting in the proposed design).

 

[Studiot]

Personally I doubt that the proposed steel bearing plate meets the codes for bearing.

The bearing code bearing capacity of new ordinary brickwork is around 2 to 5 N/mm². London building control requires less than 0.5 N/mm² for old brickwork.

Using your figures and some fag packet calculations I could not see proposed beam sizes so I have estimated my own.

Calling your end load 50kN that is 100kN distributed load on a 4000mm beam ie 25kN/m

Limiting the deflection to say 5mm (span/800) and entering table 16.4 of the Steel Designer Manual with

α = 800 extracting K =4.7

Thus I_required = KWL² = 4.7 *100*4² = say 7500 cm⁴

Entering Universal Beam tables the nearest above is a 305 x 127 x 42kg/m beam I_xx = 8159 cm⁴

This has a web thickness of 12mm and a flange thickness of 8mm.

Assuming the 50kN load spreads at 45° a point 50kN load corresponds to a pressure of

At the base of the web vertical shear stress = 50000/150*12 = 28 N/mm²

At the base of an 8mm flange this reduces to 50000/150*(12+8+8) = 12 N/mm²,
which is still too high.


Adding a 12 mm thick pad further reduces this to 50000/150*(12+20+20) = 6.5 N/mm²


This would be acceptable on high stength concrete blockwork or engineering bricks.

 

[afreiden]

OP, why do you keep mentioning connecting the spreader plate to the beam? How would that help?

Studiot, I don't understand what stress demands you are calculating.. Without the spreader plate wouldn't the flange WIDTH be the only relevant dimension that you need once youve sized the beam? With the spreader plate, why do you need to know anything about the beam?

 

[Studiot]

The vertical shear force in the web appears as a shear stress equal to the force divided by the sectional area of the web over the support. I have taken this area to be 150mm x web thickness.

This is therefore the stress at the root of the web ie the junction of the web and the flange.

This stress does not magically spread over the full width of the flange immediately on entry.

I have taken the conventional assumption of a 45° spread of this stress which allow calculation of the stress at any lower section eg the bottom of the flange.

Making the flange effectively thicker continues that spread.

You are correct there is no structural difference whether the plate and flange are connected or not.

 

[AlephZero]

The design calls for no bolting or welding - from the bottom up there is, unreinforced on this a mortar bed, then the steel plate (just bedded on the mortar, no other connections to the masonry), then the I beam sits on that.

I think the main purpose of the plate is to provide a flat level surface for the beam to sit on, and the mortar bed under it ensures that the loiad is distributed over the whole area, rather than just the points where the flange of the beam might touch the uneven brickwork.

You can ensure the plates at each end of the beam are properly aligned with each other and allow the mortar to set and fix them in position, without having to deal with the weight of the beam and whatever loads it will support.

I'm not a CEng, so no comment on whether the area of the plates or brickwork is sufficient to carry the load. I would expect that meeting the building regulations in your country is more relevant than trying to do an "accurate" calculation.

 

[Studiot]

I think the main purpose of the plate is to provide a flat level surface for the beam to sit on, and the mortar bed under it ensures that the loiad is distributed over the whole area, rather than just the points where the flange of the beam might touch the uneven brickwork.

Why do you think a plate would achieve this, better than proper bedding?

I could understand the plate as a bed to slide on if thermal considerations were important but say the temperature varies ±10°C the expansion will be less than 1 mm.

 

[TimSmall]

OP, why do you keep mentioning connecting the spreader plate to the beam? How would that help?

I had a feeling that if it was unrestrained, then under load the plate might rotate about the end of the beam when a load was introduced (since the beam is bearing on less than 50% of the plate, but the reaction to the plate is over 100% of the length). I've since done an experiment which seems to suggest that this might actually be correct:

http://youtu.be/Hoe_0PADGrs

... standing in for the I beam, I have a length of 18mm OSB/3, the spreader plate is being impersonated by a piece of hardboard (probably a bit harsh on the spreader plate, the hardboard has way too much flex - maybe I should try and find a bit of ply), and I have a piece of polyethylene foam to take the place of the masonry. Although the polythene deflects way more than the masonry, hopefully it shows where the load would be concentrated, as the masonry will deform slightly prior to cracking?

I put the plate in two different arrangements:

In the first, the beam bears onto the whole of the plate.

In the second, the plate is positioned roughly as it would be in the proposed design (i.e. the beam bears down on a little under 50% of the plate).

In both cases, the load was imposed vertically (as best I could manage), and was approximately the same magnitude (just enough to lift my knees off the ground).

For the first case, the load distribution seems to be uniform - as expected.

For the second case, it looks to me like the plate might be acting as a lever and that perhaps the load is actually being concentrated at the extreme edge of the masonry (i.e. the opposite of the desired effect). This seems like a bad idea to me, since isn't that exactly where the masonry will most prone to failure by shear cracking? When this shear cracking occurs, then the next bit of masonry will now have the greatest load, and may crack, and so-on.

With the beam covering more than 50% of the bearing plate, this effect doesn't occur, and the compression seems to be uniform over the whole polythene block once again (ignoring any plate deflection effects - which presumably ARE correctly accounted for in the design).

Any thoughts greatly appreciated!

Tim.

 

[TimSmall]

Personally I doubt that the proposed steel bearing plate meets the codes for bearing.

The bearing code bearing capacity of new ordinary brickwork is around 2 to 5 N/mm². London building control requires less than 0.5 N/mm² for old brickwork.

I share your doubts about the allowable bearing pressure onto the masonry wall (unless the stub wall is rebuilt) - and that information is very useful as the wall itself is about 60 miles from London in Brighton, and made from similarly naff old bricks with poor mortar.

Setting that aside for the time being, I was suspicious about the entire design of the bearing plate orientation, since if that's not actually spreading the load in the location and orientation that has been specified (and the engineer has confirmed that this is their intended orientation + position) then it doesn't matter how big you make the plate - all bets are off!

So, I wanted to get that established before progressing further! The fact the the engineer didn't make any comment about the original masonry bearing pressure, or give any calcs for the bearing plate made me more suspicious.

So... is the bearing plate doing what the engineer thinks it is (uniform load distribution) or not?

 

[Studiot]

Any thoughts greatly appreciated!

Well you were obviously not interested in mine before.

I doubt your building inspector will be interested in yours.

 

[TimSmall]

Well you were obviously not interested in mine before.

I doubt your building inspector will be interested in yours.

Sorry, I didn't wish to cause offence, I was working my way through the comments, and hadn't yet replied your second comment (although I did just get in there with a response to your first before you posted the one above)!

 

[Studiot]

So... is the bearing plate doing what the engineer thinks it is (uniform load distribution) or not?

Of course it is, however the thicker it is the the lower that pressure.

Look more carefully at my earlier post and try to provide the information I had to guess at. Then a more accurate calculation can be substtituted.

As a matter of interest if you cast a proper concrete bearing pad the allowable bearing pressure is 0.5 - 0.6 f_cu

 

[TimSmall]

Of course it is, however the thicker it is the the lower that pressure.

Look more carefully at my earlier post and try to provide the information I had to guess at. Then a more accurate calculation can be substtituted.

I'd be extremely happy to do that, but I'd really like to try and satisfy my doubts about the bearing plate orientation first. In the video I made, the plate starts to rotate about the end of the beam when under load. I'm quite prepared to believe that this wouldn't be an issue in the building, but I cannot yet convince myself that is the case...

It's this rotation and perhaps the potential for the load to be concentrated at the extreme edge of the stub wall which is unsettling me - what is the difference between my model and the actual proposed design which means that rotation cannot happen? What are the forces like at the extreme end of the I beam if that does happen? Ordinarily (or at least in all designs which I've seen), the centre of mass of the plate is underneath the beam (in fact it's underneath the beam web), but in this case it's 15mm horizontally beyond the end of the beam...

 

[TimSmall]

OK, how about we consider some hypothetical extreme cases, which have some relation to the problem. If we assume:

1. The steel beam and plate do not deflect.
2. The masonry is slightly compressible (and uniformly compressible), and for the sake of argument only in the vertical direction.
3. The steel beam rests on the plate, but is not connected to it by any other means.

The length (in side view) of the bearing plate is "b1"
The length (in side view) that the beam bears onto the plate is "L".
The force "f" is imposed down by the beam onto the plate.

and then try to ask the question:

What is the pressure imposed on the masonry in each case, and how does it vary if:

i. The beam bears on the plate over it's full length (L == b1)
ii. The beam bears on a small portion of the plate at its edge only (L is much less than b1 e.g. L == (b1/30))
iii. An intermediate case.

See attached pictures.

I think that:

In case i. all of f ends up imposed as uniform pressure onto the masonry.
In case ii. the pressure under the plate is greater at the edge of the masonry, and decreases as you move along the plate, away from the beam.

I also think that in the general (case iii.) the pressure under the plate is uniform, if and only if:

L > (b1/2)

(which incidentally is not the case for the design which I've been given).I'm reasonably sure of those statements, but why won't it cause a problem in the real world if L == 150 and b1 == 330?

Thanks!

Tim.

 

[TimSmall]

Hmm, illustrative files didn't attach properly the first time, let's try again...

 

[Studiot]

What does hypothetical have do do with it, I thought you were building an extension?

All beams deflect and rotate at the ends, if not fully restrained. The question is how much.

The effect of end rotation will be seen as a tipping upwards of the ends so the beam no longer rests squarely and evenly on the pads.

Without the information I asked for 'how much' is all guesswork and pointless.

Your engineer has designed the size of the plate to brickwork size for a single leaf. The only question is; is it thick enough?

The answer to that again depends upon the information I asked for.

Personally I wouldn't use a steel plate for the end rotation for reasons described above and because it would introduce practical construction difficulties that are difficult ( and uneccessary) to surmount.

 

[lvwarren]

You might want to model this in BeamCalc or some other tool. The spreader plate will tend to distribute the load across the entire 330 mm span, modulo a decrease in shear as you go across. If this is in a seismic zone I would visualize the entire structure placed on a shaking table and secure it with bolts and rebar accordingly.

 

[hokie66]

The spreader plate will do little, if anything to distribute the load or reduce the bearing stress. As the beam is many times stiffer than the plate, the load will essentially all be carried by the brick which is directly under the beam. Bearing plates, or spreader plates, are used to distribute concentric loads, not eccentric loading like this.

 

[nvn]

TimSmall: (1) The steel bearing plate ("spreader" plate) is necessary. The thickness of your current steel bearing plate is adequate.

(2) Your I-beam web crippling stress is probably OK, but we cannot be sure unless you post your I-beam size.

(3) Nominal bearing pressure, p, on brick with cement mortar must not exceed the brickwork allowable bearing stress, Sbra = 1.70 MPa. Your current nominal bearing pressure is, p = V/(b*L) = (47 300 N)/[(100 mm)(150 mm)] = 3.15 MPa, which exceeds Sbra, where V = applied load, b = bearing plate width, and L = bearing length. Therefore, the masonry currently appears overstressed.

It appears your 150 mm bearing length instead needs to be, L = V/(b*Sbra) = (47 300 N)/[(100 mm)(1.70 MPa)] = ~280 mm. Therefore, if your given end reaction force (V) is correct, it currently appears you need to increase your 150 mm bearing length to 280 mm.

(4) You might want to also see the last sentence in post 18, if you are in a seismic zone, or if you might experience high winds, tornadoes, or minor hurricanes.

(5) By the way, always leave a space between a numeric value and its following unit symbol. E.g., 12 mm, not 12mm. See the international standard for writing units (ISO 31-0).

 

[Studiot]

You need to post more detail than just the proposed beam size

Since this design is for the UK, you need to indicate how the load is arrived at.
Is it factored?

What is in the space between the brickwork and the end of the beam?

What are the deflection criteria for this beam. They are not a code requirement, they will be determined by what is in the opening eg a window frame will place more stringent limits than a clear opening in a room.

 

[hokie66]

I know that this is a physics forum, not engineering, but as an old structural engineer, I would never rely on a extended plate in that manner to redistribute the stress under the beam. The principle is that the load follows the stiffness, so the plate extension beyond the end of the beam will only see load after the brick starts to crush, or after that corner shears off.

 

[Studiot]

would never rely on a extended plate in that manner to redistribute the stress under the beam.

Agreed.

But we are only guessing at the intended purpose of the plate.

Aleph Zero may be correct in that it may be intended to bed the plate onto mortar to create a landing surface for the beam.
I have already explained why I don't consider this a good idea.

 

[AlephZero]

I agree it dioesn't look a very good design. I was just trying to rationalize what the designer was thinking.

On the other hand if the plate was slightly arched upwards, or the end of the beam was shaped so the contact was guaranteed to be at the end of the beam and the center of the plate and the load would spread towards both ends of the plate, that would seem a better idea - assuming the end of the beam would support the concentrated load without damage.

 

[nvn]

TimSmall: I would also be concerned about the stability of your masonry wall, in the lateral direction. I thought the required minimum thickness of a one-story, load-bearing masonry wall is 150 mm, not 100 mm.

Also, we do not know if your wall is properly anchored, in the lateral direction, to a stable structure, every 450 mm in the vertical and horizontal directions.

Also, ensure your house would be safe before you begin structural alteration or wall removal.

To ensure your plan sounds correct, you might want to have your entire plan independently checked by one or two additional, competent engineers who have expertise in residential building analysis, construction, and foundations, as well as expertise in load-bearing brick masonry walls, and stability.

You might also want to try some on-line searches for, e.g., brick masonry wall stability, etc.

If your existing masonry wall is load-bearing, I am not sure you can remove it, then concentrate all that load on the edges of an unreinforced, 100-mm-thick brick wall. You might consider rebuilding the end of your masonry wall, such that it would be thicker. And you could perhaps embed, within the wall end edge, a steel column. This wall needs adequate lateral support. I think this wall end also needs an adequate, independent foundation, underneath your building slab foundation.

 

[Studiot]

I thought the required minimum thickness of a one-storey, load-bearing masonry wall is 150 mm, not 100 mm.

Well 100 is not brickwork size although it could be blockwork size.

We have all said we are just guessing without more information

 

[xxChrisxx]

If this is a question where a real answer is required, and not a bunch of assumptions.

Why not just ask for the design justification from the person that come up with it?
You are the customer after all.

 

[TimSmall]

Agreed.

But we are only guessing at the intended purpose of the plate.

The engineer has implied (but has not yet confirmed - even when explicitly asked by email) that the intention of the plate is to evenly distribute the load evenly over its 330x100 area.

They have explicitly said by email that the location and orientation of the plate should be as pictured in my initial graphic.

I have been provided with factored loading calculations, and deflection calculations, but there is no calculation for the bearing plate provided - just the statement "ALL PADSTONES (sic) TO BE 330x100x12mm THICK SPREADER PLATE. ALL BEAMS TO HAVE 150mm END BEARING" - and a rough sketch of the bearing arrangement on plan.

When I asked by phone whether the bearing details were definitely right, I was told that it was (but I didn't feel comfortable or that my question had prompted much thought, so I then started this thread). The following day I put the same question in an email and I got this reply back "The steel spreader plate is required to be 20mm [my emphasis] thick for a 150mm end bearing. So, the size of the spreader plate is 330x100x20 thick." - but no calculation (which I've now asked for, so that I can see what assumptions have been made) - however from what they've said before, I'm 90% sure that the assumptions include even load distribution by the plate. There is also no explanation of why the 12 mm thick plate is now 20 mm thick.

I very much appreciate the other points made regarding wall stability, lateral loading, compressive strength of masonry units, beam deflection, beam restraint, web buckling etc. - I feel that they are all very much valid points (I've checked some of these points already using beam design software, and I won't continue with construction until all of my remaining concerns have been addressed to my satisfaction), but what I was trying to concentrate on, is the question of the load distribution under the plate - given that the plate is (as hokie66 has said) eccentrically loaded, and moreover that the centre of the plate->masonry bearing lies beyond the end of the beam.

From past conversations, I'm 90% sure that the engineer has assumed that the load distribution under the plate is the same as it would be if the plate was concentrically loaded (as I've said, I've asked for explicit confirmation of this from them - but not yet received it), and I'm now pretty sure that it isn't (I started off about 50% sure - so this thread has been very valuable, thanks!).

BTW, external to this forum, the professionals (and competent amateurs) which I've consulted have been divided on the matter of load distribution (although a rough consensus seems to now have been arrived at here - many thanks to all who've contributed!).

The reason for my focus on this one issue, is that without this having been established with some certainty, it's not really possible to check the rest of the design (or indeed any proposed modification to the design which may be required).

I intend to meet with the firm of engineers at the earliest possible opportunity, and also to meet with my local building control department's checking engineer (incidentally, the building control dept. have told me that they don't normally check "simple" designs from this firm - and a handful of other well-established local major companies - because they trust them to always get it right - I'll leave you to draw your own conclusions on that policy). I will also take the design to another independent engineer, if I don't feel that I'm getting much traction from my meeting at either of the above!

I'm also feeling somewhat up against it, as my builder is going to ask me at 8am on Tuesday (Monday is a public holiday here) how much of the original design he can continue to work to (the beam is cut, primed and on-site) - and how much he shouldn't (and it currently looks like he'll probably run out of things to do before the matter is fully resolved).

Again, this thread has been very useful in that respect.

I hope this explains my apparent tunnel vision on this matter of beam->plate->masonry load distribution, as I'm trying to get the relevant principles as clear in my head as possible (or alternatively an explanation from first principles of why my concerns were baseless).

I'm emphatically not ignoring the other points which have been raised - I've addressed some of them to my satisfaction already, and I'll ensure the others will be addressed. I'm not a builder, and this has been the first time I've had any significant dealings with a firm of structural engineers (the firm was selected-by and received their initial instructions from my architect - since I'd decided to leave all of this business to the professionals!), so I wanted to be as sure as possible of the facts before I started raising concerns - apart from anything else, if I go in as (in their eyes) a layman and immediately make a newbie-grade mistake, I suspect that they're much less likely to take-seriously, any other concerns which I might have!


Thanks once again for everyone's input.



p.s. as the graphics aren't visible to people who aren't logged into the physics forums, I've also posted them here:

http://buttersideup.com/files/steel/I-Beam-Bearing.png
http://buttersideup.com/files/steel/Bearing_Case_A.png
http://buttersideup.com/files/steel/Bearing_Case_B.png
http://buttersideup.com/files/steel/Bearing_Case_C.png


Tim.

 

[hokie66]

All the other issues are important, but as you requested, I was just trying to deal with whether or not the bearing plate achieved uniform pressure on the wall. It will not do that in the designed configuration. One way of achieving more uniform bearing would be to add an end plate to the beam which projects slightly below the beam. This would apply a point load to the bearing plate, sort of like a rocker connection. The plate would need to be checked for bending. You could also just provide a smaller plate under the end of the beam between the beam and bearing plate. Same principle, but your beam may still not be long enough to reduce the stress to acceptable limits.

 

[nvn]

TimSmall: If the I-beam flange and steel bearing plate are parallel after load is applied, then the actual masonry bearing stress distribution will be triangular or trapezoidal, extending over a bearing length, we could say, of L1 = L + t, where L = beam bearing length (150 mm), and t = steel bearing plate thickness.

However, I am about 70 % sure that the brickwork allowable bearing stress, Sbra, is an equivalent uniform bearing stress that already takes into account this triangular stress distribution for you. Therefore, you generally simply compute a uniform (nominal) bearing stress as shown in post 20. But I will say, the Sbra I listed in post 20 is for an unfactored applied load.

The current conclusion is, your masonry appears overstressed, even if you use bearing length L1, stated above, and use an unfactored applied load.

Secondly, I think you will find, for a one-story, load-bearing masonry wall, the absolute minimum wall thickness allowed by the codes is 150 mm, not 100 mm. But a better minimum wall thickness would be 200 mm.

Also, if your masonry wall condition is degraded, and/or is not average quality, it might not even meet the Sbra value listed in post 20.

 

[Studiot]

See note 11

http://www.lbhf.gov.uk/Directory/En...ions/19870_Structural_Alteration_Guidance.asp

There are two routes to satisfying the building regs.

The deemed to satisfy route as above where the code provides a figure that should be safe in all cases because of generous margins of safety.

The full calculations route where more accurate calculations or even tests are performed to achieve a more economical solution by reducing the margins of safety to known but acceptable limits.

 

[NateB]

https://dl.dropbox.com/u/1040535/Beam%20Bearing%20Plate%20McCormac-AISC.pdf

Text is from McCormac's Structural Steel Design, 4th ed. The example is from AISC Design Examples v 14.0. It is specific to the US, so it may or may not be useful...