Composite beam (H beam encased in concrete) connection to top of concrete buttress? 1

[Redacted]

Hi there, I've been tasked to design a solution to fix this problem (see attached images)

The problem:
1. Existing concrete beam-column on top of the buttresses has experienced corrosion of the inner steel member that has expanded and caused delamination of the concrete, which is now falling onto the motorway beneath it (The existing has lasted about 50 years).

The road is closed until this issue gets resolved.

2. Historical drawings of the road are not available and they are expecting a quick solution. Therefore conservative assumptions need to be used.

The solution:
I thought about potentially saving the beam by using some bolted channel sections but not sure if those would be adequate in compression and the quality of the beam concrete seems poor, so I will just replace them, which the client is happy to do.

I am thinking of using a concrete-encased steel H beam. I am assuming the buttresses are in adequate condition and plan to use those for the new beam. I am planning to replace the beam with one of the same cross-section dimensions so that the overall weight of the beam is not higher than the original.

I am wondering how I can connect this new beam to the buttresses so that the connection can effectively take the axial compression loads etc. Essentially I would like to know what the most efficient connection detail to transfer the force from the beam to the buttress is.

I'm roughly thinking of some type of system using post-installed anchors in the buttresses, bolted to the beam flange and concreted over?

I'm not sure if there is a common connection detail for something like this or not. If anyone can provide assistance or point me in the direction of helpful advice applicable to this design, that would be greatly appreciated.

Beam-column has a clear span of about 16', with a depth of 1'8" and width of 2'.

It's a strange structure as the steel inside the beam-column appears to be an old train track?

I am also assuming that these beam-columns are acting as lateral bracing for the adjacent retaining walls.

 

[Enable]

My apologies for being out of the discussion for too long. Every time I've wanted to come back to this I've found myself into a bottle of wine (or two) and I'm afraid I would have been a liability out here. It's been a week and it's only Tuesday (also it's September first today...which really means it's December. Contractors will understand)!

SlideRuleEra has covered much of my own thoughts and I have not a great deal to add. But to increase the shear resistance at the repair area interface I would consider adding dowels at the joint (use HY-200 and fully develop them, else chip deeper than required on the strut side for half depth of the wall and anchor bar with hooks / canes below the shear plane). Also, I would consider adding a W-section onto the end of the beam that extended below the repair area (instead of just a plate onto the patch area). Size the beam for the induced moment / deflection and it'll be a pretty redundant system if you couple it with the doweled repair plus headroom reduction is only at the sides and not across the majority of the member.

 

[SlideRuleEra]

Redacted - Glad you are learning, projects like this are excellent opportunities for "outside-the box" approaches. I mentored six young engineers, one at a time,on real projects like this, so they would get the career-broadening experience. Back to business, we have a lot to cover...

Beam Size: I believe you can do better than a W12x45, for several reasons:

1) You have a high corrosion environment (even considering HDG or high-performance coating). All (without exception, to the best of my knowledge) W shapes have "thick" flanges and "thin" webs to optimize structural properties for use as beams. For W12x45, flange thickness = 0.575", web thickness = 0.335".
In our high-corrosion railcar coal unloading structures, it is always web loss of thickness that compromises the members.
For all HP shapes, flange thickness = web thickness (to maximize service life from corrosion when used as piling). Yet, HP shapes work very well as beams, just not as efficient (per pound of steel).

2) I had in mind, say an HP14x73 (flange thickness = web thickness = 0.505"). The "larger" footprint (14" x 14") is a big help for baseplate design compared to the W12x45 footprint (12" x 8"), more on that later.

3) Don't lower overhead clearance for any reason, the strut will be more likely to have impact from an overheigth vehicle... it happens. The following applies even if clearance remains unchanged. When it happens, the strut needs to have substantial Y-axis strength to at least survive (although it probably will have to be replaced). How many times have the existing struts been hit.. you don't know... they have significant Y-axis properties. Compare Y-axis properties of a W12x45 versus an HP14x73... not even close.

4) There is a subtle reason, too. The traveling public is accustomed to seeing substantial struts. Even if a "small" strut is fully structurally capable, the new "toothpick-sized" replacement will raise the (incorrect) question from the public... Do "they" know what "they" are doing?

Baseplate and Hilti anchors:

1) Proposed baseplate size of 16" x 16"... for a W12x45, 2" cantilever top and bottom, 4" cantilever each side. Use those same cantilevers on an HP14 and base plate goes to 18" x 22". Larger size baseplate should help on the existing concrete shear problem you are addressing.

2) Proposed 5/8" thickness... pretty "thin" for cantilevers of up to 4". A baseplate needs to be rigid to apply a uniformly distributed load to concrete. Higher the loading, the more important this becomes. To make it rigid, the BP can either have elaborate stiffeners or be thick. Rigidity of a plate increases with moment of inertia (I). "I" increases with the cube of the thickness. I'm thinking more like 1" thick (a 1" plate is four times more rigid than a 5/8" plate). There is also a pretty long span between flanges of an HP14 (over 13") where rigidity is important. Do not take my "educated guess" as fact, do your own calcs, to see if the cantilevers have reasonable deflection and determine baseplate thickness you need.

... I'm going to take a break now, back with comments on Hilti anchors, Neoprene and the troublesome reaction anchor later.

 

[Redacted]

@ BridgeSmith, thanks - the size of the existing strut is quite large and isn't a standard section. The encased steel is most likely an old train track. Although that is a good point. When I estimated the load I tripled the factored loads due to the uncertainty and still ended up with quite a low axial compression value (see attached rough calcs). I'm not sure how the original strut was sized/why it was that big, or am I missing something obvious? My only concern is my assumption is that the base plate will not be applying out of plane moments/loads to the beam. There is an eccentricity from the centre line of the baseplate to the centre line of the beam, which is 58mm that I am considering and using for bending moments in the major axis. However, I haven't considered eccentricity for out of plane bending, as I am assuming there is none.

@Enable thanks, and no problem lol a drink wouldn't be too bad right about now but let me finish this design first. Ok, including dowels were what I was thinking earlier. Mind elaborating a bit on the addition of a W section on the end? I am assuming you are referring to the left connection for strut 1? If that connection maintains the head height I think it would solve a lot of issues, although I am not familiar with that type of detail, are you proposing something similar to what I sketched in the attached document. If you do mean what I sketched, I think it may work, the only difficulty is that it somewhat complicates the load path. The weld would need to take the applied shear from the axial compression.

 

[Redacted]

@SlideEraRule

I didn't see your response when I posted just now but thanks. Your comments are helpful as usual and I agree with the section loss comment and general perception. Cost increase should be marginal. I will use the larger section.

That's true about the base plate. Part of the HILTI Profis calculation is that it assumes a rigid base plate, which stipulates engineering judgement must be used to determine. I will run some hand checks on a 1" plate, as I would need to ensure that prying of the plate does not occur to simplify the analysis.

I look forward to hearing your further comments.

 

[SlideRuleEra]

Redacted - TIMEOUT

I reviewed your recent posting of wall load cals, they are incomplete and likely underestimate strut compression loading... significantly.

The values you have arrived at (127.6 kN Dead Load and 25.9 kN Live Surcharge Load) are not strut loading. Instead they are "unit-loading" (kN/Meter of Wall Length). Each strut will have a tributary wall length that it supports. You have to determine that tributary length... 3 meters, 6 meters, 10 meters...?

After tributary wall length per strut has been accounted for, compression force in each strut has to be calculated using statics. If the strut was located at the same elevation as the retained soil, strut compression would be 1/3 of tributary wall Dead Load and 1/2 of tributary wall Live Surcharge Load.

However, the struts are lower than retained soil elevation. Cals for this are not difficult (you already have posted enough dimensions to do this). Depending on the tributary wall length, strut compression may be >> 154 kN. This could go a long way towards explaining the large exiting strut size and fractured Strut #1 reaction anchor.

Have you considered discussing all assumptions & calcs with an experienced engineer at your firm?

 

[Redacted]

Thanks SRE, yes you are correct, the tributary area may be the reason for the larger original size. I just updated the calcs and the HP14x73 is still adequate at less than 50% utilisation in combined loading.

My firm does internal checks of calculations and yes I plan to discuss all assumptions with another engineer at the firm to make sure everyone is in agreement. I want to have a decent detail/schematic before then to have for discussion.

Thanks for pointing that out though!