Welding Rebar to Steel? 1

[SRO]

I have a 14"x14" concrete column supporting intersecting steel beams. This needs to be a moment connection from the column, through the steel top plate, and to the steel beam.

In an effort to avoid the clutter of anchor bolts to connect the plate to the conc, I wanted to bend the rebar 90 degrees, and weld it to the steel top plate, then extend the other end as long as needed for development length.

I've checked out thread507-55969 "Welding of Rebar" Which said that E309 electrodes should be used, but they were welding rebar to stainless steel. My case is rebar to regular steel.

I know in general, welding rebar is not good practice, but using this construction technique might make a common job we do more economically feasible.

Does anyone think this is a bad design, if so why? Should I weld threaded rod instead of rebar?

 

[CWIC]

SRO:
Here in seismic zone 4 - CA, tying in existing framing to new grade beam cages, etc. is very common for retrofits. Rebar to HSLA steel is performed daily for embeds as well.

Most EOR's and building depts. are now requiring procedure (PQR) qualification to confirm the compatibility of the materials. Previous WPS's that have been qualified typically are not permitted. Procedure qualification is not an expensive undertaking although some contractors whine and complain that they are requested to do so.

 

[KarlT]

If this is new construction you should specify a grade of rebar that is weldable. (I believe the normal non-weldable grade of rebar has a higher amount of carbon which makes it more brittle for welding applications)

In addition, be careful about welding a bar too close to a bend. Typically it is recommended to terminate rebar welds a minimum of 2 inches or 2 bar diameters from the start of the bend.

Threaded rod would give you less strength for standard A307 bolts or grade 44 ksi steel so I would use 60 ksi rebar steel. Anything higher strength could cause embrittlement problems.

Hope that helps!

 

[Taro]

Consider using deformed bar anchors that are butt-welded to the plate with a stud gun. The quality is more reliable.

 

[PXC]

SRO Your comment: "I wanted to bend the rebar 90 degrees, and weld it to the steel top plate, then extend the other end as long as needed for development length". If I understand you correct, then this is a bad detail. When the load is applied, it will simply "unzipper" the welds starting from the bend. A better detail is to drill a hole through the plate and insert the bar and weld both faces of the plate. It is possible to insert the bar only part way through or say 2 x the plate thckness.

 

[Bbird]

I am not entirely convinced by PXC suggestion is necessary. Sure it seems to be a more positive anchorage to weld a bar to pre-drilled holes but the strength essentially comes from the weld. A long fillet weld, mostly on both sides of one leg of a "L" shape rebar should have more weld than that around the circumference of a couple of holes. One leg of the "L" in this case could be fully welded to the steel member. The weld should be strong enough allow the bar fail first before any unzipping or any detackment from the host steel member.

Rebar is difficult to weld and the heat process of the welding process can degrade the load carrying capacity of the steel. The former can be overcome by experienced and certified welders and selection of a suitable weldable grade for the rebar. The latter can be offset by increasing the development length, say in the worst case the high yield degrades to mild steel. There is no technical reason why SRO's proposal should not work in practice, as some posts here already substantiated.

I have doubt on the end-on butt weld application capable of developing the full tensile strength of the rebar. I am aware of the stub type connection is good for shear and widely used in composite construction. It doesn't sound right here if the bar's primary function is to resist moment with axial tension.

Can an end-on butt weld by a stub gun, which I presume the weld is formed by impact and electrical current, be pulled out by failing only in the rebar body and not the weld?

 

[Koz]

Hi all,

I'm a little surprised no one has mentioned the governing welding code here, AWS D1.4.

It has sections on allowable stresses, allowable reinforcing steel and allowable filler metals. For the flare-bevel joint described above, the Code will only allow stresses of 30% of the nominal tensile strength of the filler metal, similar to a fillet weld in LRFD structural welds. Effective weld sizes must be determined per Code similar to flare bevel grooves joining sections onto pipe or tube steel.

I'm not sure of the design, but the reinforcing steel is 60 ksi YS and the structural shape is 50 ksi YS. Moment frame connection are typically complete penetration welds to develop the full strength of the materials. Code requires (I assume A706 "weldable" rebar) 70 ksi filler metal to match the lower strength structural steel. The Code makes no distinction between A706 and several other reinforcing steels available. They are all "weldable" when following a proven welding procedure in conformance with the Code. My tests have shown tensile strength to be at least 125% of yield strength is easily attained for welded A706 and A615 gr 60. But it is your money.

Remembering the structural steel and E70 class electrode, automatic stud welding of DBA's would accomplish the same thing. They, too, are complete penetration welds. Preproduction bend tests proof load these welds by bending a sample of the fabricator's daily run to 30 degrees out of normal. That's a pretty good indicator of weld soundness. These welds are governed by AWS D1.1. Buy that Code, too.

And under a moment load, won't your concrete crack out around the bar before the weld or even the reinforcing steel yields? I don't think development length governs here. We're not just pulling the rod out of the concrete. Just a thought.

The hold back of two rod diameters from the tangent of the cold bend is correct.

Running the PQR is mandatory to prove the fabricator's welding process. I don't recommend running procedures for every job though. Some jurisdictions require recent tests. My thoughts are once the appropriate PQR has been run, it validates the supplier's methods. Nobody cheats a weld test. You have to audit the supplier to verify continued compliance with the qualified welding procedure.

And don't let your fabricator use some other fabricator's welding procedures! It's your money and your fabricator's should be sophisticated enough to understand the control of their welding program. If they aren't, run them out of business. There has to be at least a modicum of honor among us thieves.

Fun!
Koz

 

[CWIC]

D1.4 is referenced and discussed in the thread noted by SRO.

 

[Bbird]

Speaking from the projects I have involved (all outside USA) the standard site practice is to disallow welding of rebar on site unless the situation is unavoidable. Welding in a fabricator yard is a totally different story because the work will be properly prepared, executed and experienced and qualified welders are available.

Unless SRO's design is prepared and included in the design of the structural shapes allowing the fabricator to implement the welded bars before the delivery of the structural steelwork, the other alternative would be to put the detail of the welded rebar on the reinforcement drawing leaving the contractor to carry out the welding on site.

May be SRO should make it clear which route he is likely to embark on. It is just not a standard practice to ask a steel fabricator to weld steel reinforcing bars routinely on structural shapes even though there is no technical reason not to and I am sure many such cases have been done in the past.

I seldom deal with structural shapes and welding myself and wish to understand a bit more from Koz's comment.

Would I be correct in thinking as the code allows only 30% of tensile strength from the filler material in the welding with a rebar this effectively requires a designer to specify an equivalent of 330% of rebar cross sectional area for the weld. Also the end result is that the full capacity of the rebar can be utilized in the design without any loss in the welding process, presumably obtainable from any site application as long as the proper procedure is followed and experienced and qualified welder is used (I am leaving out the problem of achieving the right facilities on site as a separate problem).

I am not sure why development length is irrelevant here. One leg of "L" bar is anchored to the structural shape and Koz's remark convinces me that the anchorage can be achieved by designing to AWS D1.4. The remaining leg is cast into the concrete column and surely the moment in the interconnected steel shapes cannot be transferred into the concrete column without the embedded rebar able to develop it full strength. Concrete cracks out around bars only in compression and the reinforcement problem here is mainly tensile (Because if needed we always can put rebar in the compression zone with the amount equal to the tension zone to beef up the combined compressive resistance with concrete to avoid a compressive failure).

Finally there is a potential gap to be bridged in our understanding of joining steel structural shapes with reinforced concrete. This is because, unlike steel which behaves homogeneously in both tension and compression, concrete is an inhomogeneous material not permitted to have strength in tension (or its small tensile resistance is ignored) in the design. Thus under a loaded condition the tension side of the neutral axis has no concrete but in the analysis of the structure it is acceptable to use only the full concrete section for resisting moment and all reinforcement is ignored. In other word in the frame analysis we always use the second moment of area of a RC beam not corresponding to its service condition. Hence there is an error in the analysis. Although many codes also allow cracked sections to be used but this is rarely done in practice because the uncracked (full) section approach works well as long as every member of the structure is reinforced concrete. We all know moment distribution is based on the relative and not the actual stiffness.

However if we marry up structural shapes to reinforced concrete beams and columns the established method of design may need to be revised, like using the actual stiffness of the cracked section of the reinforced concrete may be necessary.

I would welcome any comment from other engineers to my last point.

 

[Koz]

Hi all

To Bbird's question regarding weld area.

I think your numbers are right. AWS D1.1 and AISC LRFD set the same limits noted in my post for fillet welds of structural to structural. You're have to have fillet welds long enough to resist the allowable load whether they're 3/16" or 3/4". And then you can only use 30% of the filler metal UTS. This is standard practice. Anyone who designs a ledger angle runs into this same consideration. It gets slightly more complicated for a flare-bevel groove since you have to consider the "effective" throat of the weld. Look at your AWS D1.4 for accurate information.

The loss of strength because of welding that you speak of is already considered in the 30% reductions and the load resistance factors of ACI. And the weld is stronger than the reinforcing steel. Sound direct butt joints always fail in the reinforcing steel with large amounts of necking of the bar observed. The structural steel has the least strength of the joint and it is your limiting factor.

Regarding the welding of the reinforcing steel, the steel doesn't care where it is welded. Shop or field, daytime or nighttime, California or Katmandu (base material quality aside and don't weld when it is less than 0 deg F or in the rain, etc and read the Code) it doesn't matter. Follow Code parameters and the weld joint will pass the required mechanical tests. Done deal. If someone can show me why the steel is sensitive to the geographical location of assembly please enlighten me. Otherwise, this is just noise.

Code parameters include proper preheat, proper cleanliness, proper joint preparation. These parameters are generally performed by non-welders or overseen by supervisory personnel that have no understanding of welding Codes or requirements, so they cut corners when no one is looking. The mundane aspects of setting the proper power source settings and having the skills to manipulate the torch properly, all controlled by the skilled welder, are only part of getting a good weld. If the welds are failing, it is because someone is cheating. It is not due to the welding processes or the base materials.

Of course your comments regarding the tension side versus compression side of reinforced concrete are correct. The reinforcing steel takes the tension. And designs frequently consider the resistance of the section using the cracked moment of inertia.

But a seismic or ballistic event is not simple tension nor compression. It is acceleration more like waves. So that hard reinforcing steel bashes around in the concrete. Eventually you just have a steel rod sitting in a dust filled hole. That's what I think, anyway.

What do you think?
Koz

 

[CWIC]

Koz:
The "noise" you refer to is not based upon geographical location, only regional practices and standards. Shop welds typically have better control over welding parameters, environmental conditions, favorable part positioning, etc. As I travel the planet, often there is no QC in the field (again, a regional standard).

Please take the time to thoroughly read thread507-55969
Welding in CA vs. welding in Katmandu? I'll put my $ on the welds performed locally.

 

[Koz]

Hi all,

Once again we are talking practice, not material or process. I have read the referenced thread and it is built upon the same confusion. While there is some good technical welding data there, it is also surrounded in anecdote and prejudice concerning the capabilities of today's erectors and fabricators.

There is a lack of welding training in our country. I judge by the CWIC appelation that you have passed the AWS certified welding inspector examination. Did you find any questions on that exam concerning the inspection of reinforcing steel? That is because reinforcing steel is relatively easy to weld and we figure everyone will read the Code and do the job right. It's not like a moment frame, afterall. And rebar gets buried in concrete. Who's going to know? Our industry continually overlooks the control necessary to do a good job. And with the carbon equivalents as high as we find in reinforcing steel, training and verification are critical.

These are not, however, material or weldability issues. These are training and implementation issues. Reinforcing steel is very weldable. If you trust a welder to weld up a seismic resistant moment frame or power piping or an ammonia containing pressure vessel, you can trust the welder to join reinforcing steel if they are aware of the requirements and necessary techniques. Many of these same folks also carry certs for D1.4.

If I have a Bechtel or a Fluor running a job in Katmandu, I know the expertise is there to do the job right regardless of product form or chemical analysis. If Ma and Pa Kettle Welders-R-Us is doing a job in San Bernadino County, I think I would impose continuous inspection, as IBC permits, because I can almost guarantee they will botch it. But they would botch handrail, too. Pride of Craftsmanship doesn't recognize geopgraphical location, either.

To fix a problem, we have to define the problem. The problem is not the material or welding process, regardless of where that operation is being performed. The problem is in the flow-down of knowledge. This will only get worse as we continue to go off-shore for our welders and demonize what little welding training is left in our country. I expect our engineers to demand excellence and enforce it. Ma and Pa Kettle cannot be permitted to set the quality standards of our designs because we choose to save a nickel on inspection. Get those bums off the job and hire properly trained welders. Welding inspectors can help by pulling the certs of shoddy welder's, too. Inspectors are permitted, even expected, to require welder requalification if a specific reason can be found as permitted by AWS Code.

Let's not confuse the issue by spinning personal war-stories about shoddy workmanship. The data is anecdotal and not scientific. There are still plenty of good welders and contractors out there. Reinforcing steel is readily welded, using common welding processes, and common welding techniques. Implement the Code, that's all. Upgrade our industry, don't degrade it.

Koz

 

[Bbird]

Speaking from practice I have to say the posters here are all very experienced and knowledgeable in their areas.

To enforce a good welding QA and QC on some sites can be more than we can chew and it may not be possible to achieve the desired standard if the amount of welding is small. Believe me if you have to be driven 6 hours in a 4x4 on dirt tracks into a no-man's land somewhere in a dark corner of Africa you make do with what you have got.

In America or Europe it is totally different because the welders are formally trained and can speak our language. No engineer want to degarde the engineering standard and we all want to push as far as we could without being chased off site.

 

[CWIC]

Well, I'll try one more time to clarify some of my replies and by others.

Repeat from the original thread noted by SRO:
"I would not reject the idea of rebar welding as it is commonly performed (at least here in So Cal) daily with satisfactory results."

"I respectfully disagree that welding of reinforcing steel is not recommended."

In my first reply to SRO above, I also note welding of rebar is quite common.

I feel more is being interpreted regarding the replies than what is clearly stated. I am encouraging the welding process for joining the rebar to the carbon or high strength steel originaly noted by SRO.

The "war stories" are documented facts noted during QA or QC inspections where the workmanship or technique performed in the past were less than adequate and not in compliance with the code. Yes [weldable] rebar can be performed satifactorily by qualified and experienced personnel, but workmanship and technique are paramount.

A few more facts:
1.) Not all jurisdictions or agencies use the UBC/IBC. Special inspection is NOT always required.
2.) NO special inspector has the authority to "pull" a welder performance qualification. Qualification or requalification of welding personnel is the contractors responsibility. QC or QA can only make the recommendation.
3.) If organizations like the ICC, SEAOSC, etc. believe these unscientific anecdotes are just that, then there are a lot of building officials and engineers [who are my clients] very confused.
4.) I am a SCWI, there are almost 30,000 CWI's, I am one of a couple of hundred Senior Welding Inspectors.

No harm no foul, just stating "what I think" which is what SRO asked way up on top.

 

[Koz]

Hi All,

Thank you for taking the time to clarify your responses. In light of that respect I will try to keep this one short, rather than get on the soap box,again.

I asked Bbird about the reinforcing steel cracking out its developed length during a seismic event for a moment frame. I've gotten no response. They are conducting tests at our local university on full-section models of what I'm assuming is a similar joint configuration. The concrete columns are about 18" x18", but I'm wasn't really paying attention. The rebar doesn't yield, maybe a little, but the concrete surrounding the bar turns to dust. And I think they are getting some cracking of the beam, but nothing catastrophic. They are developing wrapping these joints in carbon-reinforced sheets to ensure the performance and reliability of these connections. They're getting a patent on the wrapping process, et. al. So what do you think about this connection?

To CWIC: Since we're throwing certs on the table, I'm a SCWI, too. But I was also a 6-G welder-fitter for Chicago Bridge & Iron and journeyman'd with Ironworkers in my wild and crazy youth. And as I work to finish my ME in Civil Engineering, I find science is much more valuable than war-stories for promoting good engineering. The initial question regarded welding rebar to structural and a confusion concerning deformed bar anchors. These are easy welding problems. Codes recognise that. You made it sound difficult for ancillary problems not related to materials or process. Problems common to all operations including NDT and concrete, I will add.

I will love to continue this discussion in another thread, because I think we have gotten way off track. But I think it is a very important discussion. Our engineers are not being taught key issues in welding. Even the best California structural engineering firms are struggling with how to effectively control the welding going on in their projects. And the engineering schools are not giving these folks even the rudimentary tools to ensure quality. But all of the codes make it the sole, legal responsibility of the engineer to develop quality plans and approve deviations from Code. It should be our responsibility, as SCWI's, to help these folks with the best data available. So let's move our discussion to a new thread in this discussion group so we don't continue to bore the concrete guys. I'll call it "Zen and Welding" and we can devote it to the gap between Code implementation and actual practice. I'll kick it off by commenting on your "facts" there.

I'm still interested in the development length, though.
Koz

 

[Bbird]

Koz

I thought your were just commenting on the subject but now you have provided more information.

The joint connection you described has obviously failed by compression in concrete. The design can be arranged so that the rebar fails first by using less reinforcement so that bar yielding has to take place and provide warning before a collapse. The failure you mentioned has nothing to do with the topic here because both concrete and rebar are able to develop the full strengths. By not having an adequate development length the column can plug out of the joint cleanly leaving just holes in the concrete. If concrete turns to powder then it proves that the steel bars have been properly gripped with sufficient lengths embedded inside concrete. The joint proposed by SRO works satisfactorily but fails at a higher load.

My reference on the need to bridge the gap of knowledge on marrying up steel shapes to reinforced concrete members implies engineer can make wrong assumptions in the stiffness of concrete by just applying the standard rules of a reinforced concrete code. I have already mentioned that tensile area of concrete cracked under load and the stiffness will be reduced, easily by more than half. Thus in your experiment the steel shapes are a lot stronger than the poor old concrete and it is no surprise to me the latter turned to dust because it wasn’t a balanced design. The mode of failure can change in the joint you described by just enlarging the size concrete member resulting a more even failure between concrete and steel shapes.

I have designed carbon fibre wrapping to strengthen the load carrying capacity of existing concrete works. Aramid fabric type material like Kelvar is better for resisting impact and to work with on columns and people have been doing it for years. The wrapping forms a confinement to the concrete under compression and improves the axial capacity. One can achieve the same thing by casting the concrete column with an external steel cylinder as a permanent formwork.

Good luck to the patent you mentioned.

 

[Koz]

Thanks Bbird,

I'm a passive observer on this experiment and the resulting patent (I wish I was part of that team!). Thanks for filling in my gaps on the steel to concrete and development length thing. I've lost track of the experiment, but it is very likely they designed an unbalanced condition to force the concrete to crack and "plug-out" They are proving their carbon-filament wrap afterall. Interesting.
Koz