D1.1 Backer Bars

[CTSeng]

Have a building project with CJP field welded flange moment connections where the fabricator is using 2 backer bars one either side of the beam web at the top flange rather than an access hole to allow for a continuous backer bar. I know AWS D1.1 requires continuous backers but the web is acting as a backer and I know a similar procedure (no access hole) is used when welding end plate moment connections. Is there any document exceptions that permit this?

Thanks,

 

[connectegr]

Eliminating the weld access holes and continuous backing is not recommended. As I understand your condition, the flange is prepared the full width. Continuous backing is provided to avoid residual stresses of intersecting welds. Without continuous backing, how does the weld test at the intersections of the backing and the beam fillets? Weld access holes also aid in releasing stresses that occur from weld shrinkage. These stresses will be relative the the flange thickness of your beam and the number of weld passes required. An argument could be made that continuous welds are not required for the actual moment in the beam or the flange tensile force. But, this would be a design consideration, that the erector may not have the information to verify.

Concerning endplate moment connections...
The removal of the weld access holes was the result of research on the ductility of cyclically loaded connections. Some of the test samples developed cracks in the weld access holes, as deformation of the endplates increased. AISC Design Guide 4 provides a specific welding procedure, and uses a backing weld rather than a backing bar. Even in high seismic applications of field welded flange connections weld access holes are required. And specific detail requirements are given for their geometry and finish.

http://www.FerrellEngineering.com

 

[connectegr]

Another consideration...
Is the procedure your fabricator recommends AWS prequalified? D1.1 does require continuous backing or backguoging for CJP welds.

http://www.FerrellEngineering.com

 

[connectegr]

CTSeng
Since I have answered this question a couple of times recently. I asked Duane Miller for his thoughts. Mostly his comments followed my responses above. He had an additional comment that I did not consider. "For the top flange, the access hole not only permits a continuous backing, but it also allows for a "defined" geometry of the groove weld in the center of the flange width."

Hope this helps.

http://www.FerrellEngineering.com

 

[CTSeng]

Thanks, this generally agrees with my understanding of the continuous backer requirement. I just seemed similar to the weld exception opposite the web in the end plate design guide. I have heard that the 2010 AWS has some changes to the continuous backer requirement but not holding my breath.

Thanks.

 

[connectegr]

CTSeng

Duane is the chairman of the D1.1 committee, and he did not offer any suggestions to eliminating the continuous backing bar. He also agreed that the weld access holes provide secondary benefit, such as stress relief, in addition to simply allowing access for the backing.

http://www.FerrellEngineering.com

 

[connectegr]

Just another observation.

The Design Guide 4, welded endplate without access holes, has a specific welding procedure. The welds are applied in sequence to minimize the residual weld stress. A backing weld is used for the CJP flange weld and a zero root opening is used. And when UT testing is done at the center of the weld (at the web) indications of incomplete fusion are frequently seen. But, ignored based on the overall design of the connection.

But, for field welded connections the shop welding controls are not available. Specifically a root opening is required due to erection and fabrication tolerances. Otherwise PJP welds could be used in the field not requiring backing, etc. with a R = 0. Typically a TC-U4a weld is specified in the field with 30 deg prep and R = 3/8". The beam web is also "set-back". Therefore, in your condition no weld is provided for the length equal at least 2 x k1. (see attached) If the flange force can be transferred with the provided effective weld strength, then your condition should be acceptable. The welds will be concentric so, there should not be any additional bending stresses.

http://www.FerrellEngineering.com

 

[connectegr]

My point is that in the field, the web is not acting as backing. The entire beam is detailed with a set-back for erection. Therefore it is not an issue of not AWS approved it is more of adequate strength.

http://www.FerrellEngineering.com

 

[connectegr]

Something else I just noticed. In AISC 13th Edition p12-9 #8, Design Assumptions for endplate moment connection, CJP welds at web fillets should be considered as PJP. Thus, accepting the incomplete fusion at the root.

But, still not an option in the field.

http://www.FerrellEngineering.com