Duke,
Al's suggestion to eliminate the lamellar tearing would also limit distortion. The free floating backing would eliminate the stress. Assuming a normal bevel shape could be prepared by the "buttering" passes stacked in a logical sequence.
No argument there. The backing does tend to curl upward, hence the increase in backing thickness to offer more resistance to curling. The other counter measure is to taper the thickness of the (top side) backing so that it is slightly thinner toward the column flange. That ensures the weld against the column face is no less than the thickness of the weld at the beam flange. The truth of the matter is that it doesn't make a lot of difference if the backing does distort slightly. The backing represents a stress riser where the backing is not fused to the column flange. The connection as shown is only permitted for static loading. If the joint is subject to cyclic tensile loading, the backing must be removed and the root rewelded. Hence the small amount of distortion that takes place isn't a major factor in the transfer of the load from the beam flange to the column flange. If the joint is loaded in compression, no problem. The misalignment resulting from the curling of the backing can be controlled if some forethought is given. It can be controlled to be within reasonable limits that will not affect the function of the connection. Are there limits to the size of the root opening that can be corrected in this manner? Certainly. I've used this approach to correct root openings that approached 1 1/2 inches. All the connections were subjected to UT once they were completed. The welding sequence is important. If the weld is completed in the normal manner, i.e., successive layers are deposited joining the beam to the column, there is an increased probability of developing lamellar tears in the column flange that is loaded in the through thickness direction.
The detail shown, as stated, is a field fix where it is impractical to splice a short section to the end of the beam that is too short.
Let's consider the situation where a short section is spliced to the end of the short member. Consider the amount of welding required to add a short section to the end of the beam. Both flanges must be CJP as well as the web if the welded joint is required to develop the full strength of the original member. How is the integrity of the welds going to be verified? UT or RT are the only two methods recognized as volumetric examinations and both incur considerable expense if the fabricator has to bring in a third party. At what point is it more cost effective to replace the member?
Scott's fix is reasonable as long as the dimension being corrected isn't too excessive. As the dimension that must be correcdted becomes larger and larger, the eccentricity introduced must be considered and the thickness of the connection angles become greater and greater. The fix is only practical for simple framed connections where moment is not a concern. At some point the correction becomes impractical.
Each situation is unique. There are several factors to be considered: the magnitude of the loads being transferred, whether the connection is subject to static or cyclic loads, the cost of labor, whether the fix is to be performed in the shop or in the field, whether the connection transfers moment, etc. I doubt there is a single fix, short of replacing the entire member, that is applicable to all situations. Most repairs are a compromise and few, except for the replacement of the member, is without increased cost, labor, and potential delays. The contractor's dilemma is to minimize the cost (delays in schedule, labor, material, etc.) of the correction without degrading the function of the component. The corrective action, if it is not within the limits provided by the fabrication code, requires the blessing of the Engineer.
The life of a contractor is not without challenges, complications, or risk.
The function of QC and QA is to ensure the fabrication meets the requirements of the project specifications, design drawings, and applicable codes. The goal is to inspect to accept, not to find a reason to reject. The TPI (verification inspector) responsibility is defined by the Engineer. Generally, the TPI serves as the Engineer's eyes on the job. The TPI provides the Engineer with information regarding nonconforming work (this is a good example) so the Engineer can make a reasonable determination of whether the work is acceptable "as is" or whether corrective measures are required.
A competent contractor will alert the Engineer to problems as they arise and propose corrective measures that the Engineer can approve or disapprove. It is generally the contractor's responsibility to propose a "fix" for nonconforming conditions.
In no situation should the TPI (verification inspector) propose a fix to the contractor. Any corrective action that is beyond what is permitted by the code should be "blessed" by the Engineer. The TPI should play no part in the process of developing corrective actions unless it is through the Engineer.
I've seen a situation were the inspection agency, with a P.E on staff, got themselves into a world of hurt because the inspection agency became involved with approving corrective actions without the Engineer's approval or knowledge. The inspection agencie's P.E. held the belief that because he was a P.E. he could propose and authorize corrective work. Ouch!
Sorry for getting off subject.
Best regards - Al