Creation of a weld defect to qualify a procedure 2

[JamesBarlow]

We have recently been asked to qualify a repair procedure and we are struggling with one aspect.

The customer specifically asked that we create a weld with a defect in it, then repair the weld to qualify the procedure. What we are struggling with is how to create the defect.

The weld itself will be in a 1 1/2 inch thick plate of SA-516-70 and the joint is 36 inches long. They want the defect to be at the mid point of the weld so 3/4 inches deep.

We have a number of opinions on how to do this but no one is certain that it will really work. While not required from the customer I am going to have the joint RT'd to show the defect and not everyone is certain that the different ideas will show. We have talked about placing some slag in the joint and welding over it, placing a tig rod in or using a SS filler rod.

An inspector will be present to witness what we are doing and will confirm that a defect was present based on what we do regardless of what the RT shows but my curiosity on how to do this want to know.

Does anyone have any experience with this that can offer some advise or ideas? We have plenty of welds that fail that we are not trying to but now that we are trying to make a defect we can't agree on what to do.

 

[gtaw]

Weld the plate with the process you will be using for production welding. At the appropriate location try placing a length of E7028 SMAW electrode and weld over it. As you place the weld bead, let it wash up to the flux covering, but don't let the arc actually impinge on the flux covering. You should have a nice example of incomplete fusion with a serious slag inclusion.

Best regards - Al

 

[tbuelna]

There are calibration specimens for NDI devices that are plates with a weld having a specific size/type of defect. Might inquire how the weld defect in these test samples are created.

 

[gtaw]

Weld defects are not easy for a skilled welder to produce. There is a muscle memory tendency for the skilled welder to compensate for any problems with fit-up or problems with welding parameters. Since you are attempting to produce a defect about mid thickness, you can weld the first half of the deposit using the welding process you normally use for production, i.e., FCAW or SAW. That weld should be sound and free of defects. Switch welding processes at mid-thickness. I suggest using gas metal arc welding using short circuiting transfer (GMAW-S). This particular transfer mode is characterized as a process that produces low heat input. GMAW-S uses low voltage and low wire feed speed. The actual parameters are dependent on the diameter of the bare electrode and the shielding gas used. A typical mixed gas is SG-AC-20%, i.e., 75% argon and 25% carbon dioxide. Typical welding parameters for 0.035 inch diameter for this application could be on the order of 18 to 19 volts and the wire feed speed set to 260 to 300 ipm.

The goal here is to produce incomplete fusion against the groove face on one side of the groove. Since GMAW-S produces low heat input, it does fuse well through oxides and it doesn’t penetrate deep crevices (ropey weld beads) between weld beads. So, use that to your advantage. Heat one groove face to red heat to cause oxidation. A piece of metal bar placed against the opposite groove face will prevent it from being heated and oxidizing. Next deposit stringer weld beads against the unoxidized groove face and deposit successive beads side by side. When the layer is almost complete, slow the wire feed speed and reduce the travel speed so the last stringer bead is bigger than all the previous beads, but do not direct the arc against the oxidized groove face. The last weld bead should not fuse to the groove face and because the travel speed is slow, it should produce a deep crevice between the weld bead and the groove face. Deposit the next layer using the original parameters and deposit a second layer over the initial layer deposited using GMAW-S. Now switch back to the original welding process, either FCAW or SAW, whatever you typical would use and complete the weld.

You should have incomplete fusion against the groove face on the one side. The orientation of the X-ray will have to be oriented parallel to the groove face to optimize the probability of detecting the incomplete fusion. You could also verify the incomplete fusion using shear wave UT with the shear wave oriented perpendicular to the groove face.

Let us know how it works out.


Best regards - Al

 

[metengr]

Most of the calibration standards are notches of various orientation and size produced by EDM and not actual weld defects.

 

[weldstan]

I have used sulfur containing grease and carbon arc electrodes to make defects in plate as you described.

 

[Don56]

Many moons ago I had to do something silly like this. Put some water in the groove where you want your defect. I guarantee you will have porosity.

 

[EdStainless]

EDM is what I have seen used.
We did some using pieces of ceramic rod, but it was too difficult to get the size we wanted.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[ledon]

The quickest and easiest way to create a weld defect to prove a repair procedure is to use a cutting wheel. Half way through the intermediates cut a notch in the weld 1.5 times the depth of penetration then complete your weld. RT will show a lack of fusion in that area. Effectively, you are using keyholing to produce a positive defect.

 

[kingnero]

RT doesn't show LOF, UT will. RT will only show volumetric defects.
Regardless, you'll have a weld defect there.

http://www.fusionpoint.be

http://be.linkedin.com/in/fusionpoint

 

[ll23]

With all respect, i disagree with Kingnero.

See if this material help can bring some help

 

[DekDee]

I also disagree with kingnero

 

[gtaw]

I agree with both DekDee and II23.

Whether RT with detect a discontinuity or not is dependent on the orientation of the discontinuity relative to the location of the source. If the discontinuity is perpendicular to the "beam" of ionizing radiation, the discontinuity may be missed if it does not represent a change in thickness or density. If the orientation of the discontinuity is parallel to the ionizing radiation, the probability of detection is greater.

Simply put, RT is unlikely to differentiate between two 1/2 inch thick plates placed on top of each other and a 1 inch thick plate. Thus laminations are unlikely to be detected using RT. However, two 1 inch square bars placed side by side and a 2 inch x 1 inch thick bar placed side beside the two 1 inch bars is easy to differentiate. Thus cracks oriented parallel to the beam of ionizing radiation is easily detected.

The attached sketches probably does a better job of explaining the subject.

Best regards - Al

 

[kingnero]

I know it's a generalisation, but a pure lack of fusion fault (thus without a void inbetween both parts) and photographed in any direction but parallel with the defect, will not be clearly visual recognisable on a RT.

http://www.fusionpoint.be

http://be.linkedin.com/in/fusionpoint

 

[gtaw]

I'll try to load the sketch again.

Best regards - Al

 

[racookpe1978]

Cut a small dia tungsten GTAW tip into 3x small pieces, each less than 3/8 long (10 mm) long. Cut the "notch" in the weld bead as described above, and drop in the three pieces nof tungsten in a row in the notch.
Then weld over the tungsten "inclusions". Guaranteed to show up in an RT film.

 

[ll23]

Sorry James Barlow, this is the file

http://files.engineering.com/getfil...7-e5db-4cff-8668-41abca04c199&file=TCS-13.pdf

 

[gtaw]

Interesting.

Just a note about using the copper implant; it doesn't take much copper to cause cracks in carbon or austenitic stainless steel. I usually use a "whisker" of copper wire from a welding lead. Only a 1/4-inch length is needed for a single pass weld. The cracks are intergranular and run in multiple directions. While they are useful in providing something for the student to see, they are not representative of either a transverse or longitudinal crack.

Free machining steels containing sulfur, selenium, or even lead, are usually pretty good materials for longitudinal cracks along the centerline of the weld.

Good samples for training purposes can often be found in the scrap bin in a welding school.


Best regards - Al