Testing the tensile strength of a weld - GIF

[Jon]

Testing the tensile strength of a weld.




Previously:

Welding up close - GIF
Cold welding GIF
Bolt with built-in tensile strain gauge - photo and patent
Propane tank controlled explosion test - video
Boeing 777 wing flexibility test to failure - video
Testing a 500 PSIG steam safety valve - GIF
Heavy tire controlled explosion test - videos

[McDesign]

Yep, that would be that "heat-affected zone" we hear so much about.

[Quinton 357]

Hi. I wonder what they used to pull it apart! So we can do our own testing.

[suther51]

https://en.m.wikipedia.org/wiki/Univ...esting_machine

Hope this helps Quinton 357
Eric

[Quinton 357]

Hi Eric.
Yes thanks. I thinking a bit more homemade.

Thanks . Good read.

[Frank S]

To find the true tensile strength of a weld first a coupon or several coupons are cut from a welded test sample then the surface of the weld is milled to match the surface of the base metal with no scratches or gouges in either side. the test shape profile is smooth and consistent on all sides The sample is often reduced dimensionally to match width to thickness at the weld the reduced area will have a 3 to 4 times length to thickness/ width ratio allowing for the reduces area to extend well beyond any possible heat effected zone.
A simple tensometer can be made with a lever& fulcrum and 2 clamps that are pivot mounted so as to not induce any lateral directional forces the just calculate the ratio of the force applied to the handle end of the lever against the length of the moment arm. sufficient travel of the pulling action of the moment arm needs to be long enough to create failure

[Quinton 357]

Awsome. Thank you.

[KustomsbyKent]

Another, somewhat simpler method is to do a bending test. It's important to prep the sample as Frank described, where it is ground smooth on both sides and ends. Then you bend it over in a vice. If it breaks, you get to start over. If it bends without breaking, pass!
We did this in our HS welding class to test your welding skills on a vertical up weld, multi-pass (root and cover passes).

As evidenced in the video that Jon shared, the joint failed right at the weld, in the HAZ, and may have been made worse by a slight undercut on the weld bead causing a stress riser in that area.

[Frank S]

Yes Kent is correct however when doing the bend test the coupon or sample is not otherwise reduced in size. There are as many ways to test a weld as there are welding applications and types of welds. Both destructive and non destructive testing procedures are preformed for the fabrication of anything joined by a welding process. the destructive methods are for qualifying a welder the non destruction methods are for the finished weldements.
I like an impact method for single sided but and fillet welds as well as simple bend resistance for impact just weld 2 pieces together either butted at the ends of positioned as a "T" make the appropriate weld to 1 side only then secure one of the pieces so it cannot move then slam it with a hammer from the opposite side to the weld if it bends keep slamming it until it folds onto itself. Then slice it open providing it did not break and examine the grain structure of the weld and heat effected zone. Other ways slightly more scientific is to use a hydraulic press with a pressure gage this will tell you at which point the weld begins to fail.
There is also a test called the nick test. where you make your sample according to the recommended welding fabrication procedure plane the sample on both sides then saw into coupons after saw a nick into the weld on one end of the exposed weld place the coupon on 2 supports with the nick down press from directly above When the weld begins to tear the test has reached the point of failure note the pressure then calculate the square of the weld (thickness and length) multiplied by the actual pressure of the ram area of the press exerted on the test sample this will give you a good indication of the tensile strength of the weld as well only this test can give a false reading if the length of the weld is significantly more than the thickness.
Another quite simple test is to clamp the sample in a vice then fit a socket over it and use a torque wrench to twist the sample until the weld fails. All of the above tests are done with samples having a relatively flat surface segments of pipe may be used as long as they are proportional in size to the diameter of the pipe. usually 1/10th to 1/6th the diameter or 2" wide which ever is less.