I've been rattling my two cents around in my head for a couple of days debating if I wanted to throw them in the ring. As they say nothing ventured, nothing gained.
It is my opinion there will come a time in my lifetime when both bend test and RT go out the door.
http://www.olympus-ims.com/en/ndt-application/183-id.209715270.htmlhttp://www.olympus-ims.com/en/ndt-application/183-id.209715263.htmlThat is just the beginning. EFIT, P scans, EMAT, SAFT and other methods currently developed, and in development will provide not only flaw detection, but mechanical properties testing. The theory is actually very simple. Every UT ever performed imparts mechanical vibrations into the weld. What is received are the gross return either by reflection, refraction, etc. What is missing is instrumentation sensitive enough to pick up and map exactly where it is returning from, in what mode, in what angle, etc. Consider how weather doppler radar picks up on air density, reflectivity, etc. Consider how modern Ultrasonic medical scans work.
The only thing holding this back is current processing speeds, and the correct alogrithms.
In theory, a system could map the entire spread of a weld and base material simply by a high decibel ring of the part in conjuction with an array of transducers to pick up and map through processing the mechanical properties of the weld etc. With known standards, a weld that does not match the elastic modulus, and other mechanical properties both gross and intergranular being markedly different from one that was welded correctly and clean.
Throw in EMAT or an advanced variation of it, you will end up with a frankensystem that can do nearly anything. For that matter, there are many attempting to create an AE/UT system capable of listening to the weld as it solidifys, matching it against known values, and determining if the weld is good nearly real time. There are a lot of technical challenges to that, but from all of my research, the theory of it is sound.
Don't think it possible? The same thought was held about cell phones, color tv, computers, and for that matter, the earth being round at one time or another.
"640K ought to be enough for anybody." - Bill Gates, 1981
Enough about the future of it for now.
Without regards to the specific code, I don't see either bend or RT as superior to one another, just as I don't view an apple any better than an orange.
It's two different animals.
As for Shanes question about the rest of the world using RT to qualify a welder, it depends on where in the rest of the world you are.
However; a weld engineer in Argentina once told me 'Paraphrased':We watch the welders during qualification, if they follow a Qualified WPS witnessed by a weld inspector, why waste more money on bends?
I thought about that for a while and came to the conclusion that in theory he was correct, in reality he was wrong. There are many variables that can change the mechanical properties of a given peice without showing up on an RT, and will not be noticable if the weld was not being witnessed as it were made I.E. with a hood on. If you have a hood on, how are you going to be keeping an eye on the voltage/amp readouts etc. Put simply, there are to many variables to keep a direct eye on for any given singular welder/inspector combination. That is the proverbial hole in the theory of RT is better.
The converse can be said about mechanical testing. If the strap is cut in the same place everytime as most codes decree, there is no promise that the welder actually made a sound weld reasonably free of flaws for the full length of the weld. (within the inherent limitations of RT flaw orientation POD, which can be mitigated by CR radiography)
"Again, the welding standard delineates the minimum requirements that must be met. The manufacturer or testing agency can impose criteria that is more stringent than that prescribed by the applicable standard. If you, as the testing agency or manufacturer, determine it is in your best interest to perform a visual examination, followed by PT or MT, followed by UT, followed by RT, followed by bend testing 6 samples instead of two samples, so be it."
I've worked a project that required just that with the exception of two bends, two macros etc, and PT/UT/RT combination. (for the nature of the service, I could understand the paranoia that drove that)
However, no matter who it is, what their opinion is, however educated that opinion may be, the codes are the code. Unless directly and clearly altered by a project specification, it is a mistake to stray from the documented requirements. It is not fair or right to any of the concerned parties to make up the rules as you go. Not to mention it can be a career killer to put on the "field expert at everything hat". My imagination and assumption capability takes a drastic dive during times of inspections during that time.
Therein is my two cents for what it's worth.
Regards,
Gerald