madax,
You do not state what code you are working to ?
Both AWS D1.1 and AWS D1.5 allow undercut but the acceptance criteria is based on specific variables - AWS D1.1 Table 6.1 (7) and AWS D1.5 Clause 6.26
Hope that helps,
Regards,
Kiwi
Madax:
An undercut greater than 1/32" used to be a weld defect according to AWS D1.1, look in the ‘Qualifications’ chapter for undercuts. Depending on the orientation of the undercut w.r.t. the loads and stresses, they can be a significant discontinuity or stress raiser on the surface of the parent metal, all the more so in a structure which is dynamically loaded. Start and stop craters fall in the same category of defects and can be hot spots/hard spots too. You probably won’t find a citation which says ‘Madax shalt not allow undercuts which run in this or that direction,’ you must use some engineering judgement in these matters. But, they are a standard visually inspected defect if to deep. Many welds have some very minor undercut.
Thanks guys, D1.1 table 6.1 (7)(b)allows me up to 2" in 12" as long as it's not transverse to tensil strength , but 2" on a4" weld 1/32" deep seems like a heck of a lot
Cyclic loads: The 1/32 inch depth permitted for undercut is when the undercut is parallel to the primary stress. If the undercut is transverse to the primary stress, only 0.01 inch is permitted. Basically, if your finger nail clicks as it passes over the undercut, it is rejected.
I did not mean to imply that B31.3 would apply to dynamically loaded machine type equipment, airframes, etc.; only to state that welds in severely, cyclically loaded piping (e.g.; out flow from some compressors) are not permitted any undercut. It is also noted that the Owner/Engineer defines whether the service merits the added requirements. The degree or type of dynamic loading was not defined in the query
I’m with Stanweld and Gtaw on this, you have to have some knowledge and use some judgement and maybe several different codes or good text references to make a proper determination in some cases. What do the code commentaries say about the particular situation, what is the intent of the code requirement, why and how is the particular condition detrimental to the design? Then the real experienced engineer pools that knowledge and through experience makes a design or spec. decision. God help us if we get to the point where we can’t do something/anything because we can’t find an exact code citation for some/every crazy condition. And by the nature of the way codes are being written these days, and the way engineering education seems to be going, we are heading in that direction. What we need is more knowledgeable, more experienced engineers who can apply good sound engineering judgement. We really don’ need more stringent codes which further restrict the application of sound engineering judgement. And, we will likely never codify dumb design decisions away.
Come on guys, Dynamic Structures transverse to Primary tensile stress "only 0.01 inch is permitted. Basically, if your finger nail clicks as it passes over the undercut, it is rejected." gtaw got it exactly. D1.1 for general Structural stuff, D1.5 for bridges. Don't overthink the question.
There may be Euro ISO specs with different criteria, but D1.1 & D1.5 have a longer history and they work. Reinventing the wheel is usually a futile excercise. If you want to write a spec different to a historically accurate & workable one, you better have a lot of research and testing data to back it up.
Guys,
The OP stated a "dynamically loaded structure".
What exactly is a dynamically loaded structure ?
I posted a question on another welding forum recently regarding this question and it seems like a good time to throw it out again.
AWS D1.1 2010 references "dynamic loading" three times, once in the Foreword and twice in the commentary.
However, Visual acceptance criteria as per Table 6.1 only lists Static Loading and Cyclic Loading.
My question was would a Dynamically Loaded structure fall under the acceptance criteria of static or cyclic ?
Regards,
Kiwi
"Dynamic Structures" typically are bridges. Good engineers try to keep their structures stiff, to keep the occupants from getting sea-sick. So very few structures are designed to have significant cyclic movement.
Thus, even though D1.1 has always been obscure about their terminology; Dynamic = cyclic
"Dynamic" "Cyclic" This is NOT a situation like "tomato-tomaaaato"
Cyclic loads refer to loads that increase and decrease repeatedly in service. These loads remain in static balance with the reaction forces coming from the structure that supports them. Bridges and pressure vessels must support static and cyclic loads. If dynamic loads are applied to them, you can get into trouble very quickly (eg. Tacoma Narrows).
Dynamic loads refer to forces that cause motion, acceleration, deceleration. These loads are reacted by the inertia in the structure or system (in addition to structures that may be in the way). Examples are springs, dampers, pistons etc. You can also consider dynamic loads to involve a certain amount of kinetic energy transfer, such as a hammer striking an object.
I may be speaking from my own background (which isn't exposed to AWS or pressure-vessel codes very often) but I would warn against mixing up the terms.
