Crack Defects in Welded Structures

GA and yes please start a thread on the subject!

Exactly right...GA for ya!

The domain of expertise is called damage tolerance. It's a bit theoretical and involves a lot of estimating. Any structurally significant item will experience a spectrum of stresses throughout it's in-service life. A stress spectrum of a properly designed item will be low enough that an infant crack will NOT propogate...but inevitably there are parts of the spectrum that spike higher and the crack propogates a small amount. Repeat over and over until the crack gets long enough that there isn't enough material left intact to hold the items and then it starts unzipping a lot faster.

The goal of damage tolerance is two fold: 1. Aid in design to ensure designs are actually damage tolerant so they don't need to be inspected too often (nobody can use their equipment to make money if they're getting inspected...or broken already) 2. to define an inspection interval that allows for the detection of a crack in these critical spots BEFORE it's too late, allowing for a proper repair.

The beginning of a DTA (damage tolerance analysis) assumes a microscopic hairline crack exists at the worst possible spots in order to be conservative. A crack long enough to detect would require repair...if the part starts with an inspectable crack then it has been given a HUGE head start on causing failure in the item.

money, exactly.

For example: aircraft. Thin aluminum structures designed for lots of flying, landing, etc etc. Halfway through their life they get every checked...EVERYTHING. Strip out the interior, stuff someone small into the fuel tank and start crawling around with a your magnifying glass because EVERYTHING gets inspected for cracks. This costs a lot of money AND your definately not making any money with this aircraft while it's sitting on the ground.

This is a big part of the appeal of the 787...since it's all composite the current theory on crack propogation doesn't work since the fibers criscrossing inside the laminates prevent crack propogation. Other things like inter-laminate delamination do happen though and are NOT inspectable visually so that just means the whole plane will need an ultrasound at some point? When exactly? Time will tell once they're flying around en mass.

#4 Is entirely correct

You have every right to be ashamed. I have never done an NDT, but it is obvious to me and to my colleagues that cracks will spread causing many times catastrophic results. Welcome to CR4.

Most any indication other than a crack has acceptable allowance size limits that are usually based on the size of the indication and the thickness of the base metal. When such indications such as in-line porosity, trapped slag lines and other multiple imperfections are deposited in-line or their length is 2 times greater than their width they are considered to be linear in nature. Even though the size of these in perfections either alone or as a group can be well below the reject limits they have a much greater chance of growing into a crack and/or causing a failure and may cause the weldment to be rejected. Another indication that is nearly always linear in nature that has acceptable limits in some codes is lack of fusion and lack of penetration. I see lack of fusion of root bead on either one or both walls of pipe welds to be a HUGE cause of circumferential cracks especially in stainless steel pipe welds around the paper mill. Most stainless steels do not tolerate vibration well to start with and when you have the added stress riser of a wall not being broke down it is a recipe for failure IMO.

I have always seen a crack as more of and indication of weld or base material that is actually starting to fail and thus should never be acceptable or left without repairing IMO.

would you go on a cruse on a boat knowing it had big cracks defects in weld seams ,

that might be why no acceptance limit , been welding sence 1966

The problem is that "big" is a relative thing. The same crack may be acceptable on a ocean-liner, but not on an airplane. The same crack that is acceptable on an airplane would not fly on the space shuttle.

One of the best examples of failures due to weld cracking is the Liberty Ship problem. I believe the cause for the splitting apart of the ships was crater cracks. (Also known as "spider" cracks.) I have witnessed several failures in aluminum welds on truck bodies due to this type of crack.

Before we let this person feel stupid we should explore the API 1104 code. I think his question is legitimate and there is no such thing as a stupid question. I have been in the welding industry for more than 40 years and taught at a university for 26 years. I feel really badly if I don't learn something about welding every day!