[quote=... Or should I believe a friend of my Dad's, who's a high-pressure pipelines expert - he says 8 out os 10 bursts are caused by hidden defects in the metal, which are enhanced by accumulated stress, and while the burst will probably coincide with a rise of pressure over everyday level (e.g. a slug or goose load), this pressure rise is simply a trigger, not a cause, of the burst. [/quote]
I assume he's talking about his experience after examining bunches of pipeline failures. The pipe used in his biz is likely made using good quality control techniques, yet he is telling us things still happen. Often, these 'hidden defects' are very small inclusions of contaminants. Some that I've seen were less than .003". I've seen many airplane parts made of the best materials available, using the some of the best quality controls, still have failures due to contaminant inclusions combined with high stress and repeated loads (fatigue).
One very important point, these inclusions require high stress repetively applied to generate a crack and grow it from the inclusion. How fast the crack grows is related to the stress levels, metal alloy, condition (hardness), oper temp, etc.. A crack that grows with repetive loads, then fails, can often reveal itself in the fracture face as 'striations'. Each striation is indicative of individual load cycles. The striations are usually extremely small, requireing magnification (no, usually not just a magnifying glass) to see.
O.k., that's it. That's about as far as I can take this with my single Metallurgy 101 class from college.
p.s.
But, to your acquaintance's comment, yes failures often initiate at 'defects'. In good designs, it is common to assume a defect is in the part at a high stress area. In airplane structure, we assume the defect is usually in the .050-.060" range because this is high probability "detectable threshold" we use for our NDI methods. Think about that for a while. If we use .050" because it is the smallest defect we have high confidence we can detect it, a shotgun barrel would be assumed to have a hole thru the thin section! That says to me that .050 is not good enough detection for a shotgun barrel. Heck, how many barrels do you suppose even got a NDI, ever? O.K. I'm done...well almost.
Yes, most failures originate at flaws/defects/inclusions. But my point is; good designs have considered the defect as existing in the part and have designed in sufficient margin to still do the job. Is that the case with barrels? I dunno, I'm not in that biz.
Last edited by Chuck H; 04/02/07 06:16 AM.
