Weirdest Failures

[swall]

To celebrate Friday the 13th, I thought I'd start a thread on your weirdest metallurgical failures. They can be either ones where you found root cause, or the unsolved "X-Files" type. Here is one of mine: We made some valve springs from .20" dia. ultra high tensile chrome silicon wire. The springs were coiled, given a 775F stress relief, ends ground, springs shot peened, then hot set.A few of the springs twisted off during the hot set operation, and examination of the fracture revealed an oxidized, pre-existing crack.This crack comprised approximately 20% of the cross section .OK--could be a wire processing defect or perhaps the spring cracked between coiling and stress relief. But what was really interesting was the SEM exam. Emanating from this oxidized, intergranular crack was a band of fatigue propagation approximately 150 microns wide! The fatigue ended right at the final torsional overload fracture.My hypothesis was that the fatigue propagation occurred during grinding of the spring ends from the resonance induced.

 

[SMF1964]

Sounds reasonable.

Weird 2:
A short (4" long) 'safe-end' was installed up in the penthouse during manufacture of a supercritical boiler's final reheat section, connecting stainless steel tubes (in the boiler) to low alloy (1¼Cr-½Mo) tubes (in the penthouse). Thirty years later, a boiler inspecter walking through this region noticed these 50+ safe-ends were bulged significantly. They were all cut out and sent to the lab (me) for evaluation. Microstructure revealed heavy graphitization of the base metal and some micro-fissures associated with the welds' heat affected zones. Chemical analysis showed a 1¼Cr material and, as every metallurgist is taught in metallurgy school: Chromium content in excess of 0.5% prevents graphitization from occurring. This is, of course, assuming that the carbon content is not 1.0%. Apparently, the boiler manufacturer made the safe ends in the shop using 52100 tubes, a bearing steel. A rudimentary jominy test of a strip of this tube material produced an axial crack that ran the length of the piece (1.5" long) until it intersected the stainless steel weld.

A search remains in progress for the welder who was able to weld 52100 without cracking not once, but twice on 50 separate tubes, but we are holding out little hope, since this was 30 years ago...

 

[unclesyd]

I have had welds made/attempted on 52100 wrecked bearings to facilitate removal from shafts or housings. I never pursued the problem that sometimes when the material was welded and it didn't affect the metal in anyway. The only conclusion I could draw was a SWAG that if the bearing had gotten extremely hot, tempered or even partially annealed, during the failure process that the material would be more amendable to being welded.

 

[metengr]

SMF1964;
Got to admit, never heard of any 52100 steel tubes used in a boiler setting. I have seen inadvertent use of carbon steel drain lines in Cr-Mo main steam headers and piping that lasted 30-40 years until failure from creep rupture.

 

[metman]

We had a 4" dia double ended splined shaft that deliverd torque through a CV joint so it was subjected to both torsion and bending stress. The shafts were failing in classical torsional/bending fatigue and would come apart like an exploding cigar and cause severe damage to a $40,000 gearcase. All of the stress calculations were checked and rechecked and it should have been under the endurance limit. Dynamic analysis was also considered.

They tried many types of heat treatment including case hardening, shot peening, polishing of the radii entering spline area, etc.

Finally we set-up a test by cutting a sandstone rock the size of car (this was an underground coal mining machine weighing over 100,000 lbs). A rotating coupling was added to the outboard end of the cutting head to attach electric wires which were wired to strain gages attached to the shaft. It was quite dramatic as the cutting head was lowered onto the rock, this large machine jumped up and down -WOW! It really was impressive.

The readings showed what was expected except for an eye-opener. The shafts were being subjected to reverse torsional loading on start-up. This because MSHA (Mine Safety and Health Act) limit the inrush-current so that right and left hand motors had to be stagger started. There was a LH motor and gearcase and a RH motor and gearcase. The RH motor would start and a time delar relay would then allow the LH motor to start 3 seconds later.

Between the motor and gearcase were several drive shafts, CV joints and a primary gearcase which provided lots of torisonal spring loading and plenty of mass so that when the LH motor started, the RH motor drivetrain was beginning to spring-back and since the two drives were connected at the center at the output, the LH drive would apply a reverse torsional load on the RH shaft. This reverse fatigue loading of course drove the stress above the endurance limit.

We were stuck with shaft size because it transmitted torque through the inside of a sun gear in a planetary and a redesign was impractical for the short term. Some of our customers mining conditions had a lot of rock and shafts were failing after 3 months use. They should have been good for several years.

I recommended for the short term to substitute maraging steel. This at the very least doubled the life but also a very expensive fix but nowhere near the cost of trashing gearcases unexpectedly.

Some years later we were able to redesign the cutterhead to allow larger diameter shafts and use more conventional alloy steel technology.

 

[rorschach]

Not a failure per se, but it almost was:

We had a vertical test well for a downhole tool string where we were doing multiple temperature and pressure cycles. The well was filled with transformer oil as the pressurization medium, but due to a plumbing issue with how the well was built (this was the first use of the test well and the use of transformer oil was a last minute change to procedure and the well was not designed to accomodate that.), transformer oil could not be pumped into the well, so instead downhole control fluid (water glycol based) was pumped in to generate the pressure. Excess pressure was bled from the top of the 60 ft well. The string was made of multiple CRA's (Foroni 918, Incolloy 925, and 13% Cr/420SS) the lowest part of the string was the 420SS bit. As always, this tool string was late and HAD to ship just as soon as it came out of the well. After 24hours worth of thermal and pressure cycling without problems the tool string was pulled only to find that the 420SS part was covered with small incipient pits. Luckily, the pits were so "fresh" or shallow that we were able to polish them out and ship the tool after only a day's worth of elbow grease. What happened was that as the system was temp and pressure cycled, the oil that bled from the top of the well was replaced with water/glycol which being heavier was at the bottom of the well where the 420 SS was. 420SS as you may know, has a propensity to pit in oxygenated water. This water/glycol had not been de-gassed. We ended up with about 55 gallons of water/glycol at the bottom of the well, completely submerging the 13% Cr parts. Lesson was learned, don't use water for the pressurization. The system was quickly re-plumbed to allow transformer oil to be used.