Dimensional Changes During Service

[Carburize]

Comments appreciated on the following:-
I am working on some failures which might be explained by several mechanisms one of which includes small dimensional changes during service.
Failures involve rotating equipment which has been in service for a number of years potentially seeing many millions of cycles of normal loads plus many starts and stops. The materials are quenched and tempered alloy steels.
Is there any information or references to these materials changing shape during their service life due to either redistribution of residual stresses or perhaps transformation of retained austenite?

 

[DAVIDSTECKER]

I'm just thinking out loud here, but rotating equipment means vibration and heat induced stress' can be relieved by mechanical, vibrating, means.
I do not know to what degree, if any the material moves but I would look very closely at that theroy if I was asigned a task like this.
Years ago I was working with forged gears that warped during heat treat even when they were press quenced. We ended up having to take geat care in normalizing and tempering the raw forging in order to get all the stress' out of the material.
Good luck, David

 

[metengr]

Carburize;
We have large induced draft (ID) fan elements on our power boilers that were fabricated by welding quenched and tempered low alloy steel. These fans operate at 900 rpm for many years in various temperature ranges and are subjected to numerous start/stop cycles. The weld detail used in the construction of the fan element was a combination of partial penetration and fillet welds.

A majority (99.9%) of the rotating equipment problems that I have seen and analyzed over the years are typically high cycle fatigue cracks caused by stress risers from fabrication (weld undercut or poor design). I have never seen any fan blade failure caused by dimensional changes from transformation of retained austenite or from excessive residual stresses. Although, local residual stresses from welding could indeed play a significant role in the initiation and propagation of fatigue cracks if a stress riser is present.

We take great care during fabrication of new rotating equipment or for repairs to existing rotating equipment to assure attention to detail like blend grinding weld toes to reduce stress concentration and stress relieving new, welded fan elements to assure increased resistance to fatigue crack initiation in service.

 

[bklauba]

There is no question that residual stresses could have been responsible for the small changes in shape that you see.

How critical is the balance of these components, and was that affected by the dimensional changes? Have the parts been balanced after being put in service?

BK

 

[Carburize]

The assembly tolerances of the two components are +/- 0.0002-inch. The people in the five sided building won't let me say what they are but the closest industrial comparison would be something like a pair of very accurate gears.
The failures occur by fatigue crack initiation and growth but it is not clear what the root cause might be. It is possible that the failures are initiated at the site of small non-metallic particles but I have so far not been able to identify the presence of any specific inclusions . The second possibility is a change in load condition due to changes in relative position or relative size. Instrumentation attached to the parts do not show any sign of relative position change so I was exploring the shape change theory.

 

[metrat]

There may be few things worth considering:

1. The transformation of retained austenite will cause small dimensinal changes, more significant in higher carbon alloys. Most application where this is a real concern are heat treated for maximum stability, using some combination of sub zero processin ( -80 to -120 F very common) and the have the final tempering temperature as high as practile with respect to the required hardness.

2. Changes in residual stress may also have an impact. If the final processes happen to be grinding processes, a significatn amount of stress can be created, sometime this is an advantage, sometimes not. Compressive stress assist in prolonging fatigue life. Tensile stress is generally considered undesirable.

3. If the final processes include grinding, it is also possible to damage the mateiral by grinding "burns", (i.e. overheating the part by aggressive grind processes.) These burns can be inter-related to varying stress patterns also.

If there are shafts associated with the fans, then some of these items may apply, as the fan shafts are most likely ground. The fan blaeds themselves may not be ground, but may have the effects of other forming/shaping processes.

Hope this provides some ideas.

 

[Carburize]

metrat - interesting comments - the parts are carburized and ground. A standard part of the post carburizing heat treatment cycle is a sub-zero treatment at -120F. All the parts are checked for grinding damage using the ISO 14104 nital etch procedure.

 

[alexit]

Had some high speed shafts changing runout/size after time in service here. Found to heat treat, triple temper with double cryo, finish grind, repeat triple temper/cryo steps needed as grinding causes small local heat affected zones (which are not concentric as grind takes more off some areas to make thing round/straight.) Also now perform additional testing per batch where finish shaft run in the open at full speed in oven heated to maximum service temperature and reinspected.

 

[lewtam]

Down to a 4 tenths of a thousandth tolerance? Could a possible cause be themal expansion/contraction?

LewTam Inc.
Petrophysicist, Leading Hand, Natural Horseman, Prickle Farmer, Crack Shot, Venerable Yogi.

 

[metman]

We once had a supplier of custom designed CV (Constant Velocity) joints change their heat treating process for inner races to provide dimensional stability after heat treat which obviated grinding of I.D.splindes to final size. This is a very large CV joint. The change in process included a holding period between upper and lower critcal temperatures after Austenitizing and before quenching. The inner races were carburized 8630 steel.

This was an effective solution dimensionally but produced carbide networks in the grain boundaries so that metallurgically the coalesced carbide trails acted as sites for crack initation and paths for crack propagation from Hertzian stresses. Since you have not been able to identify the presence of any specific inclusions, it is probable that you would have detected carbide networks if present. But this brings into question what metallographic practice you are using since even for aircraft-bearing-quality-alloys there are allowable inclusion count and size.

Not only can grinding burn result in cracks but if LSGT (Low Stress Grinding Technique) is not employed, the burn can generate untempered and overtempered Martensite which later can act as crack initiation sites when loaded with Hertzian stresses. Does ISO 14104 nital etch procedure provide for detection of untempered and overtempered Martensite in the case surface?

 

[Carburize]

There are no significant carbides present.
The ISO 14104 test procedure can identify various levels of grinding damage it is the international version of AGMA 2007

 

[Maui]

Was wire EDM used in the fabrication of these parts?

Maui

 

[aviat]

Just a thought.
Have you considered creep. It causes "growth", but I believe that there has to be heat involved.

 

[Carburize]

Maui - no, EDM was not involved.

Aviat - I think the temperatures are too low - less than 150C