Hot Rolled Screws - decarburisation?

[kclim]

Received some failed screws from a turbine bearing the other day - was requested to perform a failure analysis. See:



The failure mechanism is more than likely to be fatigue:



However, there is some contention as to whether the failure was due to some operational fault (perhaps a screw tightening issue), or a materials defect.

I took some cross sections and microstructures - the decarburised layer indicated that the bolt threads were hot rolled.

However, I managed to pick up two indications -

1. Branched cracks at tip of thread:

(note scale bar)

2. What looks like a lap:


My question is are these indications considered normal; or would similar indications have been significant in the failure?

Regards

 

[unclesyd]

My question is are these indications considered normal; or would similar indications have been significant in the failure?

The answer to the first question is NOand the second question is YES. It definitely appear to be classic fatigue failure.
The branching crack appears to come for the bar stock used to manufacture the fastener and enhanced by the heat treatment or the cleaning process. Depending on the location the lap could have come from one of several operations.

I would immediately replace all fasteners of this type, same head markings as the failed ones. I wouldn't bother with testing the remaining fasteners.

Can you comeback with more information as to the size and material and if possible a better picture of the head markings and purchase specifications.

Make sure you document all you cost associated with these failures.

 

[kclim]

Thanks for the response Unclesyd.

In response to your queries, there are a 'U' and 'H' on the bolt head

See:

http://img130.imageshack.us/img130/5927/dscf01038rq.jpg

I couldn't get a standard size/schedule for the cap head screw. Rough dimensions are:

Head diameter 18 mm
Head height 11.5 mm

Threaded diameter 12.5 mm
Non-threaded 9 mm

Don't have purchase specs; apparently these are OEM bolts installed during commissioning of the turbine.

Apart from the two indications highlighted in the previous images, is the microstructure normal? Should you expect cracking in the decarburised layer?

 

[unclesyd]

Not knowing the material and thermal history the microstructure doesn't look too far off. The cracking appears to be in an oxide layer. There does appear to be some decarb, around the branched crack.

Could you comeback with a hardness, taken on the end of sample.
What was the preparation of samples shown of the crack and lap. You should be able to zoom in on the origin of the two failed samples and hopefully identiy the cracks to the failure mode.

What temperature does the fastener see? If over 450°F I would get a quick analysis for Boron.

Both decarburization and carbizuration causes a reduction in fatigue life of fasteners. It can be up to 50%.

 

[redpicker]

What temperature does the fastener see? If over 450°F I would get a quick analysis for Boron.

I'm curious of the significance of the Boron content. Would its presence just suggest that the original tempering temperatures may have been less than 450F (so that service temperatures above this would soften the material) or is there some other effect I am not aware of.

rp

 

[dbooker630]

I agree with unclesyd, we need to see the hardness data. What is the fastener specification? I have seen certain class 10.9 fasteners that are rolled after heat treatment, and boron is often used (ex. 10B30, 50B46) on fasteners that I receive from Europe.

 

[unclesyd]

The tempering temperature would probably be a little higher than the 450°F I mentioned but we started seeing effects on Boron containing fasteners at this temperature and had catastrophic failure of the aforementioned fasteners at 650°F. If you go above the tempering temperature of any in service fastener you are subject to have "stress relaxation" with all it's resultant consequences.
This is one potential mode of failure, based on the location of the failure, seen by kclim in his fasteners.

Like a lot of people we got caught in the Grade 8 and 8.2 trap and the rush to Boron additions of several years ago. During this period we also found that fasteners containing Boron do not like even short exposures to oxidizing atmospheres.

 

[yates]

I am surprised you assume the bolts were hot-rolled because you see a decarburized layer. Hot thread rolling is normally done at temperatures which will not cause decarburization. Furthermore, threads should be rolled after heat treat and sandblasting/machining so there should be no decarburized layer anyway on threads.
I do not have access to your links, my Company computer does not accept images from internet. Was the rupture in the thread or under head ? Some socket head cap screw are not fatigue-rated, and as such, will not have received underhead fillet roll - a sure failure point in fatigue-inducing turbine conditions. In my experience, rolled threads with laps have just the same resistance to fatigue testing than perfect threads.

 

[kclim]

In response to your queries comments:

High temperature catastrophic failure should not be an issue; the operating temperature is 120 C tops.

I performed hardness tests on the cross section - values ranged between 256-285, with an average of 273 HV.

No specification is available. The bolt isn't too critical in the scheme of things (i.e. bolt fracture did not cause catastrophic failure. In fact, they only picked up the failed bolts when they looked at the turbine during a scheduled outage).

Prep for metallography included grind and polish to 3 micron, followed by etching in 2% nital - standard for carbon and low alloy steels.

I think the surface cracking is in the metal, rather than oxide, layer. The outer layer might appear darker due to etchant seeping out from cracks/pores.

Yates: Thanks for your comments.

One thing I have not considered - is it possible that they've used some compound to get the black surface finish seen in the images. And if so, would that result in the surface microstructure we are seeing?

 

[metengr]

kclim

1. Branched cracks at tip of thread:

I have seen this type of defect as well in lower grade fasteners. It seems as through the material fissures at the thread tip as a result of thread rolling. I personally have not seen this type of defect result in failures.

I would agree that the cap screw failures were caused by fatigue crack propagation. If you have no specification requirements related to these cap screws, I would assume you also had no installation torque requirements. Inadequate preload will result in failures under turbine bearing service conditions.

I wasn’t sure based on your information - are these cap screws for the main turbine bearing caps or for some other function related to the bearings? I would recommend you specify ASTM A 193 Grade B7 or equivalent for this application. All of our large steam turbine components have specific fastener grade and torque requirements for this very reason.

 

[yates]

Black finish is no doubt black oxide finish. Acid baths used should not introduce any embrittlement, particularly at strength levels corresponding to 275 Hv.