Titanium Bolt - SLC

[ella0509]

Hi all, hope somebody can give some good pointers. We are looking to use a Titanium Gr.5 (Ti6AL4V)fastener bolt within our assembly that will be in continuous load for >20years (subsea app). I can see that this material suffers from SLC by predominantly the H2 content coupled with temperatures <20C and the sustained load. One way of eliminating/radically reducing the embrittlement risk is to minimise the H2 content possibly by Vacuum annealing the material, I've seen text stating 10ppm is achievable. My supplier states that 60ppm is more their usual requirement and anything lower will be difficult and/or very expensive. Further, more research is stating 50ppm as the 'tipping' point.

If I state material supply to the supplier value of 60ppm, and assuming I can't alter the application service temperature, what else can I do during design analysis to ensure theat SLC does not become a mode of failure - without embarking on some massive test programs. The obvious thing seems to be to reduce the sustained load or stress intensities within the design, but what I'm struggling to find is to what value I need to reduce these to. Can anyone with greater experience of this material and phenomena shed any light? Ti6AL4V ELI (Gr.23?) has been mentioned in pasing as potentially better alternative for this application - again, can anyone tell me why??

Thoughts??

Thanks all!

 

[CoryPad]

It sounds like you understand the basics well enough. Low hydrogen and low stress are good. The ELI grade provides higher fracture toughness because it reduces oxygen, which is known to affect crack initiation and propagation. Make sure there is no alpha case (oxygen enriched surface) which also leads to cracking. Lastly, beta processing produces an acicular microstructure which has better toughness than the equiaxed microstructure from alpha-beta processing.

 

[ella0509]

Hi Cory, thanks for your reply. Do you have any case examples/past knowledge of what sort of reductions to allowable yield stress, I might apply to aid prevention of the SLC problem?

Lee

 

[CoryPad]

There are so many factors, and a dearth of information on your application, that it would be difficult to provide a number. I recommend testing for your application.

 

[rb1957]

from MMPDS-01 (successor to Mil-Hdbk5) ... both available online

"The material is marginally susceptible to aqueous chloride solution stress corrosion, but is considered to have
good resistance to this reaction compared with other commonly used alloys. Under certain conditions,
titanium, when in contact with cadmium, silver, mercury, or certain of their compounds, may become embrittled.
Refer to MIL-S-5002 and MIL-STD-1568 for restrictions concerning applications with titanium in
contact with these metals or their compounds."

I'd talk to a specialist supplier/metalurgist/university to see if this is a big problem, and how to mitigate it (low working stresses, residual compression surface stresses?, protection??, inspection???

Is Ti the best choice material ? (I'd think so)

Quando Omni Flunkus Moritati

 

[TVP]

You haven't mentioned some important details, namely, will the titanium fasteners be joining titanium alloy members, or some other dissimilar metal? Will this assembled joint be exposed to cathodic protection? There is some good information in the publication "Guidelines for Materials Selection and Corrosion Control for Subsea Oil and Gas Production Equipment", with the 2nd edition published in 2004 by The Engineering Equipment and Materials Users' Association (EEMUA). Things to consider:

Fastener size (larger requires more material with more toughness)
Joint materials (best to isolate Ti if connected to steel, etc.)
Use of cathodic protection (best to limit the voltage to -900 mV Ag/AgCl/seawater)

Grade 23 is usually used in the annealed condition with YS ~ 110 ksi (760 MPa), UTS ~ 120 ksi (825 MPa), and Elongation ~ 10%. Another alloy more recently developed, TIMETAL 5111 (UNS R5111, Ti-5Al-1Sn-1Zr-1V-0.8Mo) has about 3 times the toughness of Grade 23, but with slightly lower strength (YS90ksi/UTS105ksi/El10%). Rolling the threads and underhead fillet introduces compressive residual stresses which are beneficial for delaying crack initiation. The rounded root of the MJ or UNJ thread is better than standard M or UN thread.

 

[ella0509]

Hi guys, thanks for the responses - I'll pick up on some of the references you've dropped in.

The fastener prototype consists of a single STUB ACME threaded bolt running into a corresponding nut. All metallic components are currently of the same material, and the intention is to keep this approach when moving towards Titanium usage. There are no CP requirements. Its about 1" in diameter.

 

[TVP]

With the trapezoidal-type thread, that will obviously have to be cut, but you may want to investigate rolling the surface after cutting. Ecoroll makes tools for this. Check out Application Example 506 on page 82 of their catalog:

http://www.ecoroll.de/en/service/downloads.html

first link [6.7MB]

 

[ella0509]

So, assuming I do as much as I can up front to reduce this problem: i.e.:
[ul]
[li]Select a Ti6AL4V ELI (Gr.23) material.[/li]
[li]Forge and then finish machined to the required stud geometry (with generous radii in thread roots and under head).[/li]
[li]Vacuum anneal / beta process it to have Hydrogen <60ppm and improve toughness.[/li]
[li]Thread surface rolling where possible.[/li]
[/ul]

However, at this stage how do I initially size the bolt? I know its service load, I know the material yield strength, I know the application FOS required etc, so this would give me the required size in 'normal' conditions. However, the failure caused by this mechanism can begin to happen at values less than this presented yield strength, so the stud would effectively be undersized.

Do I need to start looking at Fracture Mechanics and stress intensities - the assumption is that the component will contain a crack of a certain shape and size (i.e. the size of the limit that we can detect through NDT etc). This may need to go out to specialists if that is the case as I doubt we have the experience in house to do the testing / analysis required.
FEA will give me stress concentrations, but these do not necessarily correlate with stress intensity figures I believe.

We're still in early development and therefore at this early component sizing stage is there not some simple 'rule of thumb' or yield stress reduction factor than can be comfortably applied to give confidence that the size is valid before committing to endless hours of testing etc. Basically if this SLC effect causes the studs to be bigger and not then cost effective we would look at other materials - I'd hate to commit too much resource and compress the schedule for it to be 'wasted'.