cracking mechanism idenstification

[Chumpes]

Dear Experts
Our client found a leaking valve in a hydroprocessing unit. Valve body material is forged stainless steel 321. Service in normal operation is hydrocarbon containing sulfur with Pop = 160bara and Top = 380°C. No water is anticipated during normal operation and also not anticipated during shutdowns. Other items in the circuit (other valves, piping, ...) were characterized (Dye Penetrant Test) and no other indication was found. The valve was manufactured by a valve manufacturer with good reputation and experience. Fabrication certificate indicates the cracks were absent after fabrication. Manufacturer had never experienced such a failure.

The cracks are being characterized, but the conclusion on the actual damage mechanism is not evident.

Looking at the crack mophology, do you have any idea on the possible origin of such transgranular cracking ?
Does it look like fatigue ? Or does it looks more like Cl-SCC or another SCC mechanism ?
Is corrosion-fatigue possible in hot sulfiding environment ?

Thanks in advance for your advise
Regards

 

Looks like transgranular SCC, with very little branching. Not good.
I suspect there was indeed moisture present. Shutdown is the most vulnerable time for equipment, so I would investigate the particular circumstances around your normal shutdown practices. The steel is not necessarily in a compromised condition (heat treatment wise), but do not rule it out.


"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

[cap1a79]

Check if the corrosion product contains chloride by metallography. If yes, then do not limit the source of chloride only in the water. Although this is my personnel opinion, the source of chloride could be "organic chloride" which used to be attached to hydrocarbon but broken by high temperature and become free chloride, inducing clscc. (Similar case has been reported in NACE.)
Check organic chloride for the relevant streams especially for makeup hydrogen which is connected with overhead of separator(downstream of aircooler) whether there was chloride carry over and boot section of the separator aa well.

Lee SiHyoung,
WorleyParsons Oman Engineering,

 

[Dhurjati Sen]

Dear Chumpes,

First things first. Carry out chemical analysis (not PMI) of the valve body. Take the sample from the unaffected portion. Your answers might lie in the result.

Regards.

DHURJATI SEN

https://www.nace.org/people/dhurjatisen

 

[mrfailure]

Looks like an environmentally assisted cracking mechanism. Besides SCC, you may want to review the feasibility of polythionic acid stress corrosion cracking. I'm not saying it is more likely, just that it should be looked at to see if it fits operating conditions.

As a note: API RP 571 is a very valuable reference for identifying petrochemical failure mechanisms.

 

[cap1a79]

I believe PASCC can occur on "sensitized SS" as a result of operation at above 427 C while the temperature mentioned is 380 C. Anyway, it would be helpful to check operation history whether it has operated in the range of sensitization.

Lee SiHyoung,
WorleyParsons Oman Engineering,

 

[MagBen]

Have you observed surface anomalies, e.g. pitting/crevice corrosion, rough surface etc.? lack of branch and dominated transgranular cracking lead me think this may not be Cl-SCC. sensitization seems unlikely because the grain boundaries are not affected, also alloy 321 is Ti-stabilized to protect grain boundaries, and operation temp is not high enough.

There might exist fine cracks before serve. Dye Penetrant is not always a reliable way to detect small cracks.

Note the grain size is pretty big, some are ASTM 0. Is material heat treated properly? Any gamma phase that can acts as a corrosion site? anyway do not rule out the possible material issue.

 

[btrueblood]

Any chance of low melting point metals being contained in the fluid stream?

 

[Chumpes]

dear all
analyses of the opened crack show some indications of striations probably produced by fatigue.
EDS and SEM analyses into the craks showed thin black deposit composed process fluid without corrosion products and NO CHLORIDES.
@MAGBEN : no other damage (pitting, ...) was found on the adjacent surface. you are right, some cracks present before service in the bulk material (not opened to process fluid or atm) were detected but are attributed to stress relaxation during cutting.
thank you all for your comments and suggestions

 

What is shown in the micrograph is almost certainly not a fatigue crack. There may be another portion of the crack where fatigue cracking was dominant.

"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

[Chumpes]

@ironic metallurgist : picture does not look like a fatigue crack to you because of there is too few branching ?

 

Because fatigue crack branching is extremely rare. More typically you will find multiple fatigue cracks with slightly displaced origins that merge before growing very far. This gives rise to 'ratchet marks' visible to the unaided eye on the fracture surface near the origin.
Were there more cracks than are shown in the picture? With SCC you often find multiple cracks.

"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

[Chumpes]

Sorry, I only posted high magnification view of the main cracks

Please see the following picture with a broader vieaw of the cracks.

It pretty looks like Cl-SCC for me... branched transgranular cracking...

Is there any way / characterization / ... that could be implemented in order to validate or invalidate root cause og cracking by Cl-SCC ?

 

[EdStainless]

You need to figure out if the origin was the process side or the exterior.
One issue with any of the EAC mechanisms is that the crack morphology is related to the stress level and rate of crack growth.
While you high mag shot show little branching and likely not CL driven that does not mean that Cl didn't play a role at all.
I am very suspicious about the grain size, though the micro looks good (lots of small TiC scattered about).
These cracks could easily have 2 or 3 different drivers at different locations.
I would look into PA also.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[mrfailure]

Penetrant may have diluted any chlorides that might have originally been present. If you have a section of visible cracking that was not PTed, then you can look there for presence of chlorides.

 

[blacksmith37]

Chloride SCC , whether or not you find chlorides now. Not polythionic because the cracks are not exclusively inter-granular and 321 is resistant to some degree.

 

[Chumpes]

Thanks for your advise, also on possible lack of chloride detection after Dye Penetrant Test.

@Ed : The origin is confirmed on process side. I'm a bit tired... what is PA ?

Regards

 

[mrfailure]

I think PA is polythionic acid, mentioned earlier in the thread.

 

Good luck finding chloride ions; they are extremely elusive after the fact.

"Everyone is entitled to their own opinions, but they are not entitled to their own facts."