Pipeline damage/failure root cause 1

[Mahmoud Khalaf]

Hello everyone,

I work in an upstream oil and gas site in Egypt.
Recently, 18 months ago, we introduced into service a new 2 phase crude oil 12" subsea carbon steel pipeline run from a gathering platform of nearly 40 meter water depth to the onshore oil terminal process.
The pipeline designed to the DNV-F101-OS offshore standard and the subsea installation carried out through the S lay barge.
Unfortunately, Three weeks ago, we have an oil spell over the line, we got forced with unplanned shutdown and switch over the production stream to another 3phase 20" pipeline.

The divers did finally catch four leakage points / ruptures observed sequentially after repetitive hydrotests.
The ruptures located at different distances apart from each others, the initial investigation showed nearly a same configuration for the rupture defects @ weld HAZ area.
The welding process at the barge got strictly controlled through three fully supervised welding inspectors that followed the DNV code instructions, hence we fairly exclude a sharp welding defect that emerges repeatidely nearly at the same locations relative to the girth welds, both axially (@the HAZ) and circumferentially (nearly the same O'clock side)

By the way, after completing the successful hydrotest and geometry pig, the pipeline got preserved about 6 months before introducing into operation with a doubt in the efficiency and control of the presevation process.

On the other hand, few months ago, the onshore section connected to this subsea pipeline has suffered a lateral movement of about 1.5 meter over the resting concrete supports that caused a damage observed for the pig trap supports.
We,here, guess that operation surges might cause the damage shown at the onshore section and hidden at the offshore subsea one. I guess that the heavy concrete coating the subsea section preclude its lateral movement, and initiate a huge fatigue stresses that released in a rupture form.

At the same time, the line convey two phase crude of %70 water cut with a relatively low flow speed, and the corrosion of the internal surface might accelerate heavily but how we could believe that corrosion leads quickly to this catastrophic failure.

We are planning to call an experienced third party to present an engineering failure root cause analysis to stand on the most likely failure reason and help support Safe future operation.

Also, we have to garantee all the following conditions together:
- Assure the pipeline integrity via confirming no other severe cracks left in the line after doing pigging using the compo tethered solution including both crack detection (TOFD technique) and corrosion detection technique.
- Confirm the failure root cause analysis to assure future Safe and reliable operation without the repeat of this catastrophic failure.

Finally, I would be grateful if someone advise the potential failure root causes that might be relevant to this kind of catastrophic failure.

 

[LittleInch]

Corrosion at the bottom of the pipe is quite telling, especially with your statement "the pipeline got preserved about 6 months before introducing into operation with a doubt in the efficiency and control of the preservation process."

Getting bugs in the system can just eat your pipe and with an internal weld bead you may have just developed a little pool of microbes in this area and had them leach acid onto your HAZ.

Once you get a corrosion hole going it doesn't take much of your slow moving oil in water mix to get accelerated corrosion if you were down to 2-3mm by that time. You don't say what, if any, corrosion inhibitor you're using? Or maybe I missed it?

Process plants have had to be re-piped in the past due to hydrotest water sitting in them for 6 months before operation.

Your other pipelines may have started off transporting much more oil than water and hence tend to get protected a bit by the oil. once you go to 70% W/C you're in a different land.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Mahmoud Khalaf]

@Steve
The TWI study target the normalized residual stress readings (residual/yield - Y axis) versus a specific location relative to the sample thickness (normalized thickness-X axis).

This is different issue that I would present.
In metallurgy science,the below is a rule:
Increasing the carbon content (%carbon, %Mn, Cr, etc..) will increase the yield and tensile strength values and decrease ductility.
This means also harder metal forming process (weldability in our case)and increase the tendency to generate more residual stresses regardless its type (hoop or axial, tensile or compressive, etc..).

Therefore the majority of the high strength alloy steel families mandate PWHT, regardless its thickness, as it has relatively higher carbon content induced by the added alloying elements (Cr, V, Mo,..) and when welded generate higher residual stress that mandates PWHT to soften the harder produced microstructures,

 

[SJones]

@MK
You may want to talk to a welding engineer about some of those statements. Irrespective of composition, the simple fact of having a liquid metal pool freeze between two very large members is going to generate some residual stress somewhere. Whether the stress was sufficient to assist mechanical cracking, or stress corrosion cracking, only your failure analysis will tell. The “repetitive hydro testing” may also have served to propagate defects to failure point.

Steve Jones
Corrosion Management Consultant

http://www.linkedin.com/in/drstevejones

All answers are personal opinions only and are in no way connected with any employer.

 

[Mahmoud Khalaf]

In the above main thread, I did mention the below status:

On the other hand, few months ago, the onshore section connected to this subsea pipeline has suffered a lateral movement of about 1.5 meter over the resting concrete supports that caused damage observed for the old pig trap supports.
We,here, guess that operation surges might also cause the damage observed at the offshore subsea section. I guess that the heavy concrete coating the subsea section preclude its lateral movement, and initiate a huge fatigue stresses that released in a rupture form.

The question is; to What certainly level, we could suppose that damage relevant to fatigue cracks iniated due lateral loads induced by operation surges could be a root cause? I know that it needs deep investigation but might someone has a relevant experience do guess?

Does the fatigue failure occur in many locations distributed along the pipeline and probably in relatively same time or just one fatigue rupture could cause full relief to the operation surges induced stresses? If it is nearly fatigue, why did it repeat at many locations nearly distributed within 500 meters along the pipeline?

 

[GD2]

Mahmoud,
It's hard to believe that the pipeline will have a fatigue failure only after a run of 18 months and that too at a pressure of 6/7 bar over a design pressure of 40 bar at 40 degree C.
You said that you had some operating surges. What exactly is that? Or operating upsets?
Do you have data on cyclic Loading time histogram on pressure and temperature cycles? 15 -20% on pressure variations of the design pressure shouldn't be a contributor to fatigue failure. I wouldn't also think that you had a lot of temperature variations.
I still suspect that your 6 months preservation played a big role in initiating corrosion. All failures in the HAZ indicates that Non-PWHT of the welds have contributed to the stress level.
Did you find the MTR for the pipes? PSL2 pipes are primarily used for sour service. It requires controlled welding.
Last but not the least, did you check if the onshore plant saw frequent process loading variations. If there was slugs coming in the flow, with no slug catcher, they should see it.

You also said that there was a crack at 4 O’lock position. What is this location’s elevation compared to other locations? Is this the lowest point in the pipeline profile?

GDD
Canada

 

[Mahmoud Khalaf]

The 4 O'Clock rupture is the nearest one to shore line so nearly it has the highest elevation (-11 m)relative to the upstream remaining three ruptures as the line runs from -40 @ riser seabed-touchpoint to the downstream onshore terminal of a relatively sea level elevation.

Another issue that lets me doubt about fatigue is the rupture morghology which looks like fish eye, the most defect morghology relates to fatigue rupture. Rupture image is attached.

Again if the doubtful presevation procedure had greatly been a powerful factor in the failure initiation process, so, in my opinion, it should be logic that the failure morghology appears like Corrosion of uneven shape dimensions rather than these observed sharp edge appearing rupture located everytime at HAZ?

 

[Mahmoud Khalaf]

Conversely, another important information that disrupt my guess in fatigue failure is that the diver points in a diving vedio to a confirmed hole beside one of the ruptures with dimensions of 2.5 cm length in 1 cm width!

Video is attached; the hole is clearly seen at the second 71 (01 minute:11 second)

 

[TugboatEng]

In galvanic corrosion of carbon steel we often see the weld acting as the anode and will have substantially higher corrosion rates than the surrounding metal. If galvanic corrosion is occuring internally, the entire weld bead may be wasted away but the cross-section is thinnest at the toe which is why the corrosion appears localized in that area. Given more time, the cap would corrode away as well.

Here is an example of a weld bead hanging from a pipe due to galvanic corrosion. It corroded through at the toes on each side of the weld and the crown fell off.

 

[GD2]

Mahmoud,
Can you send us the process stream condition when it enters the pipeline after the offshore separation plant with the H2S partial pressure?

I will rule out fatigue failure. Reason: Even if there is one process upset cycle a day, 540(30/month x18 months) cycles is not enough for a fatigue rupture given that it is operating only at 18% of the design pressure.


GDD
Canada

 

[Mahmoud Khalaf]

Come back,
We have informed recently via verbal confirmations, that many times when instrumentation guys do their routine PM at the production platform, they reported verbally a non conformance events, they found the level control loop of the production separator upstream the pipeline routinely switched override!, and even the automatic shutdown valve and pressure switch high (PSH) connected to the pipeline found overrided too!!

When they were asked, the process guys replied that due to the upstream process disturbance(wellhead flow regime) we were forced to do so, otherwise a complete shutdown would have been initiated so much,even daily!.

Regardless this attitude and its relevant refused justification, we expect that due to this reported miss level control, considerable amounts of gas surges would have been introduced into the liquid phase inside the pipeline and might initiate a slug flow regime. when we do a relationship between this event and the confirmed pipeline snaky movement seen for about three hours and reported four months ago before the accident, I would guess the displacement overstress ( huge bending stresses) that might contribute to this disastrous failure.

I do, here, differentiate between irrelevant fatigue stress and the displacement overstress we talk about.

I would inquire, does this guess seem true ?

 

[GD2]

Mahmoud,
It appears that slugs are flowing from the wells making the separator units unstable. This is one of the major challenges in operating offshore facilities. One of the things Operations have done is override the separators controls. If you ask around, you might also find load and vibration issues with the gas compressors that flares the gas.

Nevertheless, the root cause of these problems - the slug flow - must be smoothened and controlled. Operation is right, if they didn't override, it should been more trips and shutdowns.

Several approaches are adopted to get way with this this kind of problem. Onshore slug catcher is one which you don't currently have. Sometimes, they make the first stage separator bigger or by adapting the separators to receive the slugs.

With regard to the subsea pipeline, it must have seen intermittent liquid plugs and even stalled flow. This on-off flow patter and the energy the liquid plugs dissipated must be the cause for the pipeline rupture and the onshore damage/displacements at the pig receivers.

In my perspective, analyzing for fatigue failure is not going to remove the root cause of the failure. It will come back again. The issue must be looked from an organization point of view with operations, process, inspection, engineering groups and a corrective action o smoothen the slugs must be planned.

Slug flows are common, it's a challenge but there are also means and ways to control it. Overriding controls and not taming it, is not a solution.

GDD
Canada

 

[Mahmoud Khalaf]

GDD

Thanks alot for your long breath discussion and valuable comments that i would highly appreciate.
I would ask about your advice regarding such below issues.

Till we, based on an organisational scale investigation, catch the most likely failure root cause and although this is a hard mission due to socio-cultural issues, we should deal with the disturbed operation as new emerged challenge.
The wellhead flow regime might get new harsh flow characteristics that should be modelled in a different profile that differs from the old original one.

This new flow modelling has to be designed through special engineering organization and should conservatively suppose the relevant flow parameters in its extreme figures/values.
Concequently this wellhead flow model might force substational modification in the downstream precess onboard the production platform like upgrading the heritage pneumatic control system, resizing the separators, equip new system, etc..

But as we know this is a hard challenge and might require much funds that the concerned shareholders are not willing to pay, in such brown field, so the other alternative that could be studied, specially if it gets incidentally forced to switch the production stream back to 12" main oil line and after considering the economic impact of accompanied potential deffered production, is to keep gas lift operation either completely or partially outage and maximize the production dependency on ESP solution only, even with its relevant operation upset which might be less disturbant and more safe when compared with the exising slug flow regime that disturbed by uncontrolled gas surges mainly feeded by gas lift operation.

Mahmoud Khalaf
Asset integrity.

 

[GD2]

Mahmoud,
I am not onto Downhole engineering.
What I see from your perspective is that management wants to fix the broken line ASAP and start production. You can't do much in doing a good engineering job right from downhole to re-engineering of the whole system. This kind of project will atleast take upto 3 years from approval to construction. You guys being in a tight spot and in operation, management is tracking how soon the line will be repaired and up and running again.
So my advise it, within your limited scope and responsibility, fix the line and bring it to the shape back. Let your engineering team work on the root cause issues, which I am sure will continue through tens of discussions and deliberations until a decision for corrective action plan is taken.

Not much help from engineering stand point but this is exactly what happens in an operating plant environment once the project is handed over to the operation and maintenance. And you are right- socio-cultural issues more than what needs to be planned and done.

GDD
Canada