Well casing corrosion caused by ESP

[Roberto Salvi]

Good afternoon,

This is a question regarding Water Wells but the matter (casing corrosion caused by production Electrical Submersible Pump) should also be of interest in the oil Wells.
We decided to ask such a questions to Oil Well engineers after a useless search in the Web of this topic in the water Well sector.
All the discussions and articles we found are dealing with corrosion and damages suffered by the ESPs in Wells.

We are facing a fast corrosion problem (2-5 years) in high yield Water Wells.
The casing corrosion is clearly caused by the production ESP (400 HP); video camera shows always corroded casing in correspondence of pump installation depth, nothing above or below this depth;it is something that, surprisingly, seems to be a very uncommon case.

We have found in the web only one article correlating this type of oil Well casing corrosion with the stray current caused by a partially grounded ESP and one more paper regarding increasing casing temperature caused by eddy currents induced by the magnetic field of the pump motor; we believe there could be at least three more reasons for this phenomenon: 1) galvanic corrosion between the SS pump and J-55 Casing (if in contact each other), 2) increased flow velocity will increase corrosion rate in pump motor-casing annulus and 3) perhaps pump vibrations (if in contact with casing).

Main data:
Pump SS-304, OD: 12 ½”, 400 HP
Pump depth: 990 ft
Production rate: 800 gpm
Casing 16”, API J-55, 75 lb/ft

We have a detailed set of information (3 pages) but before to send it we would like to know if anyone of you gentlemen have already experienced such a case and if you are interested in this topic.

Many thanks in advance for any contribution and suggestions (field case, if any, causes and remedial works) on this topic.

Roberto Salvi
Technical Consultant
Hajjan Drilling Co.
Dammam Saudi Arabia

 

[hacksaw]

Is this for a 3-phase 4-wire motor? You should be able to assess and ground fault currents.

What are your salinity levels?

 

[Roberto Salvi]

Thanks for the response.
Yes, it is a 3-phase 4-wire motor, 4,160 V, 60 Hz, grounded by means of 10 ft steel pole (rod) driven into the ground at about 10 ft distance from the Well head.
Water salinity: 1,300 TDS, no H2S, Ph: 6.8

 

[sraesttam]

Robert, SPE paper #16924 indicates that ESP pumps tend to have high vibration compared to surface centrifugal pumping due to the equipment design. Are you pumping any type of corrosion inhibitor? If so, it is possible that the pump moves around down hole during operation, especially during starts and stops, and scrapes off this corrosion inhibitor. Does the ESP have some type of sprayed metal coating? It could also be possible that dissimilar metals are causing galvanic corrosion. I'm sure there are many other possibilities as well.

 

[Roberto Salvi]

Thank you Sraesttam,

We do not have any inhibitor in the Well, only water.
Galvanic corrosion is most probably the corrosion process which occurs between the SS pump top flange and the CS casing (annulus 19 mm, 3/4"); at 300 m (984 ft) depth and without centralizers it is most probable that the flange and the casing are in contact with each other.

Nevertheless something else must happen down there to have such a fast corrosion rate; pump vibrations may facilitate and accelerate the corrosion on conditions that the pump assembly is in contact with the casing and I am quite sure they touch each other.
The eddy currents caused within the casing by the motor and cable magnetic fields may increase the casing temperature, thus accelerating the corrosion rate; and what about the external corrosion acting on the casing ?
There is no CP system; is the motor acting as a catalysator for ground stray currents ?

Sorry for the many questions.

 

[georgeverghese]

Maybe you've got high localised erosion corrosion close to the pump suction due to high localised fluid velocity near the casing wall? Can you devise something that will smoothen out / reduce the fluid entrance velocity close to the pump suction?

 

[Roberto Salvi]

Thanks George,
we know exactly the depth of the pump in relation to the position of the corroded points: these are in correspondence of the pump top flange and bottom of the motor (connection with the cooling exchanger); in one Well the corrosion has also taken place in correspondence of the 2nd flange (20 ft) above the pump.
The flanges, which most probabbly are in contact with the casing, could definetly cause a galvanic corrosion;
pump vibrations may also concur to the abrasion/corrosion at the casing/flange contact.
The motor should not touch the casing: geometrically it is a few millimeters away.
No corrosion appears opposite to the intake (entrance).
But again, why the corrosion is so fast?

 

[hacksaw]

The case ground should be earthed as you've stated, at issue is neutral grounding.

If the neutral is grounded at the MCC, you might be dealing with ground potentials between the neutral and the case ground in the field.

 

[georgeverghese]

If the pump or its motor is in contact with the casing, then agree galvanic corrosion would be possible. Erosion corrosion due to high localised fluid velocity may occue a short distance upstream of the pump suction as the fluid flow profile orients itself towards the suction.
High galvanic corrosion rates may be due to the saline, high temp environment here ?

 

[Roberto Salvi]

Sorry fo rhe delay in answering

Hacksaw: you look very familiar with electrical stuff but I am not; the thing seems to be quite close to the concept of stray currents caused by a faulty pump phase: could you please explain your idea in an accessible way ?
Thanks

Georgevergese:
High localized fluid velocity: the flow direction is upwards when you look up from the motor exchanger (cooling device), which is the lowest component of the pump assembly; high velocity is therefore only foreseen below the strainer ( pump intake ); above the pump (upstream)the fluid does not move that much.
High salinity water: TDS = 1,300 ppm
Water temperature: 30 deg. Centigrade
As mentioned above to Sraesttam the eddy currents could sensibly increase the casing temperature.
Thanks

 

[hacksaw]

Roberto,

So you have a 3-phase 4-wire motor. The fourth wire is your electrical common or neutral. You motor connections will be either Y (uses the neutral) or Delta (floating)

Some systems use a grounded neutral, grounded at the motor controls and ultimately to your substation ground, you can also have a floating neutral or a resistively grounded neutral.

One thing to check is the voltage from the neutral wire to the motor case (grounded locally). That will tell you how much differential between the substation ground and the local ground.

You also need to do a grounding resistence test to findout the effectiveness of the local ground.

Megger test of the motor winding to case ground (disconnected power wiring) will tell you if you have current paths to motor case.

 

[Roberto Salvi]

Thank you Hacksaw, I have understood the concept.
Now I have to convince the pumps owner (our client, a government organization) to test all the pumps they have in the Wells (about 60 Nos) to see how many of them show a ground potential sufficient to cause stray currents on the Well casing. Also I think it would be interesting to know why 42 pumps have been retrieved and then reinstalled in only 1.5 years; perhaps some or many of them had grounding problems.

Anyway, I thank the guys who answered to my insert, it was very helpful; as soon as we get some results of the investigations we asked for I will inform you; but I do not know how long it will take.

 

[hacksaw]

Roberto,

Good luck, hopefully a simple fix solves the problem. Ground potentials are always an issue in the field.

AC potential in brackish water releases, O2, H2, and HCl on both electrodes...and the corrosion can be severe.

 

[sraesttam]

Roberto, I calculated an average fluid velocity of 4.5 FPS in the annulus based on the data provided. This is likely adequate to cool the equipment (>1 FPS) and will not likely cause erosion (<7 FPS for fluids with abrasives). Any additional heat in your casing caused by induction would likely be negligible and transferred to the fluid.

I am not an electrician, but it sounds like you and hacksaw are on the right track.

I understand you have replaced 42 pumps in 1.5 years. Are you saying that you have 60 wells giving you a failure rate of 0.5 failures/well/year? I agree - that seems very high for pumping water in that big of casing from less than 1000'. What percentage of your pump failures are caused by corrosion? The pump failure mechanism and casing corrosion could be related.

 

[Roberto Salvi]

[a]Hacksaw: good to know it; on top of that any stray current entering the casing close to the surface (cathodic point)would corrode the casing at its exit point (anode); with no CP system in place, which should protect the external side of the casing, I could imagine that the current exit point is located somewhere at pump depth and this could explain the fast corrosion rate; but the question remains: why at that depth ? Is the motor, with its magnetic field, responsible for causing, at its depth, the outflow of the stray current ?
As you rightly suggest O2, H2 and HCl could develop at both electrodes, the casing being one of them: what are the probabilities that the casing “electrode” is located 1000 ft below surface, just in front of the motor ?

Sraesttam: yes, you are right with flow velocity calculation but unfortunately, as mention above, the pump is not centralized and probably eccentric to an extent that flow lines are “restricted” on one side and “enlarged” on the other side; it is therefore difficult to quantify the flow velocity.
The large majority of the retrieved pumps, to my knowledge, had electrical problems, some motors were also burnt out and the rest (I estimate about 10%) had sand problems.
This does not mean that casing corrosion has happened only in the Wells where sand pumping has occurred; I am quite sure that if we could run a casing integrity log or a camera log at pump depth the results would show a thinning out and/or a perforation of the casing.
However, not always a casing perforation may lead to sand inflow, even in a sandy formation; the formation we are dealing with is made up by alternating clay/claystone (not shale) and loose sand (not sandstone) layers; if the casing perforations are located opposite to clayey strata, sand, if any, might appear very late during the Well production life.

Sorry for the lengthy explanations; I appreciate your help.

 

[hacksaw]

This may help, mention is only made of leakage currents and corrosion, but little detail

http://www.slb.com/~/media/Files/technical_papers/2010/2010_esp_geothermal_applications.pdf

sounds like the pump manufacturer needs to get involved

 

[Roberto Salvi]

Hacksaw, this is exactly what we also tought a couple of months ago; we sent our request of "help" to five (5) well known pump manufacturers, including the one of your link: no one has answered! It seems they do not like to get involved.

 

[hacksaw]

Sounds like they are all too familiar with the problem!

Bottom line is that you have a current path from the pump/motor to the downhole casing.

The fixes are not going to be cheap, in fact possibly prohibitive.

Options:

1. Take one well and instrument it to research what is going on. Not likely you find any business doing that in this day and age, even though it is an industry wide issue according to the electrical journals.

2. Install an isolation transformer. Once done, measure the voltage of the "floating neutral" relative to the casing.

3. Perform routine pre-construction electrical tests on a de-energized motor. Megger, ground resistance of the casing to a proper electrical ground,

4. Just keep replacing the failed systems...you may have to do that even if you identify the cause!

 

[Roberto Salvi]

Hacksaw, many thanks for your conclusive comment.

We are planning to convince our client to:
1. Use plastic centralizers for the pump assembly (contact with casing will be eliminated)
2. Install a CP system for each Well

If the CP system is not accepted we will propose your suggestion:

1. Install an isolation transformer. Once done, measure the voltage of the "floating neutral" relative to the
casing.

2. Perform routine pre-construction electrical tests on a de-energized motor. Megger, ground resistance of the
casing to a proper electrical ground

plus the plastic centralizers as above.

Pump can be replaced (expensive) but Wells like these ones affected by the corrosion are difficult to repair unless you are prepared to accept a reduced yield.