Odd corrosion pattern [at least to me]

[hailtopitt]

Hope someone can help brainstorm an idea with this one.

Running a 20" carbon steel pipe about 600 ft in length which ends up dumping into a separator at the end of the line. Gas is, on a dry basis:
80-85% H2 and CH4
10% N2
<2% O2
3-4% C6H6 and derivatives
3-4% H2S and HCN
(Percentages don't add up perfect but that's the general range)

Gas is also saturated with water due to the gas compressors. The separator in line is meant to drop the water out (so I guess you'd call this a wet gas line). The line is 30 psig and 180 Deg F.

The problem we've seen is that the piping has recently begun corroding at a very high rate, and only at the top of the pipe. The wet area experiences little thinning and the water system itself also sees nothing.

Almost nothing has changed, the only notable thing being the removal of excess NH3 from the system. The water in the separator used to run a pH around 7.5 because of the extra NH3 but since then, it has dropped to about 5.5-6. I could see this as a problem, except the wet area isn't what's seeing the corrosion.

Is there an anomaly where the condensed water (ie) droplets on the top and side) could be even more corrosive than the bulk water? I feel like I'm grasping at straws there. Any bright ideas?

 

[Latexman]

Can you speak to the flow rate and regime of the gas and liquid phases in the 20" pipe?

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[hailtopitt]

Gas flow is about 25-28 MMSCFD. Liquid Flow is about 1500 gpm.

The gas flow is certainly turbulent and I would think the water is as well. We don't have thermo data for where bulk water sits in the pipe.

 

[sshep]

When a drop of water forms, it is a magnet for any acids in the vapor phase. The first point of condensation is the worst for corrosion. The bulk liquid is largely out of contact, and has more dilution capability. If the bulk pH is 5.6, then localized pH seems well able to corrode carbon steel.

Good luck,
Sshep

 

[SJones]

What morphology is the corrosion taking - pitting or thinning?

Steve Jones
Corrosion Management Consultant

http://www.linkedin.com/pub/8/83b/b04

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

 

[Latexman]

When NH3 was in the gas, there was a pH check and balance (acid/base) on the local condensation areas.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[hailtopitt]

Thanks for the help everyone!

My first instinct was that because of the added surface area and much smaller volume of liquid, the film / droplets could have a much higher rate of absorption, but I also wasn't sure if the water would be saturated after running through the gas compressor. I couldn't find an accurate way to calculate the rate of gas absorption into liquid as a function of T, P, and concentration gradient, etc. Anyone have an article about this in their back pocket? Figuring out this, I'd be able to get a relatively good idea what the local pH of the droplets would be.

Given the constituents in the gas, what would the mechanism for the corrosion be? I've seen a lot of research saying HS- and CN- can directly attack the iron, but also can form a hexacyanide complex with iron.

SJones - It's pitting on the inside and making pinholes on the outside.

Latexman - I was going to add NaOH to the water prior to being re-exposed to the gas. Assuming the acid-base reaction happened upon absorption into the water, I'm not sure this would work, as the bulk water would increase in pH but there wouldn't really be any way for it to affect the condensates that form.

 

[SJones]

A rough view of the velocities suggests stratified flow which is ideal for top of line corrosion. Ideally, with that amount of H2S, a protective sulphide film would form, but this would be prone to break down with the cyanide content. The oxygen is puzzling as that will generate elemental sulphur from the H2S which would act as a nice pitting agent. The condensing water will have an initial low pH that will rise as it becomes saturated with corrosion product. It may be that the removal of the ammonia has tipped the balance in favour of the corrosion product film becoming marginally protective and prone to localised breakdown.

Steve Jones
Corrosion Management Consultant

http://www.linkedin.com/pub/8/83b/b04

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

 

[chem1357]

I think the problem is removal of excess nh3

 

[hailtopitt]

I think the problem is removal of excess nh3

That's not really an option...

A rough view of the velocities suggests stratified flow which is ideal for top of line corrosion. Ideally, with that amount of H2S, a protective sulphide film would form, but this would be prone to break down with the cyanide content. The oxygen is puzzling as that will generate elemental sulphur from the H2S which would act as a nice pitting agent. The condensing water will have an initial low pH that will rise as it becomes saturated with corrosion product. It may be that the removal of the ammonia has tipped the balance in favour of the corrosion product film becoming marginally protective and prone to localised breakdown.

Typically the gas portion of the piping is coated with a layer of black sticky substance. This seems to be partially organic and partially salt-based. in water, some of it will dissolve, and solvation increases at higher pH. At low pH, it breaks apart quickly in water and settles to the bottom within a few minutes. In benzene, the sample dissolved readily (meaning it's certainly organic too). Don't know if this gives any additional insight.

 

[BenThayer]

erosion-corrosion? velocity in the "gas" portion could be stripping off the passivation layer and then it keeps the corrosion active.

also, it sounds like you are having wetting/drying of the "gas" portion. never a good thing. especially if you are getting localized concentration areas of the H2S in the water as it evaporates/drys.