erosion of stainless steel piping with steam and oxygen 2

[frostrobn]

There is a process in my facility that uses live steam (160 psig saturated) mixed with oxygen which is piped to the process in 316L stainless steel piping. The piping just after the point of oxygen injection in the steam line is eroding rapidly. The line has to be patched every 3 months or less. Does anyone have any experience with this problem? The piping upstream of the oxygen injection point is fine, it's just the pipe after the introduction of oxygen that is troublesome. Other information: The line is under high frequency vibration caused by injecting live steam into the process a few feet after the oxygen injection point. I have pics if anyone is interested. Question: Would 2205 (duplex) stainless steel be a better metallurgy in this situation? Any response would be greatly appreciated. One response I've received pointed at Moly as an alloy in 316L as the culprit due to rapid oxidation due to the oxygen and high temps. Any ideas?

 

[BronYrAur]

The metallurgy discussion is over my head, but I'm thinking that the O2 is condensing steam at the injection point and this is causing the problem. Why are you desuperheating the steam? It would seem to me that keeping it superheated will allow for cooler O2 to enter without condensing any (as much) steam. By the way, what is the temperature of the O2?

 

[frostrobn]

We are now doing a trial with no desuperheating. This is for two reasons.

The first is to eliminate the addition of chlorides since the desuperheating fluid is simply treated water, not boiler condensate. By "treated" I mean treated with sodium hypo to kill the bugs. If the process requires desuperheating in the future we will probably route boiler condensate to the desuperheating nozzle instead.

The second reason is to eliminate a possible issue with condensation of the steam caused by the O2 injection as BronYrAur has suggested. The oxygen temp is 90 degf.

 

[davefitz]

Believe it or not, there could be other trace elements in the steam that are contributing to the damage to the 316SS. If the steam was not sourced from a pure steam supply , there could be traces of Chlorides in the steam, and there also could be minute pieces of iron from exfoiliation from an upstream superheater. Either of these would contribute to corrosion of the SS316 at 500 F + temperature.

Also, depending on the water treatment method used from the upstream boiler, there could be other volatile species in the steam, from a breakdown from the water treatment chemicals.

I am not convinced you need to use SS in this application. Have you considered an inconlnel lined or 9Cr alloy lined ferritic pipe?

 

[frostrobn]

Not a lot of experience here with the use of inconel or 9Cr lined pipe. Can you think of any ASTM specs for these? I have in the past used chromemoly in saturated steam line applications with erosion issues.

 

[BronYrAur]

What happened when you eliminated desuperheating?

 

[frostrobn]

Can't say much for eliminating the desuperheating as far as the erosion/corrosion problem goes because the unit is in service so I can't readily measure the rate of material loss. What I can say is there is too much superheat to leave the desuperheating off. The temperture without superheat can go to 400 - 450 degf which is too hot for the mixing device according to the OEM. The OEM specifies a maximum operating temp of 392 degf. My guess now is we look for a way to get boiler condensate to the desuperheating nozzle. I'll need a small multistage pump to get the pressure up over 160 psig. The desuperheating water we use presently is little more than potable water with some sodium hypo in it to kill the bugs.

 

[BronYrAur]

Do you need the 160 psig for the process to function? If not, perhaps you can install a PRV in the steam line to reduce the pressure. This will reduce the overall temperature, which will help, but it will also increase your degrees superheat. Then you can use your desuperheater to trim the temperature and still be above saturation. When the O2 gets injected, you may not condense - assuming that's the problem.

I may not have thought that all the way through, but it seems like a good idea as I sit here and type this.