316L Vessel Cracks near welds 6

[RGCook]

Hello,

I am a chemical engineer with no materials science expertise. I have a problem with a 10,000 gallon 316 Stainless steel reactor. It is 10' diameter with ASME heads top and bottom. This may sound strange but the vessel normally runs with only 12 inches of liquid (sidewall depth). I sparge a hot CO2 gas into water to perform adiabatic evaporation. The rest of the tank is vapor and the tank never goes above 180F.

My problem is that I have cracks developing near (not at) all welded nozzles and parts IN THE VAPOR SPACE OF THE TANK ONLY. The cracks DO NOT appear where the there is liquid. Only in the vapor phase. And the wierd part is, the cracks are very consistent in appearing about 4-5 inches from the welds. Not at the welds. I wonder if it might be stress due to pipes connected to the nozzles, but the cracks even appear on parts like lifting lugs and ladder brackets that have no real loads or stresses on them. Like I said, NO CRACKS at or on the welds....all cracks appear about 5" from the welded piece.

My feeling is that some strange hardening occured during welding but I don't understand why there are no cracks at the head seams or why no cracks where there there is liquid?

Any and all thoughts appreciated.

Bob Cook
Houston, TX

 

[hacksaw]

sounds like carbide depletion in the heat affected zone.

you need to check the alloy of the tank, 316 vs 316L and have a corrosion specialist examine the tank.

good luck,

 

[metengr]

The attached web site is by Outokumpu and contains a technical paper (acom 1998 Ed:3) titled SCC of Stainless Steel under Evaporative Conditions. I would recommend you download this paper and review it.

http://www.outokumpu.com/applications/documents/start.asp

Typically, SCC (stress corrosion cracking) of austenitic stainless steels requires three variables - tensile stress, susceptible material and environment. All three of these variables must be present - remove any one of them and the damage mechanism will stop. What are the constituents of the fluid in the vessel? The evaporative service conditions as mentioned in the article would be similar to "wet wall" conditions we have seen with austenitic stainless steel tubing in heat exchangers. Over time, even small levels of contaminants in the fluid deposit by evaporation and concentrate on the austenitic stainless steel tube surfaces. If the residual or service tensile stresses are high enough, you will form transgranular stress corrosion cracking (TGSCC). Under certain situations, intergranular stress corrosion cracking can occur. However, in this case exposure to aqueous solution with contaminants is typically required to cause intergranular cracking..

In addition to hacsaw's comments, it would appear you have SCC under conditions of “evaporative service” driven by either applied or residual tensile stresses from welding. The location that you report is consistent with this being a stress driven damage mechanism.

 

[RGCook]

hacksaw and metengr:

Thank you both for your insight and reference. The solution treated varies but generally consists of extremely high (saturated) levels of sodium sulfate and some sodium chloride, although generally the chlorides are below 100 mg/L.

The vessel is 316 and NOT 316L as I mistakingly put in my original post. We are taking the insluation off the remaining areas and inspecting it. I have obtained some digital pictures as well. We are in the process of finding an expert to insepect and analyze it. Some folks think that fixing it is a lost cause --that the cracsk will just occur in a different spot and not really solve the problem. I think we have the wrong metallurgy for the application. I just hope that titanium isn't the only option!

 

[stanweld]

When you stated that cracking is seen in ladder to shell attachmnet welds, were the ladders external or internal? Have you concluded that cracking originates from the internal or external surface? Note that chloridea can concentrate under insulation and chloride SCC has often occurred/initiated at welds under insulation on 316 SS at 180F. Also nozzle to shell welds have high residual welding stress and SCC is often first seen in those welds.

 

[EdStainless]

This sure sounds like CSCC from residual stresses related to the welds.
I like Stans question, are you sure that the cracks start on the inside?
Any system that has locations that see intermitent wet/dry service can have significant contaminant concentration.

If the 316 had good enough general corrosion resistance then you might want to consider a duplex alloy. The duplex stainless alloys have decent cscc resistance and high strength. The workhorse alloy, 2205 has much better pitting resistance than 316. If you don't need that consider the lean grades such as 2304, 2101, 2003. By using their strength (lighter walls) you can keep the cost reasonable.

= = = = = = = = = = = = = = = = = = = =
Corrosion never sleeps, but it can be managed.

http://www.trenttube.com/Trent/tech_form.htm

 

[RGCook]

stanweld and EdStainless,

Interesting points about potential for external origination. That is very possible since the original leak ran down the sides of the tank and dried salt can be seen in many areas. We are proceeding with expert analysis. I will keep the group posted.

In my research, I also came across the 2205 duplex candidate. This is great input guys...I really appreciate your help. When I learn more, I will respond to the group about what we found so hopefully others can benefit from the lesson.

 

[EdStainless]

There are many chemical plants on the Gulf Coast that paint the outside of all stainless equipment. Some companies even have standards for it.
I have even seen it 'up north' with salt off of roads being the Cl source.

= = = = = = = = = = = = = = = = = = = =
Corrosion never sleeps, but it can be managed.

http://www.trenttube.com/Trent/tech_form.htm

 

[unclesyd]

4" to 5" from a weld sounds a little much for any effect from welding.

What type insulation do you have on the tank and when was it applied?

If you can take a piece from a cooler section or around a support and run a water extractable CL.

 

[stanweld]

Good point unclesyd! I'm wondering if distortion of the shell due to improper welding/fitting of the nozzles into a thin wall shell may be the culprit. I saw this problem once about 25 years ago when some heavy o'let branches were welded to some thin wall 316 SS piping. In this case cracking initiated in the pipe wall about 1 1/2" from the weld.

 

[EdStainless]

My thought was that the welded attachments cause places on the skin to be cooler. This would result in more local condensation. Around each attachment I would expect a region that met three cirteria:
1. Cool enough for condensation
2. Warm enough for CSCC
3. Enought stress for CSCC
Since the last two are very easy to meet, this could happen at just about every attachment.
I don't think that the welding itself has anything to do with this. Even if it turns out to be ID cracking, the mechanism could still involve preferential condensation near the attachment points.

= = = = = = = = = = = = = = = = = = = =
Corrosion never sleeps, but it can be managed.

http://www.trenttube.com/Trent/tech_form.htm

 

[RGCook]

Folks,

Our corrosion expert has inspected the tank. A piece is being removed for testing. His first thought is carbide depletion SCC and, as stanweld suggested, many of the cracks appear to have originated from OUTSIDE the tank.

On removal of the insulation (simply 2" of fiberglass with aluminum jacket), the vertical weld seams have cracks running through them, perpendicular to the welds. The tank is an eggshell.

Testing for chlorides in the sample and running other tests. I will keep the group informed. Thanks for all your input!