Vessel Repair and Failure Analysis 1

[steris]

Hi All,

My company has a Section VIII Div 1 vessel that has been in steam service for 2 years and has developed a crack in the jacket. I attached a picture of the defect for reference. The vessel is a flat wall cross-section with a stainless steel inner wall, 516-70 jacket (1/4" thick) and A36 support bars between the vessel walls. I've attached an illustration of the cross-section of the construction. The picture from the field shows that the defect is about 1" below the seam weld. The seam weld is backed by an A36 bar that is 1.5" wide. The defect runs parallel to the seam weld/backing bar and is about 3/4" below the edge of the backing bar. I'm trying to figure out what happened here. We have thousands of vessels of this design that have been in the field for 40+ years and we have never seen this type of failure before. In terms of operating conditions, the jacket sees saturated steam service at 20-40 psig and has a proof test pressure over 1000 psig.

My intuition says that this crack is likely a material defect - either an impurity or void that was captured in the rolling process. The material receiving inspection reports don't note anything peculiar during the visual inspection about this plate. I was hoping that if something was rolled into the plate that there may have been a tell tale sign on the surface of the material but this doesn't appear to be the case. Has anyone else seen anything like this happen? What could be the cause?

In terms of repair, we were going to grind out the length of the crack plus about 2" on either end. Then we were going to drill the ends of the ground area to blunt the crack in case there's a hairline fracture that we can't see. After cleaning and other weld prep, we were going to perform a weld repair. Are there any suggestions of other items to examine or procedures to perform prior to repair?

I appreciate the assistance!

Best,
Steris

 

[fegenbush]

"...in case there's a hairline fracture we can't see..."

This is why you MT or PT the crack. I strongly recommend that you perform this to ensure that you have removed all of the cracked material prior to weld repair.

 

[Duwe6]

Strongly agree with MT/PT prior to welding. Most failed welded repairs of cracks happen because the welder left one or both tails of the crack behind. And be aware that cracks usually propagate when being welded on - thus the 2.5" crack may be 3" long after welding. If the welder stopped at 2.5", you still have some crack left.

The best way to excavate cracks is with arc-gouging - CAC, 'air-arcing'. The crack opens up due to the heat stress and the welder can see the crack fairly well, and can see when the crack has been eliminated. Additionally, the arc-gouge leaves rounded ends, unlike the sharper excavations you get with a grinder or a burr-bit. And if there are internal material flaws, like slag-lines and delaminations, the arc-gouge will find it.

 

[metengr]

I am guessing it is probably a thermal fatigue crack based on appearance. Either way, you need to remove the crack, verify crack removal using either wet MT or liquid PT and weld repair. After welding, blend grind and perform a final wet MT or Liquid PT. You do not need to drill the crack ends because it is not needed and will complicate weld repair.

 

[steris]

Hi All,

Thanks for the tips! PT is a great idea and you are absolutely right that it should be performed.

In terms of the potential causes, I would be surprised if this had experienced fatigue failure. We are running at temperatures of 60F-280F and all temperature changes are very gradual - there is no thermal shock on this vessel. Additionally, prior to failure, this vessel could not have seen more than 2,000 pressure/temperature cycles. Given that the stresses are low and the cycle count is low, I don't know what the mechanism would be to create fatigue failure. Is there a reliable way to determine if thermal fatigue was the root cause? Are there any other explanations that could describe why this failure occurred?

 

[metengr]

steris;
Residual stress from forming or handling coupled with the temperature variation in service (despite gradual), can result in this type of cracking. You would need to extract a small boat sample capturing the crack and have a proper lab analysis performed to characterize the cracking mechanism - straight, branching, transgranular, intergranular, etc.

 

[davefitz]

I agree with Meteng- it looks like theremal stress initiated fatigue. Check the coefficient of A36 vs the liner and determine the imposed strain due to DT* DA, where DT= total change in temp begtween assembly ( room temp) and operating, and DA= difference in coeficient of thermal expansion. A rough indicator of the thermal stress may be St= E(max) * DT* DA, where E is the max young's modulus. As I recsall, this same estiame is included in sect VIII div 1 related to liners.

"Whom the gods would destroy, they first make mad "

 

[weldstan]

Agree with metengr & davefitz. It would appear that free movement of the backing bar relative to the ss liner is not occuring. The ss liner is expanding at nearly twice the rate of the carbon steel so you need to design for the differential.

 

[steris]

Hi All,

Thanks for all of the feedback! I really appreciate all of the information.

I am reluctant to believe that this crack is a design flaw mainly because our company has been using this exact vessel design for 40+ years and has 1000+ vessels of this exact design in service without ever seeing this happen before. If this was a design issue, I would expect that we would have seen this failure mode occur many times before. I don't see how I can make a case for a design flaw when there is such a long and successful service history of the design. As such, I'm forced to believe that something unique happened on this vessel whether with the materials, manufacturing or use. Could this have been caused by an excessive HAZ either from plate cutting or welding? I'm wondering if a higher welding heat was used, could there be embrittlement of the base metal which led to premature thermal fatugue? Is it possible for this to be a defect in the base metal from the rolling process? Are there other explanations for this other than design flaw?

Thanks again!

 

[metengr]

steris;
Unless you remove a material sample capturing a portion of the crack, whihc you can weld repair the slightly larger excavation, you cannot determine root cause, period. Why beat your head against the wall looking for an answer????

 

[steris]

Hi Metengr -

Thanks for the advise. I apologize for repeatedly asking the same question. We will have a sample analyzed however in the meantime management and the customer are pushing hard for a possible explanation. I was just trying to give them a few possibilities to tide them over while the the analysis is performed.

 

[GenB]

Steris. I fixed several of these vessels. It could be possible the internal bar is wider and the line corrossion is along the bar, U.T. it to be sure. If condensa ion is being concentrated horizontally along the line, it will eventually corrode the entire vessel. it seems time; hat this vessel is well corroded not at the line but on other places. We scraped a few vessels(sterilizer chambers) corroded beyomd repair. Genblr.