Out of Roundness on Pressure Vessels After PWHT

[izamil]

I am sorry if this is long, but please bare with me.

We are currently facing a slight problem with one of the vessels that we have manufactured for a ýclient. The problem lurks in the results of heat treatment performed in a way where the vessel ýturned into an egg shaped object. ý

For your information, because the vessel is so long, we converted the vessel into a furnace. The ývessel size is the following: Inside diameter 4,877mm X 48,768mm tangent to tangent in length X ýý44 mm thickness. ý

The plates used are supplied by the mill with simulated PWHT.ý

The vessel was treated in the horizontal position with four units high velocity burners. Two on ýeach head end, and the other two from the top spaced equidistantly on the remainder of the ýshell with heat being diverted horizontally by means of a specially made 22 degrees diversion ýelbow. ý

We have placed 13 units outside saddle supports with none inside and started firing. Upon ýcompletion @ 50 degrees Celsius an hour for a total temperature of 610 degrees Celsius areas ýof the top third of the vessel fell to about 210mm from its own weight and nozzles. Maximum out ýof roundness per code is 48mm. ý

I kindly seek your expert opinion for a method statement and or a repair procedure to resolving ýthe current situation so as to bring this vessel back to code compliance requirements. ý

The code is the latest version and addenda of ASME section IIIV Division 1.ý

Our engineering manager stated the following:ý

The dimensional reports attached with NCR P01-1432/02 (Q11-D-204) for out of roundness of ýthe vessel after heat treatment indicates the following:ý

ý1.ý The heating in my opinion was not uniform at 4 burners as the result of this there may ýbe thermal pockets generated due to differential expansions at the burner’s locations. ý
ý2.ý This report does not indicate shifting of TRUE VESSEL center line of vessel towards 90° ýor 270°. The vessel might have bent like Banana or at some portion bent towards 90° ýand remaining portion bent towards 270°. If the vessel is bent towards one fixed degree ýthen turning the vessel by 90° say and doing heat treatment would reduce out of ýroundness. If the top half portion is bent and maximum deflection is at the center of ývessel length and towards one fixed degree than it indicates the introduction of internal ýspider supports. These spider supports shall be inter-connected so that it acts as a ýhomogeneous structure. This is one of the possibilities. Complete readings have to be ýtaken at 1 meter interval or at the shell course weld joints. ý
ý3.ý Jacking of shell would not help unless vessel is jacked beyond yield strength. To yield ýý44 mm thick shell, it requires a very big jack. Jacking of vessel would lead to jacked ýspots and it is not recommended by the writer. ý
ý4.ý Heating rate and cooling rate shall be faster as to avoid prolonged thermal stresses. ý
ý5.ý How this material behaves at 1100° F (HEAT TREATMENT) temperature is very difficult ýto assume. ý
ý6.ý Cutting the shell weld joint would be one of the disliked alternatives. ý
ý7.ý It is ad visible to take expert opinions particularly from metallurgical engineers pertaining ýto support etc before it is too late. ý
ý8.ý It is beyond my engineering capacity to deal with thermal expansion calculations, ýstiffeners outside or inside the vessel. ý
ý9.ý This problem may be new to us, but definitely not a new problem for thick vessel ýmanufacturers or thick vessel inspectors and they may give vital information about ýarrangement of internal stiffeners. We do not want to do it wrong this time. ý


I would appreciate any and all assistance you could provide, ý

 

[unclesyd]

From your post and my understanding of it I think you have lost the upper 1/3 of the vessel and probably haven't PWHT the other 2/3s'.
The area in question got well above 610°C for an extended period of time.

Before I go any further please post the material of construction?

Where are you located? There are people that do this on a daily routine.

 

[izamil]

Thank you kindly for your very prompt reply. ý

The material of construction is SA 516-70 HIC tested with successful results on ýsimulated heat treatment carried out on coupons of the material at the mill. ý

We have placed 36 thermocouples throughout the vessel. None of which indicates a ýtemperature above 610°C.ý

I am located in Jeddah, Saudi Arabia, and I sure do hope that I have not lost the vessel. ý

 

[unclesyd]

I assume the you have your insurance representative involved.

You mention taking measurements at 1 meter intervals, has this been done? These need to be taken of the center line taken from the bottom support. This will determine what section is the most affected. Hopefully you will be able to replace a small section if this is necessary.
How many feet of the column are affected. It would be of interest to know where the top head center line is in respect to the base.
How many nozzles/sizes are there in the area that caved in? How much weight was acting on the shell while heating?

The AI that I’ve worked with for years just told me that from your description and as he sees it the out of roundness section will need to be replaced and the entire column re-heat treated. I asked him about overlapping sections and his comment was he was afraid that the PWHT wasn’t uniform throughout the vessel.

A coworker that flame straightens steel fabrications said it might be possible to pull some out of straight section, but he wouldn’t try anything with the shell in the areas of the nozzles.

Being the weekend in the US it’s hard to get people’s advice. There are others that normally respond to code issues that should come on board with some more opinions.

 

[Metalguy]

This is a realitively thin-wall vessel considering the large dia. I'd make sure that there were T/C's that were accurate where it sagged, and if it wasn't overheated past 1,300 deg F. I'd heat it back up a bit and jack it round. I think you'll find the amount of strain is pretty low. Then hydro it.

 

[izamil]

Thanks unclesyd. ý

I hope we will not need to deal with the insurance representative. ý

The 1 meter measurements have been done, and results of which will be counter ýchecked tomorrow morning (that is our morning/your night time). I am sorry if i did ýnot clear this matter before, this is a horizontal vessel and not a column. ý

I will obtain from our engineering the number, size, and approximate weight of ýnozzles on the top portion including the reinforcing pads and clips and I will post ýthem here.ý

I would appreciate all of your assistance, and I do not know how to post the reading ýresults on this website. ý

I cannot thank you enough people for rendering a helping thought, ý

Kind respects, ý

 

[unclesyd]

Being horizontal might help you.

What are you design conditions for this vessel?

Metalguy,
Heating and trying to jack at that diameter in proximity of shell penetrations is going to be tough. Can you add any thing in the ways and means to accomplish this. I’ve done it on a much smaller scale 60" and 3/4" plate, this was chore as we had to hold one side at the correct curvature.
It required some very big contoured shoes.

 

[Metalguy]

We have made several 516-70 cylinders of ~5,000 mm dia. and 32 mm thick. They sag at room temp. about an inch elastically at room temp. We use a spider arrangement-12 legs, spaced about 1,500 mm apart to keep them round during welding.

Now, since the YS of CS drops off very fast with temp, if he heats his vessel to ~300 C he will probably have a YS around half of the RT YS. With a bunch of work and calcs., he can make that vessel round again.

We once had to jack a much thicker piece of 50 ksi YS P3 steel, about 5,500 Mm dia. and 165 mm thick. One 100 ton jack worked fine at 300 deg C.

 

[izamil]

Design Conditions

Design Code ASME Section VIII DIV1, 2001 ý
Internal pressure ý ý300psi gauge ý
External pressure Absolute @ 300º Fý ý7.5psiý
Temperature ý ý240 º F
MDMT @ 300 PSIGý ý35 º F
MAWP @ 240 º F ý300psigý
MAWP @ 73 º F ý300psigý

 

[deanc]

See boiler post. Scrap it and start over. This sounds like the more you work on it the worse it gets. So--cut your loses now.

Internal spiders are good,get someone that knows PWHT to help you.

I probably would not accept this as a inspector or buyer.

 

[Metalguy]

On what grounds would you reject it, unless it has actually been overheated?

 

[deanc]

Read the post completely about the problems. To repair this will require major rework. Whats the point? This supposed to be new. I think that this vessel is so distorted,esp. in the nozzle area,that to meet the original requirments would be unrealistic.

Would you want it at your location? Now-yes repairs are permitted for new product. It will have to meet all the requirements. As an inspector I would first have to agree to the repair methods,then buy off on the result. It would need to be next to perfect after all this. But this is an opinion group and my opinion is that it is not cost effective.

 

[Metalguy]

Sounds like there's a lot of steel and welds in this vessel, with the problem limited to one area--"Cutting the shell weld joint would be one of the disliked alternatives".

The above statement does not make sense to me, unless it was supposed to be plural (welds).

In any event, if this company has made a few large vessels and just screwed up this one, they should be able to fix it. But it they're just starting out building such things, I think we need to wish them good luck-"It either meets the code or it doesn't".

 

[deanc]

We are not so much different in opinion Metalguy. I would like to thank you for your posts as you have some good advice.

This vessel? I have become hard nosed over the years. Yes,think they got in too deep. But thats how one learns.

Try for a fix,yes that would be acceptable,but with reading all of the post think it may be cheaper to start over.

 

[unclesyd]

Starting over is always an option. We have built, heat treated, and installed the top sections of 12' dia columns in place, with very few problems. A horizontal vessel should be any problem.

The owner always has the right to refuse the vessel if not to code as this one apparently is not. Depending on some good measurements to ascertain the amount of out of roundness and its' extent, repair may be a valuable option if the owner and jurisdictional authority will agree. It's something that needs to be looked at and conferring with the owner and jurisdictional authority. The nozzles usually are a high cost in the fabrication of a vessel, so they deserve a close look.

People on this end have said the vessel maybe repairable and if economical and no undue time restraints they would accept it after a complete PWHT. They feel that the PWHT was accomplished. It was also stated that any problem of a metallurgical nature uncovered during repair would be cause for immediate rejection of the vessel in its' entirety.


izamil,

Didn't you have a thread concerning another vessel that got out of shape during PWHT?

 

[izamil]

Gentlemen,ý

Thank you dearly for all the help you have been providing. I greatly appreciate it. ýThis is not the first vessel we do, we did many before. As a matter of a fact, on this ýparticular contract, this vessel happens to be the first one out of three identical units. ý

I have PWHT the first with the posted results. Obviously, I have stopped PWHT the ýother two units until I find the route cause of this unfortunate and costly incident. ýOnce we determine the route cause, then we can start checking on a fix. I recon I ýshould be asking a "heat treatment support engineer" for assistance too. I wonder if ýyou could help me with this.ý

I have posted the same issue on different forums (each forum with its specific area of ýspecialization). ý

Naturally, this specific post concerns Metal and Metallurgy engineering Forum. I was ýhoping to find ways and means from within this forum to check where I went wrong. ý

The material is SA 516-70 HIC tested and the thickness is 44mm. We taken some 290 ýhardness readings (after heat treatment) on circumferential/longitudinal and base ýmetal seams. Results of which is a range starting from 165 – 192 BHN. ý

Whilst PWHT, we had 24 thermocouples (K-Type) spread all over the vessel at ýpertinent locations. I am wondering if the 24 thermocouples were too little? ý

The charts show very good results. ý

I fear that we may have been substantially slow in performing the PWHT, as the ýwhole cycle of PWHT took about 5 days. I think we should have been finished with it ýon the first 2-3 days. This could be a cause of the problem, but I don’t know this for ýsure. Maybe you could kindly help me with this. Does such deformation have to do ýanything with creep (temperature versus duration)?ý

I made a drawing showing the vessel required dimensions prior to manufacture. In ýthis same drawing I have tabulated and sketched out the as built dimensions showing ýgraphs of actual diameter readings versus required and allowable. Similarly, in this ýsame drawing I have plotted the dimensions after PWHT. I wish I am able to post it ýon this website for your viewing to give you a better vision of the situation. ý

The client wants me to fix it due to time constraints, but he wants me to have it per ýASME VIII Div 1 code. ý

Having said all this, I list the repair scenarios available to us at this present time: (feel ýfree please to comment and or add)ý

ý1)ý Turn the vessel 90º, stiffen it with temporary supports from the inside (to the ýrequired dimensions) and perform PWHT again at 620ºC. this time however ýto complete within 2-3 days. ý
ý2)ý Normalize the vessel by elevating temperature without a stop to 950ºC within ýa recommended amount of hours, then open up the insulation ýimmediately to the open atmosphere (30ºC) and allow it to cool down in ýthe hope that it will revert back to its original preformed shape.ý


Any input you give me will be very much appreciated.

 

[unclesyd]

Stop I don't think this vessel will survive the 950°C temperature even with supports.
Let's get a hold of some people in the business that you have as a consultant.

One that I know is Cooperheat that does field service work. They did have an office in the Mideast. I don’t know if they will do consulting work or will want to do the whole thing.
They are a couple of more that work in you area.

 

[deanc]

izamil-Strongly suggest you stop now on this repair. You must read SecVIII Div 1 and all of UCS-56.
1.Contact your Insurance Inspector
2.Contact a consultant as unclesyd suggested.
3.You may be able to salvage the ends and nozzles of this vessel now. Unlikely after your next step.

With all due respect,you need experienced help with this. You need it "on-site". As is,you risk a major loss or incurr a long term liability.

 

[Metalguy]

If ever a company needed outside "expert" assistance, this is the one! To even think about trying to normalize something like this is amazing to me.

OK, not to change the subject, but let's see who's the first to correctly state the reason for the code out-of-roundness requirement. Ready, go!

 

[5240]

I think you must do an overall assessment of the actual out of roundess condition and the straightness.
Do you have internal tray support rings? If yes, things are even worst.
ASME Sec VIII Div.1 allows 1% dia for out of roundness and 2% for nozzles section. Use those figures to evaluate your current possition.
Sorry for your current possition.
Have you given sliding condition during PWHT? The vessel must be able to slide during PWHT and the 13 saddles may have restricted the expansion which caused it to bend like a banana.
fabricator