Tricky Scratch fix 1

[jakejas]

We are trying to fix a scratch on the sealing surface of the head of a stainless steel pressure vessel. It needs to be blended to within 1 mil/inch. Now comes the difficult part. The vessel is huge, and due to schedule restraints, we are trying to fix it while it is on its stand which is 15” above the ground. There are multiple scratches that need to be fixed, and for various unrelated reasons, the workers can not be working for a long time near the vessel head. We made a mockup and have tried various grinding/buffing methods to try and remove scratches we put in the mockup, but are having difficulties blending out the scratch within 1 mil/inch. Any ideas would be greatly appreciated. We may have to do a weld buildup on a few of the scratches.

 

[racookpe1978]

Get a photograph of the scratches.

Each will have to be addressed as a separate case: what will work for a few will not work for all. What will be needed (potentially a pre-heat, weld-buildup, PWHT (maybe) and then machining) for the worst "scratches" will not be needed for the easiest.

Have you done a blue-check for the cover to know exactly where the low spots are?

What is your cover and how do you seal the surface? (Gasket material is?) What torque on the fasteners?

Radiation area?

 

[jakejas]

Radiation area, bingo. I don't have photographs of the scratches, but I am working night shift and will see if we have any and will post them if we do. The vessel is sealed by two stainless steel seal rings. The scratches happened when one of the rings got stuck and the maintenance guys tried to use metal objects to remove it... Our vendor technical document says that any scratches over 7 mils will have to be weld built-up and machined down, but anything under that can be buffed out. The issue we are having is that is it is difficult to machine out to within 1mil/inch in the field when we are trying to use ALARA practices in a high radiation area.

We have not done a blue-check, and I am not sure if we can or not because of the possibility of getting some of it in the primary system. The studs are pre-tensioned. I just took a look at the vender manual, and they want the studs stretched 77.6 mils which I believe is a force of just under 2,000,000 lbs per stud.

The main issue we are trying to work out now is how to blend out the surface to a smooth taper for 1" for each mill of indication.

 

[hydtools]

Old school file and scrape?

Ted

 

[jakejas]

@hydtools, we tried using a hand-tool that had a rough polishing wheel (the kind that looks like a circular scour pad that you use in your kitchen) on it, but when we put a dial indicator on it and moved it around, the repaired area depth changed more than 1 mil in a distance of less than an inch. Do you know of a file and scrape that can perform better than a rough polishing wheel?

 

[jakejas]

The hand tool that we used was a pneumatic tool.

 

[MikeHalloran]

I think hydtools was asking about a hand tool, not a handheld power tool, which is likely to do more harm than good.

See if you can find an old machinist to teach your people how to properly draw file a surface to flatness.]



Mike Halloran
Pembroke Pines, FL, USA

 

[GregLocock]

I was going to suggest asking an automotive tool and die fabricator, as they'd be used to repairing the horrible things we do to stamping tools.

1 thou per inch doesn't sound too onerous, but that's on a workbench, not 15 feet in the air and glowing in the dark.



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[jakejas]

Ok, thanks. I will ask our machinists if we can draw file the surface. Sorry for the confusion, I am a young engineer and am not yet familiar with all of the terms.

 

[jakejas]

One of the options they are considering is using a stone.

 

[racookpe1978]

I regularly use stones for turbine seating (metal-to-metal) contact surfaces. Power tools - unless locked into precise alignment as with a portable milling tool fixed to the surface - are too likely to remove dips and valleys.

Use a long flat stone (10 inch long x 3 or 4 inch wide, 1 inch "deep" or more to be stiff against the bending pressure as you push it down against the metal). The problem will be the sealing surface: If you use a file it will produce hundreds of small parallel scratches as you wear off the metal.

Unfortunately, both stones and files (any grinding method for that matter) doesn't remove scratches - they remove the high parts and wide areas between the scratches so the net average metal height is slowly reduced to the bottom of the scratch. But that "slowly" term means greater radiation exposure.

Weld buildup put a very thick "bump" into the "scratch" - which then means that new "mountain" then has to itself be ground down so the final surface is more nearly flat. But any undercut or edge effect at the sides of the weld to the base metal at the weld will form yet another groove in the surface.

Of course, any scratch that does not cross the whole surface - anything that does not provide a leakage path from inside to outside of the pressure boundary - will not (theoretically) leak.

 

[hydtools]

MikeHalloran, exactly. Hand file and scrape. Drawfile with a single cut mill file to clean up and successive fine finishing with a scraper for necessary finish. Use a file card frequently to clean the file teeth so as not to ruin the file work.

Ted

 

[potteryshard]

Assuming that the sealing surface is horizontal, might it be possible to rig a rotary drive to turn the head cover? Guide pins or rollers around the edges might be needed. The weight of the cover should be enough to allow lapping compound to flatten out the mating surfaces over time.

This might appear to come from somewhere in loonie-land, but once rigged, it could be left to run with only remote observation.

I'm envisioning some sort of drive spider made to fit the cover holes, a motor and gearbox either adding its weight to the cover or suspended from a crane hook, and with a torque arm (or arms) restrained by chains or links to nearby structural supports.

Okay; I'll go back to my corner and talk to myself.

 

[MiketheEngineer]

Fill them with BONDO - sand and paint and you are done.

Do these scratches structurally affect the tank. If not - BONDO!!!!

 

[vpl]

MiketheEngineer

I don't know if you were being serious or not, but I don't think that jakejas should be using BONDO on a nuclear reactor.

Patricia Lougheed

******

Please see FAQ731-376: Eng-Tips.com Forum Policies for tips on how to make the best use of the Eng-Tips Forums.

 

[Ceramicguy]

Considering potteryshard's suggestion; can you get a known flat surface plate up there, invert it and lap the scratched areas? It will give a nice blend meeting the .001 inch per inch localized flatness requirement. It could be taken off-site between lapping operations to reflatten the lapping face using traditional surface plate truing methods and then put back in service. You can vary the abrasive size to select from high material removal rates to fine polishing with virtually no material removal.

Bruce

http://www.accuratus.com

 

[SnTMan]

jakejas, a hard stone in experienced hands will take down a stainless surface faster than you might think. Of course in your situation it might not be "fast".

potteryshard's lapping idea might have some merit, if possible.

Regards,

Mike

 

[jakejas]

Thank you all for the input. We are going to try a bunch of methods on our test specimen that we have rigged up and go from there. I also suggested J-B Weld or wait until it starts leaking and then use furmanite injection to stop the leak, but those ideas didn't go over so good...lol