O-RING Groove design for hydrostatic testing 1

[NBR90]

I'm having trouble with coming up with the right groove design. Briefly, I do hydrostatic testing on valves from 2" to 36". For 2" - 12" I need to test them for leakage up to 5000 psi water pressure.

I have a 200 ton hydraulic clamping system with 200 mm cylindrical rod in the middle, sandwiching the valve are two plates (36" wide) top and bottom (see pic attached).

Success so far:

* Last year, I cut groove to 7.0 mm wide, 5.0 mm deep (Using NBR 70D, 7.0 mm wide O Rings). The grooves are cut "round" as opposed to "retangular U shape" , giving it even less volume.

I was able to test to 3000 psi water pressure for 12" valves using about 90-100 ton clamping force.


* This year, I cut another plate with 6.5 mm wide groove and 6.0 mm deep (Using NBR 70D, 7.0 mm wide Orings). The grooves are cut "rectangular" giving it more volume than 'round' cut. I was able to test up to 4000 psi water pressure for 4" valves using about 50 ton clamping force.

But with the 6.5mm, 6.0mm cut, anything above 6" fails miserably at 1800 psi and below. Sometimes it pops out, most of the times it just leaks.


So I need to re-cut my grooves.

Some have suggested that I go with a groove that is 8.0 mm width and 5.0mm deep (7.0 mm Oring) - enough volume in the groove to swallow all of the Oring.

The problem I may encounter is that my clamping force is not uniform & exacting like a bolting series. I'm relying on a 200 mm wide cylindrical rod in the middle to 'clamp' it together. Therefore, any gap inside the groove may cause all sorts of unplanned leakage, especially at higher pressure (2000-4000 psi). I'm sort of relying on the extruding O ring for 'tolerance' adjustment, sort to speak.

On the other hand, I know that elastomer are incompressible. And that I may have different problems such as extrusion, swelling, etc, etc, if I don't give it enough volume for it to do its work (I've read that Oring should only be 80-90% maximum volume of Groove Volume).

That said, I do have empirical data that putting the Oring inside groove it will not fit and under 100 ton of compressive force DOES NOT destroy it. Nor does it make the seal fail. In fact, it was sealing 12" valves pretty nicely up to 3000 psi (I'v never tested beyond0. Again, in that set up, I cut the Oring groove 'round', 7 mm wide, 5 mm deep using 7 mm NBR 70D Orings.

So, I know I need to re-cut my plates. The question is, how?

I'm thinking either:

7.0 mm wide, 5.0 mm deep, rectangular U shaped. Using 7.0 mm NBR 70D Orings.

or

8.4 mm wide, 6.0 mm deep, rectangular U shaped. Using 8.4 mm NBR 70D Orings.


Any help, suggestions is appreciated.


Yes, I should use my old plates, but they are too small, I designed the new bigger plates according to a 'suggestion' for better seal by a specialist... well..

 

[israelkk]

To my opinion the plates are too flexible. An FEA analysis of the plates clamped by he bolts under pressure will reveal how much the plates deflect thereby allows extrusion of the O-ring or the leak. The plate/clamping should be rigid enough to prevent deflection of the plates between the bolts.

Why not connect the valve to the a short piece of a standard compatible pipe with a welded standard flange and close the other end of th pipe? If the combination of valve and flange doesn't hold the 5000psi then the system is self protected against a pressure eise of 5000 psi. The o'ring/packing then will be the fuse.

 

[NBR90]

Yes, perhaps plates are the problem. The plates are about 60 mm thickness. 960 mm in diameter. The older plates are only 450 mm in diameter but 65 mm in thickness.

We do use blind flanges for emergencies but when you are testing hundred of valves, it becomes impractical to bolt them, especially 2500# Class valves - pus the gaskets are expensive, single use.

I've these seen these plates at work with similar set-up, they even used 50 mm plates. I don't expect the plates to hold much pressure past 12" - but at least they should suffice for 100 psi air pressure tests.

 

[byrdj]

the deepth of the groove should be a given for the cross section of the O-ring, to give % compression. then the width to allow for the o-ring deforming when compressed. I'm not accustome to dealing with metric desiginations so forgive if the values you state are per design. another importaint consideration is the durometer of the o ring material. for applications where the joint gaps and durometer of the o-ring are liable to extrude, a backing ring is needed. I would suggest consulting an o-ring design catologe like Parker's

http://www.parker.com/literature/ORD 5700 Parker_O-Ring_Handbook.pdf

 

[GregLocock]

Agree with byrdj, read the book. Your grooves are too tight.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[unclesyd]

I'll have to agree that your end plates are way too thin for the larger valves. The plates on our valve tester capable of 2400 psig for valves up to 24" were around 4" thick and were supported on a frame using 4" guide rods. The machine was horizontal built like a filter press frame.
As stated above use the catalogue dimensions for a particular O-ring. Most O-ring manufacturers have excellent technical service departments that will give you assistance in selection of material and design of the seating groove.

Take a look at some of the flanges on the Barbee Testers.

http://www.barbeetesters.com/s5.htm

 

[btrueblood]

The "new" flange grooves (7mm ring dia., in 6mm groove depth) were too deep, you only had about 15% compression, which is marginal for a static seal. The o-ring should be compressed to a nominal 25% for high pressure static seals, at least that is the value we design to, so a 5mm groove depth is probably going to work a bit better. As others noted, use the groove dimensions and details in the mfgr. handbooks. If the rings are extruding, consider using a backing ring, or contouring the o.d. side of the groove, or increasing the seal durometer, to help the ring hold its shape better.

We've used thinner flanges than what you've described, but not above about 900-1000 psi. We also don't rely on a single element to provide clamping for hydro tests, but then we tend to be a bit cautious about human lives.

 

[rob768]

25% isn't too much either, you might go further, up to 30-35% imo. Also, have you used a harder ring (higher shore hardness)? you need high hardness for high pressure. The mentioned Parker site is a very good one.

 

[NBR90]

So for 7.0 mm Oring, should I cut it 7.0 mm wide and 4.5 mm (or 5.0 mm deep)?

I would like to cut it 8.0 mm wide, 4.5 mm deep (Similar to what Parker suggest I tihnk) but then all the O ring will stay inside the ring. There is also the Dovetail design, but again, this design will swallow all the O Ring when compressed. I think this will work great for bolted clamping system. But for hydraulic clamping, I don't know if this will work.

I want some extra rubber outside the groove to help the seal. If some parts of O Ring are ouside, there's no popping out because it will have free space to expand laterally instead of being confined. Well, up until internal pressure overcome the hydraulic pressure. Then it is just matter of increasing the hydraulic pressure.

I had success with this set-up up until 12" valves and 3000 psi water pressure. But then I had to use about 90-100 ton of clamping pressure. Groove: Round cut - 7.0 mm wide, 5.0mm deep. Oring 7.0 mm NBR 70.

 

[byrdj]

I read that you are using the oring like a gasket, and your claming device results in unequal clamping force to compress gasket around its perimeter

if you cut the groove as per the dimensions for a "Normal" oring groove so that when the oring is compressed, the flanges go metal to metal. then the internal working pressure that makes it past the inside metal to metal will press the oring against the OD of the groove to flange and seal.

If your clamping device doesnot result in the flange going metal to metal on the full perimeter, you can adjusts its orientation so that equal loading around the perimeter is achieved

 

[GregLocock]

NBR90 'So for 7.0 mm Oring, should I cut it 7.0 mm wide and 4.5 mm (or 5.0 mm deep)? '

Design it like the manual says. High static pressure O ring design is trivially easy- do it like the book.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[byrdj]

page 98/292 of the parker design handbook previously posted gives the following recomendation for a static gland

a 7mm cross section oring would be -400 series oring (0.275")
the gland depth L would be 0.201 to 0.211" (5,2mm)75%
the gland with W would be 0.342 to 0.362' (8,9mm)128%
the gland OD would be the O-ring OD +0/-0.060"

 

[NBR90]

Thanks all for your kind help. I will eventually follow Parker's guide - but then we have to do a Dovetail groove because else the Oring will just fall out from the top plate. Our machinist is not so fond of doing Dovetail, says it is very difficult. Waiting on the bit....

Right now, I'm going to do a little experiment and just shave 1 mm off the plates to 7.0 mm wide, 5.0 mm deep and let the O ring be a 'gasket' as byrdj noted.

Yes, I will stand far away while testing. Although, all the times that it went 'pop', only splashes of water came out. But I understand high pressure is nothing to be messed with. Air - anything above 100 psi - is what I'm afraid of, not water (under 5000 psi).

 

[btrueblood]

Agree. But it's sometimes hard to get all of the air out.

 

[Compositepro]

Using an O-ring as a "gasket" at high pressure is not a a good idea. I'm not clear on your geometry but the gap that the ring is to seal should be as close to zero as possible. And that gap should not be changing with pressure.

 

[byrdj]

Your pic is not visible, so I'll do some more assuming.

you say the
the inservice gasket is a one time use
so I'll assume the valve flange are flat and use a steel ring backed flexatalic type gasket.

you also say the oring goes in an upside down flange.

why not make a backing ring plate, that is 5mm thick. then the manifold and valve flanges can be flat. if OD backed flexatilics are what they use, you could afford to destroy one to get the ring, clean the ID and use an oring that gives about 25% compression.

 

[Compositepro]

I think byrdj has a very good idea.

 

[byrdj]

 

[EdDanzer]

The Parker dimensions usually work well. Sometimes a quad ring or square ring will work better than an o-ring in face sealing.

You may find Inch size seals a better choice as many more sizes, materials and configuration may be available. You should also measure the metric o-ring to make sure of its dimension are correct for the groove.

I don’t think NBR is the best material for sealing water.

Ed Danzer

http://www.danzcoinc.com

http://www.dehyds.com

 

[NBR90]

Ed,

What would be the best material for sealing water? Viton?

Thanks.