Find the flaw on this design 7

[vonz33]

Hi there

I'm building a few pressure vessels and having some issues with one of them, here is a video of a failed hydrostatic test

https://goo.gl/photos/d3mvpcHzmzNXnp1V6

Here are the drawings, hope it makes sense

https://drive.google.com/open?id=0By1l7HueUo__SXJyYlB2QnZkVzQ

Here is the sequence of the test assembly in the video

EndCap(10mm)>Gasket(3mm)>Spacer(25mm)>Gasket>Plate(0.7mm)>Gasket>Spacer>EndCap

Note that I have not recorded the torque on the studs, all I can state is that my DW059 was maxed out with a full battery here are the specs

http://www.dewalt.com/products/powe...m-18v-cordless-xrp-impact-wrench-kit/dw059k-2

I think the issue is from the plate, this was not my design just tried to save costs on engineering... I have no clue why there are slits instead of holes for the fasteners, if anyone have an idea why the designer of this plate would have made it that way, please let me know... It canot be to ease the assemblies because if its a hole (fully surrounded by plate metal) assembly is not really an issue even for a big stack.

Cheers

 

[prex]

From your drawings it appears that the middle plate is withstanding the pressure on one side, as the holes in it are blocked by the spacers. As this plate cannot of course withstand the pressure it bulges so the gasket leaks and the leak comes out from the bolts because the gasket is still compressed at the periphery.

prex
[URL unfurl="true"]http://www.xcalcs.com[/url] : Online engineering calculations
[URL unfurl="true"]https://www.megamag.it[/url] : Magnetic brakes and launchers for fun rides
[URL unfurl="true"]https://www.levitans.com[/url] : Air bearing pads

 

[vonz33]

UPDATE: I just ran a test under 30psi without the plate and water is still pouring out from the inside of the nuts, so it means that some fluid is jumping from the inside of the gasket through bolt holes.

I shall conclude that this means that there is not enough pressure on the studs.... Studs are tightened at 300 ft-lbs and inside of gasket fails below 20psi

I will give it a good night of sleep and try to think fresh in the morning....................

 

[vonz33]

prex, if you look at the top corner of the gasket you will see a notch this is where the fluids can equalize from one side to the other side of the plate, the gaskets have to be assembled face to face, of course.

Compressed at the periphery how is that possible since the end caps are supposed to distribute the torque on the bolts evenly at every point of contact between all parts... or am I missing something...?

I have not studied mechanical engineering obviously.... but keen to learn more on the subject!

 

[LittleInch]

What's the purpose of this thing if it allows flow across it, but only by a strange method of a hole in the gasket. Gaskets compress and I can't see what the purpose of those holes are in the corner of the mid plate and gasket.

what sort of pressure are we talking about here for design.

Those bolts look pretty weedy to me. At 20 psi and approx. dimensions you've got a force equivalent to 200 - 250 kg on that plate.

Some of the nuts aren't even fully engaged on the bolt.

You need to make sure the bolts don't interfere with any of the holes - top plate looks misaligned.

I'll be honest here the words "I'm building a few pressure vessels and having some issues with one of them" and "I have not studied mechanical engineering obviously" look like a potential disaster in the making.

Pressure vessels are not for the uninitiated, even small ones in a lab can kill you or someone else. Failure under pressure can be very violent, even when only filled with water.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[SnTMan]

vonz33, I won't tell you I understand your design, I cannot work up the interest to wade through all the material you provided. But I can share some of my experience with non-circular pressure vessels.

First, check the obvious things. Bolts properly tightened? Parts reasonably flat? Gaskets in good shape? Gasket surfaces in good shape? Machining of non-circular gasket surfaces usually results in tool marks directly across the gasket surface, a potential leak path.

Second, you appear to be using a kind of "full face" gasket, that is it extends outside the bolt pattern. This design can be more troublesome than a "narrow face" gasket, that is, one that is contained within the bolting. Two things happen: a) The gasket material that is outside the bolt holes absorbs bolt load, but provides no useful sealing. Fluid gets to the bolt hole, you are done. So, you should assure yourself you have sufficient bolt area to properly seat the entire gasket.

b) When pressured up, due to their deflection flat plates tend to rotate about either the bolt pattern centerline or worse, the gasket outside edge. Either way, it loads up the gasket outside edge, which provides no useful sealing, and unloads the inside edge which is where all useful sealing takes place.

In a narrow faced gasket design the plate tends to rotate about the gasket centerline. While this can unload the gasket inside edge, it further loads the outside edge, helping to maintain a seal.

I am not going to tell you this is what is happening, or that I can prove any of the foregoing, but I've seen it.

If you can assure yourself that all else is in order and you still have leaks, you might try making a set of narrow face gaskets by cutting along the line of the inside edge of the bolt holes, discarding the outside portion.

Regards,

Mike

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[mk3223]

The problem could be the combination of the gasket surface roughness issue, gasket type, bolting location and tightness, etc. The typical standard flange has a surface roughness of 200 AARH for "biting" into the gasket. The location of the bolting with the correct tightening procedure and torque may be required too.

 

[Compositepro]

Did you even use any particular procedure for tightening the bolts? With this type of seal you must finger tighten all the bolts and then in small steps tighten bolts on opposite sides so that the plate compresses the gasket evenly.

 

[chicopee]

Didn't a similar problem occured with flat head V8 engines on Ford vehicles whereby the bolts had to be tightened in a pattern similarly described in Compositepro's response?

 

[vonz33]

In response to LittleInch:
>What's the purpose of this thing if it allows flow across it, but only by a strange method of a hole in the gasket. Gaskets compress and I can't see what the purpose of those holes are in the corner of the mid plate and gasket.

I am unable to reveal more of the design any further, I apologize.

>what sort of pressure are we talking about here for design.

Sorry I thought I had mentioned this variable previously and it was omitted. The working pressure would be only 150psi but we have to rate the unit for 300psi.

>Those bolts look pretty weedy to me. At 20 psi and approx. dimensions you've got a force equivalent to 200 - 250 kg on that plate.

We had to TIG weld the nuts to the rods to ease the process of tightening the assembly, it make them look odd. It's normal. However you are perhaps bringing my attention to the fact that the studs are too narrow for a vessel that is meant to work under pressure, is that what you meant?

>Some of the nuts aren't even fully engaged on the bolt.

Well.. I perhaps done a pretty rough job at tightening them and should have used a torque wrench to make sure we're all even. Meanwhile, that end cap is pretty rugged and should help the force to be spread in an even manner as it would not disform that easily.

>You need to make sure the bolts don't interfere with any of the holes - top plate looks misaligned.

This is a relatively simple assembly and we should not need laser precision to assemble it successfully. If we do then we may look into a different shape for the whole unit. Perhaps something cylindrical with disc flanges as end caps, 16mm bolts... It looks like we may have to go down that road..

>Pressure vessels are not for the uninitiated, even small ones in a lab can kill you or someone else. Failure under pressure can be very violent, even when only filled with water.

Thank you for your advice, we understand your concerns. We provide safe work environment and comply with OHS rules, we keep all records, everything is in order.

 

[vonz33]

>SnTMan (Mechanical)9 May 17 20:23
>vonz33, I won't tell you I understand your design, I cannot work up the interest to wade through all the material you provided. But I can share some of my experience with non-circular pressure vessels.

Thank you.. This design is not patented and we can't say too much about it, but it is indeed a pressure vessel and responds to the same mechanical rules.

>First, check the obvious things.
>Bolts properly tightened?

VERY TIGHT

>Parts reasonably flat?

POLISHED STEEL vs CLEAN RUBBER

>Gaskets in good shape?

JUST MADE FRESH

>Gasket surfaces in good shape? Machining of non-circular gasket surfaces usually results in tool marks directly across the gasket surface, a potential leak path.

I have just ran on one of our older prototypes (the same 30psi test) that has very similar parts but is not to work under more than 5psi, the bolts were sweating from the inside just the same way. It is indeed a design flaw.

>Second, you appear to be using a kind of "full face" gasket, that is it extends outside the bolt pattern. This design can be more troublesome than a "narrow face" gasket, that is, one that is contained within the bolting. Two things happen: a) The gasket material that is outside the bolt holes absorbs bolt load, but provides no useful sealing. Fluid gets to the bolt hole, you are done. So, you should assure yourself you have sufficient bolt area to properly seat the entire gasket.

The older design I mentioned above has full face gasket and plate, and it fails the exact same way.

>b) When pressured up, due to their deflection flat plates tend to rotate about either the bolt pattern centerline or worse, the gasket outside edge. Either way, it loads up the gasket outside edge, which provides no useful sealing, and unloads the inside edge which is where all useful sealing takes place.

Would this assembly be more naturally prone to work more reliably under pressure if it was a cylindrical by any chance?

>In a narrow faced gasket design the plate tends to rotate about the gasket centerline. While this can unload the gasket inside edge, it further loads the outside edge, helping to maintain a seal. I am not going to tell you this is what is happening, or that I can prove any of the foregoing, but I've seen it.
If you can assure yourself that all else is in order and you still have leaks, you might try making a set of narrow face gaskets by cutting along the line of the inside edge of the bolt holes, discarding the outside portion.

If all this esoteric activity can happen like that in a non cylendrical vessel, then we may let this "boxy" shape behind for cylendrical, what do you say?

Still... my next cue in this troubleshooting will definitely to cut some new gaskets with a wider distance between inner lip and stud (x4 the distance) and give it a squeeze.

 

[vonz33]

>Compositepro (Chemical) Did you even use any particular procedure for tightening the bolts? With this type of seal you must finger tighten all the bolts and then in small steps tighten bolts on opposite sides so that the plate compresses the gasket evenly.

As I said earlier, the bolts were done up with no particular procedure in mind, would you have something to suggest?

 

[TugboatEng]

If the flange is distorting causing the gasket to lose compression you need to make the assembly stiffer. You can either go gasketless and use an anaerobic sealant. Use a thinner gasket, expanded PTFE tape is really nice. Use a firmer gasket like Garlok 3000. Or, if you want to stick with your current materials and thickness, add some compression limiting spacers to the bolt holes in the gasket. These will be a metal washer that is just thinner than the uncompressed gasket. You will have to enlarge the holes in the gasket slightly for the last case.

 

[racookpe1978]

As I said earlier, the bolts were done up with no particular procedure in mind, would you have something to suggest?

Yes. Get a "millwright" to tighten them properly.

/sarcasm. A little.

1. It is wrong to TIG weld nuts to a stud to tighten them.. That assembly os called a "bolt" . Use it.
Use a proper gasket, as mentioned above.
2. Give us the bolt pattern you have. There (might be) a way to tighten them in some pattern that will evenly compress the gasket evenly by increments. Finger tight, 50% toque, 75% torque, 95% torque, then final torque. Ensure all bolt s and threaded surfaces are clean, lubricated, evenly tightened.

 

[vonz33]

TugboatEng

To go gasketless for a test unit prototype is not ideal to us as this means sealants that requires cleaning the parts or thin stuff that you have to cut over and over again every-time the unit is being pulled apart, and this happens a lot at this stage of R&D.

Now for the "stiffer" material, we have previously experimented with BS7531 Grade Y on disc flanges (16mm bolts vs 8mm on this design) for another unit and it was found that the pressure required to get that gasket to fully seal the assembly was too high for the tooling we have at hand. The gaskets BS7531 Grade Y failed the pressure test at around 100psi, so we tried with neoprene and we went past 300 easily.

So in this square shaped flange vessel situation, it would mean that the torque to apply on these poor 8mm bolts vs fibre type gasket will take us down perhaps to close to the no go zone, unless this Garlok 3000 is softer, we could give it a chance but this would mean getting them stamped to our specs and blow some more budget in risky business (we are at the end of financial year...)

I'm seriously considering a complete re engineering of the whole unit like I said in one of my earlier responses, as spending time trying to apply internal pressure to a cube is perhaps an unnecessary challenge.

Tomorrow I'm going to suss out some disc flanges for capping and draw the internal parts to match them, a cylinder shaped assembly will be a lot more friendly to pressure than cuboid... mechanics 101 or is it not... and also may not pass the compliance inspection (this cuboid design was an attempt to save time and funds by modifying a non pressure vessel and reinforce it to hold pressure, and we may have to scratch it)

If I am right, perhaps the quickest response to my question was: Don't waste your time fixing this unit, make it cylindrical.

 

[TugboatEng]

It can be square, there are many square plate type heat exchangers.in service. They just have very thick end plates.

Cleanup with anaerobic gasket makers like Loctite 515 only takes a few seconds with a razor blade. Dry time is in minutes after assembly.

Give the spacers a shot.

If you just need it to get through the prototype phase, consider getting some 5/8“ or thicker steel plate and just sandwiching the whole thing with it.

 

[Compositepro]

http://www.gasketresources.com/technical-info/gasket-installation

 

[kingnero]

Welding to the studs can, possibly, lower the yield strength enough (while hot) to lose all preload.
I don't see how welding to a bolted connection can add anything useful. I can see however several disadvantages, like altering the bolt and nut material characteristics, damaging any coating, ruining the gaskets, making it impossible to undo the connection, or to check torque/preload, ...
If it's done to prevent loosening over time, consult handbooks like Bickfords or others to gain knowledge on better alternatives (like using effective preload, which is not evident on gasketed joints).

 

[SnTMan]

vonz33, changing to a cylindrical design will not, of itself, prevent problems.

Regards,

Mike

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[vonz33]

TugboatEng (Marine/Ocean)17 May 17 13:52
It can be square, there are many square plate type heat exchangers.in service. They just have very thick end plates.

Thanks for your response

This vessel has very thick plates already, the issue is all the fiddly bits in between

Cleanup with anaerobic gasket makers like Loctite 515 only takes a few seconds with a razor blade. Dry time is in minutes after assembly.


A few seconds on one parts but when there are 50 of them and you have to clean both sides it is a real problem

Give the spacers a shot.

Spacers? Sorry I must have missed that one.

If you just need it to get through the prototype phase, consider getting some 5/8“ or thicker steel plate and just sandwiching the whole thing with it.

That's the first thing I tried the next day I quashed the assembly in the biggest vice in the shop and it did not really do that much.

I'm really tempted to go cylindrical, my gut is telling me that we have to go there for now, it really does make more sense to me that way. Started sourcing parts for it, no budget spent yet just some time. Will update this post once I have done some pressure tests.