Hydrostatic Work Hardening Question

[Toymaker]

Question: Could Hydrostatic testing a copper boiler work harden the copper making the boiler permanently stronger?
This question occurred to me as I work through the process of Hydrostatic testing my monotube boiler.

Given that copper hardens and becomes stronger when it's hammered or stretched (drawn), giving it higher yield and tensile strength.
Than, if hydrostatic pressure inside a copper boiler, of any shape, exceeds the yield strength of the annealed copper boiler thereby allowing the copper to be stretched ever-so-slightly, would this not stretch-harden the copper? My guess is, Yes.

Also, since the annealing temperature of copper is 400 C (750 F) than as long as the boiler never reaches 400 C the copper would remain permanently hardened, and permanently stronger.

Any comments? Guesses? Knowledgeable experts that know the answer?

 

[Charles Lamont]

The boiler test code used by all UK model engineering societies states that during the first, 2 times working pressure, hydraulic test on a new copper boiler, the test pressure should be raised and relaxed in increments to work harden the soft copper.

 

[Steamchick]

Similary, my post #176 in your monotube boiler thread.
Well done on working this out unaided, Toymaker.
On annealing, consider the length of tube you use to make a coil. I assume you do not heat it uniformly in an oven to make your brazed joints at the ends, etc? Therefore you will have heated the parts of tube at the joints to annealing temperature. This hot zone will then form a temperature gradient to the cooler parts of the tube. Anything hot enough to anneal will be annealed, but cooler areas will not.
Therefore there will be a transition zone between annealed and harder copper. This in effect causes a stress-raiser, which is unknown, both in location and "effect" (value in mathematical terms). To avoid the highest pressures and stresses of the hydraulic test causing any stress cracking occurring at the transition zone, the mechanical work hardening of the copper by the cyclic progressive application of hydraulic pressure is applied.
But you had it all worked out anyway.....
Cheers!

 

[ninefinger]

My take on this subject of copper boiler work hardening during pressure testing. I rewrote this several times as I worked through the math and my references.

I will admit I had to rewrite my viewpoint a few times, as initially I jumped to the conclusion that the metal should not be yielding, that just seemed wrong given the large safety factors that at first appear to be applied, but read on.

The current AMBSC codes appear to apply a hefty safety factor for the stresses versus hardened copper, however AMBSC code part 1 issue 8 - 2012 uses max allowable stress in annealed copper (they specify annealed) of 26000 kPa (~3770psi), and that brings the expected stresses during a 2x Working Pressure test to quite possibly above the yield strength of annealed copper, so you could very well be work hardening the copper. Is this a "miss" on the part of the code?
Quite frankly I was disappointed that the code was not more upfront with the design points and safety margins as opposed to forcing you to reverse calculate the safety factors from tables and such.

Annealed pure copper tensile strength = 4830 psi
MatWeb - The Online Materials Information Resource

ASTM B370 copper yield strength in various tempers: 20000 - 35000 psi
Fundamentals: Types of Copper and Properties

Taking for example the boiler design by Martin Evans for the B1 Springbok I am building:
6" inner diameter, 1/8" thick.
@100psi the max hoop stress in the shell is ~2450 psi (ok by AMBSC code) but, at 2xWP that stress rises to 4900 psi - just exceeding the yield strength of annealed copper (per matweb link above, and exceeding by a good margin that value used by AMBSC). I haven't looked at the other parts of the boiler but I'm sure that similar values will pop up (i.e. exceed the yield of annealed copper).

This brings up the question, where should the design stress point be relative to the material condition and test pressures? Should we be designing around one of the "hard" conditions of copper with its higher yield strength, or the annealed copper yield strength? And the fact that a 2xWP test exceeds the annealed copper yield - should this be a concern or design consideration? Or is it just noise and doesn't matter - the boiler will work harden itself and settle in and we can all be happy that the material got stronger by our pressure testing.

 

[Toymaker]

My take on this subject of copper boiler work hardening during pressure testing. I rewrote this several times as I worked through the math and my references.

I will admit I had to rewrite my viewpoint a few times, as initially I jumped to the conclusion that the metal should not be yielding, that just seemed wrong given the large safety factors that at first appear to be applied, but read on.

The current AMBSC codes appear to apply a hefty safety factor for the stresses versus hardened copper, however AMBSC code part 1 issue 8 - 2012 uses max allowable stress in annealed copper (they specify annealed) of 26000 kPa (~3770psi), and that brings the expected stresses during a 2x Working Pressure test to quite possibly above the yield strength of annealed copper, so you could very well be work hardening the copper. Is this a "miss" on the part of the code?
Quite frankly I was disappointed that the code was not more upfront with the design points and safety margins as opposed to forcing you to reverse calculate the safety factors from tables and such.

Annealed pure copper tensile strength = 4830 psi
MatWeb - The Online Materials Information Resource

ASTM B370 copper yield strength in various tempers: 20000 - 35000 psi
Fundamentals: Types of Copper and Properties

Taking for example the boiler design by Martin Evans for the B1 Springbok I am building:
6" inner diameter, 1/8" thick.
@100psi the max hoop stress in the shell is ~2450 psi (ok by AMBSC code) but, at 2xWP that stress rises to 4900 psi - just exceeding the yield strength of annealed copper (per matweb link above, and exceeding by a good margin that value used by AMBSC). I haven't looked at the other parts of the boiler but I'm sure that similar values will pop up (i.e. exceed the yield of annealed copper).

This brings up the question, where should the design stress point be relative to the material condition and test pressures? Should we be designing around one of the "hard" conditions of copper with its higher yield strength, or the annealed copper yield strength? And the fact that a 2xWP test exceeds the annealed copper yield - should this be a concern or design consideration? Or is it just noise and doesn't matter - the boiler will work harden itself and settle in and we can all be happy that the material got stronger by our pressure testing.

I ran into similar problems trying to find the yield strength of the copper tube I had purchased here in Thailand, but that I know was manufactured in China. Of course, the hardware store I buy from has no idea what the ASTM # might be or even what manufacturing company made it.

Bottom line for me,...the only way I will know with any certainty, yield and ultimate tensile strengths, will be to test samples of the tube. I managed to get part way through testing a short sample section of tube when my brazed joint used to seal the tube sprung a pin-hole leak. I will continue hydrostatic testing after I fix the leak using a proper end cap instead of the sloppy crimping I first used. I've already determined my sample's annealed yield strength to be above 13,600 psi.

I'll post my final results, but I don't know how useful those numbers will be to this community as I cannot tell anyone which alloy my tube is made from.

 

[peterl95124]

This brings up the question, where should the design stress point be relative to the material condition and test pressures? Should we be designing around one of the "hard" conditions of copper with its higher yield strength, or the annealed copper yield strength? And the fact that a 2xWP test exceeds the annealed copper yield - should this be a concern or design consideration? Or is it just noise and doesn't matter - the boiler will work harden itself and settle in and we can all be happy that the material got stronger by our pressure testing.

you ask all the right questions, and I wish I had answers but don't, all I can say is that IIRC not only does copper work harden, it can also become brittle, so another reasonable question is how do you know if the copper will work harden or become brittle during pressure testing.

 

[Steamchick]

The UK test code (Federation of Model Engineering Societies) uses progressively higher pressure returning to zero between each, so 6 to 10 cycles is supposed to harenew annealed zones to be adequate for the final - Full - hydraulic test pressure. Because "those boilers tested have not failed yet" the Model Engineering bodies think it is adequate...
I understand the principle is that while NOT ensuriNG a uniform or known state of hardness for all the boiler components, at least the highest stressed zones that were annealed become partially work hardened, reducing the "yield variation" to unknown areas of unknown hardness....
All a big fudge! IMHO...
K2

 

[Charles Lamont]

K2, may I politely suggest you have a tendency to overthink.

The UK test code was discussed and agreed with the HSE and the usual UK model boiler insurer and is deemed to satisfy H&SAW act and PSSR.

Generally, localised plastic deformation may occur a very long way from failure. You get it every time you tighten a new bolt and nut.

 

[Charles Lamont]

deleted

 

[Steamchick]

Hi Charles, I accept that and you are right. I DO over think problems. Always have and criticised at work for doing so.
I'love shut up on the matter.
My only problem is that I have have design a few boilers over the last 20 years, and compared to calculations used back in the 1980s on pressure vessels the boiler design calcs are sparsely covered in most (possibly all?) the text books etc that I have seen, and used, in the last 5 years.
I have basically done calculations, of which I am confident, that have de-rated mine and others' boilers, not because they are considered "unsafe" by the regs, but that an hydraulic test of 2 x NWP has not proved the boiler capable of a factor of safety of 8. In one case, a 20 psi boiler collapsed a flue tube at 1 20psi (6 x NWP), another failed hydraulic test at 1.9 x NWP. AND, people thought it was "good enough" because it was so close to 2 x NWP....
Hence, in the absence of an expert to stop me, I have continued to blunder on in the darkness of boiler calculations....
Thanks for your advice,
I need it.
K2

 

[Charles Lamont]

K2, I don't have a copy of the AMBSC codes so I am somewhat in the dark too, but I do have the old books by KN Harris and Martin Evans that you have misgivings about.

I would suggest that the design safety factor of 8 is intended largely as a factor of ignorance. It is intended to allow relatively simple calculations to be performed by people without engineering degrees. For example, stress concentrations don't need to be analysed or looked up in Roark's, because (a) the safety factor allows for them (and (b) the ductility of annealed copper means they are largely self compensating).

Given that, I am not at all surprised or worried that a flue, designed with simple theory with a safety factor of 8, collapses in practice if you are daft enough to take a hydraulic test to six times working pressure. I think I would be impressed if it didn't.

And if it didn't fail, I think I would regard it as scrap, having been probably overstressed and subject to unknown internal stresses upon release of pressure in the now excessively work-hardened material.

OTOH, if it fails a test at 2 x working pressure then there is definitely something wrong with the design, the material, or the construction.

 

[Toymaker]

Has anyone used or considered using one of the Copper Beryllium alloys? Information on the web claims 6 times the ultimate tensile strength of pure copper at room temp.

Properties	Metric	Imperial
Hardness, Rockwell B​	80.0 - 85.0	80.0 - 85.0
Tensile strength, ultimate​	1280 - 1480 MPa	186000 - 215000 psi
Tensile strength, yield​	965 - 1205 MPa	140000 - 174800 psi

From other charts on the web the Yield at 500 F is 140,000 psi which is roughly 10 times higher (or more) than garden variety copper. A very impressive copper alloy,...wonder what the cost is?

 

[Steamchick]

Ok, I accept I was daft, as there was probably some degree of imperfection in the circularity of the tube.... but it should have been OK for a factor of safety of 8 according to the eminent designers you quoted, who's texts I had used.
Reading ASME, of which I first used Koto Hiraoka's magazine article as a guide, I learned that copper deteriorates in strength with temperature. ASME therefore take a Factor of safety, added to which they derating the allowable stress for temperature, then, so I was advised by one professional, they use a "global" stress concentration factor of 3.3 - despite there being lower determined text book factors - where there are any stress concentrations in the pressure vessel.
All be it these added factors sound crippling to models built to UK MODEL standards, they have been added by modern eminent engineers as a minimum requirement for the USA.
DO I KNOW BETTER? I Don't.
Therefore, and this also ties in with my pressure vessel design work in my job during the early 1980s, I follow ASME thinking.
Call me daft, over-cautious, or whatever, I CANNOT honestly say, "less safe is a good thing".
My friend's brother nearly lost an arm from a steam scald, so I have seen evidence of an "escaping steam" accident.
Someone I discussed this with, from the Southern Federation, suggested that I was doing more than they had decided was adequate, but didn't tell me what NOT to do in my calculations.... as it was "outside their agreed remit". He suggested my designs would "probably be safe", but not appropriate for someone "who has a boiler that has been in service for many years and it hasn't failed yet".... that was built to a known design from before the standards were written...
And I accept that "the pyramids have not fallen down yet".
But what permitted stress level, Fact of Safety, Stress concentration factor, etc. Should I be using for boiler designs in the UK?
K2

 

[mrehmus]

Has anyone used or considered using one of the Copper Beryllium alloys? Information on the web claims 6 times the ultimate tensile strength of pure copper at room temp.

Properties	Metric	Imperial
Hardness, Rockwell B​	80.0 - 85.0	80.0 - 85.0
Tensile strength, ultimate​	1280 - 1480 MPa	186000 - 215000 psi
Tensile strength, yield​	965 - 1205 MPa	140000 - 174800 psi

From other charts on the web the Yield at 500 F is 140,000 psi which is roughly 10 times higher (or more) than garden variety copper. A very impressive copper alloy,...wonder what the cost is?

The problem with Beryllium copper is the great danger of getting even the tiniest bit under your skin. Makes one very ill and until it is removed, the illness remains. Sometimes the victim has to be examined with magnifying tools everywhere to find it. I'd leave it out of the equation were I you.

 

[Toymaker]

The problem with Beryllium copper is the great danger of getting even the tiniest bit under your skin. Makes one very ill and until it is removed, the illness remains. Sometimes the victim has to be examined with magnifying tools everywhere to find it. I'd leave it out of the equation were I you.

Are you sure you mean Beryllium-Copper and not pure Beryllium? I know that Pure Beryllium metal is highly toxic, but various Heath and Safety Data Sheets say it's safe to handle, just don't breath any dust, as getting it into your lungs can cause real trouble.

Your post has me believing you have first-hand experience,...can you tell us more, please.

 

[mrehmus]

I have no personal encounter with the metal but my shop teacher who was good enough that his work was left on the moon showed the class documentation of what happens when beryllium in any form gets under your skin. I stay away from metals with that in the alloy. He also showed us the shop rules for anyone who had to work with those alloys. Not something to take lightly.

 

[Toymaker]

Final destructive hydrostatic test results performed on a short sample of the copper tube I'm using in my boiler are now complete and posted: Here

 

[Charles Lamont]

Another alloy I have vaguely thought might be useful for flash boilers is Cunifer tube as used in car brake lines, but I have never bothered to look into it.

In UK practice, full size steam locomotive inner fireboxes are made of C107 arsenical copper, which has higher strength at elevated temperatures.

K2, I am not sure, but I think the AMBSC codes are accepted in the UK. I would like to see a copy, but don't know were to get one. Until I have I don't think I should pontificate further.

 

[Steamchick]

Thanks Charles. I'll do a search on AMBSC codes later, as I'm too busy right now...
But I do worry a bit when someone thinks their boiler is "Good" when the hydrostatic test has passed, but calculations suggest they have a LOT less than a Factor Of Safety of 8.... (FOS of ~2.6 on the NWP was one I calculated after the event!! ... Draw you own conclusion? It was one I had used... Because the Boiler Inspector said "I don't need to examine it, just Hydrostatically test it. - If it is OK I shall certify it, so you do not need to remove the cleading for examination". - Later, when I did remove the cleading (very difficult), for a different inspector to examine for updating the certification, we discovered the boiler ends were domed from earlier hydrostatic tests.... hence the calculations. The boiler is now out of service awaiting re-design and rebuild with the fitting of more stays... or SCRAPPING.).
K2

 

[Steamchick]

Hi Charles,
Just re-reading your advice, and seeing this point: "I am not at all surprised or worried that a flue, designed with simple theory with a safety factor of 8, collapses in practice if you are daft enough to take a hydraulic test to six times working pressure."
Based on my industrial experience, we designed the pressure vessels using the recommended (text-book and regulation) calculations, then tested the first manufactured version to prove the design - "to the pressure used to create the stress proposed for the factor of safety to be determined" - which by design was below the UTS of materials. When the object did not burst, we understood the design to be good, so the NWP was then approved at that test pressure less the factor of safety. The equipment was to contain highly toxic gas, so had to be absolutely gas tight under pressure, (40 year service life), and any failure could result in catastrophic problems due to the volume and pressure contained, and possible energy release from internal combustion should a fault occur. Materials included Aluminium welded chambers, Cast aluminium chambers, steel chambers (welded and castings), ceramic chambers, bolted neoprene joints, and regular stainless steel pipework and fittings. - But I am not an expert on boilers... The equipment was tested by component/sub-assembly at 8 x NWP, to prove the design, but only at 2 x NWP as a full assembly before assemblies were shipped. Hence I was stupid enough to expect the "Text book tubes of adequate size" should not collapse when tested at >6 x NWP" in my boiler... I therefore no longer use those "recommended sizes for flue tubes". - Frankly, those standards are obsolete IMHO.
Better to test and not-use if it fails, than to use "blindly" without testing... as I do not have the expertise to judge otherwise.
K2