Boiler Design – issue 4765

[MEinThailand]

On MEinThailand Said:

 

We will be able publish these conclusions when the model engineering community is a listening mode rather than a criticising mode. I don’t feel that we are that point yet, which is shame because our intention is always to help the model engineer with our research and do I think we have a lot of value to offer.

 

 

I would say that people have listened (read) and have offered their opinions based on what has been presented

Your Summary says

Further Observations About The Yield Point Method​
Due to the inherent characteristics of the YP formula the resulting boiler shell thickness will always be slightly above that calculated by the UTS method, therefore, arguably, a safer boiler would result.

So, if the benefits of using the YP method include avoiding permanent deformation during testing no matter how much of the boiler has been annealed during brazing, and a safer boiler, for only a very slight increase in boiler shell thickness, why not adopt it for future boiler designs?

You say it will be a safer boiler but so far there is nothing to really show (failures) that the current methods result in boiler that is not safe enough. As most model engineers seen to dislike spending money that would be one reason why not, as they will not want to spend out on thicker copper than has already been shown to be upto the job. If you were able to show failures and that a similar size boiler made to the YS method did not fail then they may convince people to adopt it or at least run the two sets of figures.

I’m sure the members would also be interested in past experience in this field not just the calculations but what boilers you have both previously made, were they to published designs/methods and did you find that they distorted and if so by unacceptable amounts.

[duncan webster 1Participant @duncanwebster1]

There has been a lot of hot air talked about this. To suggest that the design of a simple pressurised shell requires FE analysis is approaching the ridiculous. I believe that Messrs Smith and Brown has made 2 fundamental errors. Firstly, they have used the wrong value for proof stress, secondly they have not applied a safety factor for the 2* pressure test.

Modem practice in designing pressure vessels is to use proof stress as the criterion (BS5500, BS2790, both now superseded), older references in ME use UTS. In both the recommended safety factor on proof is 1.5, BUT then BS5500 requires a hydraulic test of only 1.25 above working pressure, whereas ME codes ask for *2. With a SF of 1.5 and a test pressure of *1.25, the stress at test pressure will be 1.2* proof, so if we adopt the same SF at *2 test we get a SF at working pressure of 2.4. I got a figure from the Copper Development Association (data sheet A6 – Cu-DHP) for annealed C106 at 24C of 9370 psi. I think the CDA should know what they are talking about. Applying this SF gives an acceptable stress at working pressure of 3904 psi

TDWalshaw (Tubal Cain,Model Engineer’s Handbook page 10.2) gives a method of arriving at acceptable stress at working pressure which allows for the fact strength drops off as temperature rises. This is based on data from the American Society of Tool and Manufacturing Engineers. Interpolating the data gives an acceptable stress at 100si steam pressure of 4193 psi

KNHarris, Martin Evans and Keith Wilson all give figures based on UTS as tabulated below. The Martin Evans figure is exactly the same as the KNH figure, so may have been copied

The Australian code gives 3771 psi

Jason’s post of 25 May gives 4000 psi

Perhaps it makes it easier to get one’s head round if these figures are tabulated

 

Source                                                                    Acceptable stress psi

DAW (derived from British Standards & CDA)                    3904

TDWalshaw                                                                    4193
KNHarris                                                                  2500 to 4166
Martin Evans                                                            2500 to 4166
Keith Wilson                                                                  3500
Australia                                                                       3771
Smith and Brown                                                           2415
JasonB                                                                          4000

 

I’m not sure of the meaning having a range in the Evans/Harris figure. According to TDWalshaw and others, a tube made by rolling plate and joining with a but strap would require an extra safety factor, but let’s not go there yet.

On the basis of the above, any suggestion that current practice is inadequate seems to be incorrect. I am not aware of a properly constructed and tested silver soldered boiler ever giving way by bursting the cylindrical shell, being a pragmatic sort of chap I would say if a boiler passes a 2* hydraulic it is safe. The only purpose of a design code is to give boiler inspectors an agreed standard against which to judge novel designs (as required by the SFED code), and to help avoid non standard designs failing the test, which could be expensive. Some minor permanent distortion, eg quilting around stays, wouldn’t cause me to lose any sleep, and no I’m not going to get into a discussion on what is minor.

My CDA booklet gives a figure for the Fatigue Endurance Limit of ±10800 psi, and we are not going to repeatedly perform the 2* test, so that need trouble us no further.

Let me re-state what I’ve said before, if something is strong enough, making it stronger doesn’t make it better, just more expensive. In the interests of international harmony I’m in favour of adopting the Aus code.

I have performed similar comparisons on stay spacing/sizing, and there is similarly little to choose between any of the standards, again FE analysis would be overkill, it is likely that small areas would be shown to be above proof stress, but they would just yield and redistribute the stress, so unless the program allowed for this (some do), undue alarm would be caused (been there before)

 

[JasonBModerator @jasonb]

Well all this is better than the usual tea room fare!

On a slightly different note and knowing that all this stemmed from queries about the boiler for a traction engine to one of Julus de Wall’s designs (not Julian) I thought I would have a quick look at a few of the traction engine drawings that I have.

Now the UK code says that a new boiler can be based on calculations or based on a similar existing design to the discretion of the inspector. Taking the latter I would say that the suggested new YP method would NOT result in a safer boiler. May be cheaper though but is the cost worth the risk?

Since Compiling the table below I see Duncan has posted showing a near 50% difference in previously accepted and the new methods Acceptable psi. This would seem to be born out by the figures that consistantly show the YP method gives a thinner result than all the established designs.

[duncan webster 1]

On duncan webster 1 Said:

To suggest that the design of a simple pressurised shell requires FE analysis is approaching the ridiculous….

I agree!  But designing a simple pressurised shell is not why I suggested FEA!  My point is the formula are good enough for practical purposes, but not good enough to design a new generation of lighter, stronger, boilers.  For that consider FEA.

As always the answer depends on the goal.

• To get past a boiler inspector, do whatever makes him happy, probably building a historically proven design to one of the codes.  Gets the job done, but not innovative, tends to demand lots of Copper and mitigates against steel, which could be cheaper and safer.
• If the need is a better boiler or changing the codes, then it’s necessary to prove the new approach is sound.  I suggest FEA is the easiest way to provide the proof, and not ridiculous for that purpose.

Dave

[duncan webster 1Participant @duncanwebster1]

Slapping a cylinder on top of the boiler and using the boiler as part of the engine structure adds extra loading. I’d have to think quite hard about what effect this has on the stresses as it would not add to the hoop stress, just the longitudinal, both a direct end to end tension and a hogging/sagging moment. That’s probably why the Aus code asks for thicker shell. Adding a winding drum adds yet another complication. It could all be worked out on paper, but when therd is a simple code why bother. Ill try to work out what the extra FS in the Aus code is, it would be easier if someone who was involved in its preparation told us.
I very much doubt whether FEA would result in a significantly cheaper boiler, you might get away with slightly fewer stays, but it wouldn’t change the shell thicknesses, and you wouldn’t use thinner plates on the flat surfaces. TIG welded copper would be cheaper as you dont have to make flanged plates and I guess TIG welding is less labour intensive than silver soldering, as well as saving the cost of the solder, which ain’t cheap. However I foresee all sorts of issues with amateur non coded welders getting past amateur boiler testers who are not familiar. Commercially produced TIG welded should come with all the necessary documentation

[MEinThailandParticipant @meinthailand]

JasonB many thanks for explaining your calculation. With regards to whether the inside or outside diameter of the boiler is used in the calculations, it makes no practical difference at all. Sure, the calculated value will be different but there are other factors that render this immaterial.

The first factor is what is the nearest thickness of boiler shell material available? Whether the calculated desired thickness is 0.112 or 0.122 is irrelevant if the closest boiler material is 0.125″ = 1/8″.

There are other factors also that I have alluded to in my subsequent study of the subject that make such minute differences in the calculated shell thicknesses irrelevant.

It seems that by focussing in the minutia of the calculation results the big picture has been lost.

With regards to flat plates and stays I have done a lost of research there too and with the aid if AI created two calculators to determine flat plate thicknesses and stay distances together with all the associated stresses, dimensions and deflections. This is an extremely complex subject and one which I am not ready to enter discussion on at the moment at a time when the members are unable or unwilling even to appreciate or to understand the extremely simple Yield Point Method.

[MEinThailandParticipant @meinthailand]

JasonB I think you are missing several points here and your table is deliberately misleading your readers.

1) Usage Factors

The YPM calculates the theoretical minimum thickness of the boiler shell for an assumed value of the Yield Point of annealed copper the value of which (as stated in our article) varies depending upon the source of information.

From this calculated thickness the designer then needs to apply factors as stated in our article that need to be taken into account.*

* “and also allows the designer to add an appropriate safety factor based on the environmental situation the boiler will be subject to during its operation.”

For example, the boiler designer may decide to add a safety factor if the boiler shell has to take external stresses such as a traction engine or more extremely a ploughing engine.

The AMBSC Code also makes a distinction between boiler shell thicknesses depending upon whether there are cyclic (external machinery) stresses or not. This is quite correct which is why we mentioned it in our YPM article.

2) Actual Material Thicknesses

Based upon a calculated boiler shell thickness and after applying other factors as mentioned above, the boiler designer then choses the actual boiler thickness to used based upon what material is actually available. For example if the calculated thickness (plus any other factors) is 0.112″ then a 1/8″ (0.125″) material may be selected.

3) Calculated Versus Specified Thickness

You are not comparing like-for-like in your table. You are comparing the minimum calculated YPM value with a value shown on a drawing. Who knows what the calculated value of the thicknesses given in your table are? Who knows what other factors have been allowed for and by how much above the calculated values for thicknesses the tabulated thicknesses are by virtue of the nearest actual material size being selected?

Please provide these details if you want a fair comparison versus the YPM calculated values.

4) Published Designs are Not Proof of Correct Design

Just because particular design of boiler has been published is absolutely no proof at all that the deign is correct. It is fallacious to say that because a design has been published and boilers built to such a design have not failed (or deformed under hydrostatic pressure testing) then the design is correct.

Even worse, and quoting your statement “a new boiler can be based on calculations or based on a similar existing design to the discretion of the inspector” is completely wrong by any measure of proper engineering practice.

We covered this in our YPM article:-

“If proper engineering procedures and practices are to be followed it is unacceptable to say that because there have been no failures, then the formulae and assumptions used during the design process must be correct.”

5) Your Table Is Misleading (Again)

As stated above, You are comparing the minimum calculated YPM value with a value shown on a drawing. This is misleading.

Also, what you are not showing is the YPM value compared to the UTS value. UTS calculation being the traditional method as used by Greenly, Harris and Evans.

Taking two examples here are the results:-

Minnie

YPM 0.026″ (As your table)

UTS 0.02″ (Slightly less than YPM)

Drawing 0.064″ (Much greater than either of YPM or UTS methods)

Comment. The theoretical thickness required (both YPM and UTS methods) is less than 1/32″. This is a very fragile thickness and I’m not even sure if it is available in the diameter required. The designer, LC Mason, has sensibly chosen a realistic thickness of 1/16″ which is shown in your table.

Allchin

YPM 0.078″ (As your table)

UTS 0.06″ (Slightly less than YPM)

Drawing 0.092″ (Much greater than either of YPM or UTS methods)

Comment. As for Minnie, the designer, WJ Hughes, appears to have recognised that the calculated thickness of 0.06″ (1/16″) is rather fragile for a 1 1/1/2″ scale traction engine and has chosen a sensible thickness of 3/32″)

All this bears out my comments above of calculated value versus actual practical chosen value for boiler shell thickness.

For the doubters, here is a screenshot of my calculations:-

 

[MEinThailandParticipant @meinthailand]

I refer you to the Post by Paul Lousick

Deformation of the boiler when undergoing testing is a sign of weakness in the design or manufacture of the boiler and is one of the first things that is checked when doing a hydraulic test and is one of the requirements in our design code for model boilers in Australia.

[MEinThailandParticipant @meinthailand]

Reply to duncan webster 1

duncan, you said:-

“There has been a lot of hot air talked about this. To suggest that the design of a simple pressurised shell requires FE analysis is approaching the ridiculous. I believe that Messrs Smith and Brown has made 2 fundamental errors. Firstly, they have used the wrong value for proof stress, secondly they have not applied a safety factor for the 2* pressure test.”

I agree with your first statement “To suggest that the design of a simple pressurised shell requires FE analysis is approaching the ridiculous.”

FEA produces an infinitely more accurate set of figures than a simple spreadsheet can. But the meaning of the figures output are dependent upon the assumptions made particularly the values of strength  of copper which as I have said many times vary widely according to the source of the information.

Garbage in = garbage out.

Your second sentence “I believe that Messrs Smith and Brown has made 2 fundamental errors. Firstly, they have used the wrong value for proof stress, secondly they have not applied a safety factor for the 2* pressure test.” is completely false and misleading. Go read the article again and point out to me the paragraph and sentence to which you are referring.

[MEinThailandParticipant @meinthailand]

JasonB your comments are again misleading and slanderous.

I have pointed out repeatedly that the thickness as calculated by the YPM have to be modified by the designer to suit the conditions of operation and the materials available. Whether the ‘traditional and accepted’ UTS method is used or the more accurate YPM method is used the same applies.

Also the outputs from the two methods depend greatly upon the physical properties of copper and vary widely according to the source of information.

As to cost this is a non issue. The boiler has to meet the safety requirements regardless of cost.

[MEinThailandParticipant @meinthailand]

Luker, just in case readers are not familiar with your many articles in Model Engineer (now MEW) magazine, can I say that you are a prolific contributor and ardent supporter of new methods and new trends in engineering that can be and are being increasingly adopted by progressive model engineers?

I am therefore disappointed and perplexed why you should chose here to apparently slander the Yield Point Method article when by your own admission you have not even read it!

To quote your post:-

“Unfortunately I have not read the article in question”

“an article that is fundamentally flawed will dissuade builders making their boilers.”

I acknowledge that you are not directly saying that the YPM article is “fundamentally flawed” but it could be inferred by the average reader that that is what you meant.

I would be grateful if you could clarify whether you think that the YPM article is “fundamentally flawed” or not and give your reasons.

I have permission from the Moderator JasonB to give you access to the Yield Point Method article as posted on the ModelEngineeringInThailand.com forum and look forward to your positive comments when you have read and digested the article.

[JasonBModerator @jasonb]

The first factor is what is the nearest thickness of boiler shell material available? Whether the calculated desired thickness is 0.112 or 0.122 is irrelevant if the closest boiler material is 0.125″ = 1/8″.

Copper sheet has gone over to metric in these thicknesses so 1/8″ is hard to come by therefore 3mm would actually be the nearest readily available. It is above the lower value but below the higher at 0.118″ so it could actually have a bearing on what is used. If tube were available then 10swg at 0.126″ would cover both

Also if you are going to be showing that existing methods are not suitable it would give your new proposed method more credability if calculations were done using the correct numbers.

 

JasonB I think you are missing several points here and your table is deliberately misleading your readers.

1) Usage Factors

The YPM calculates the theoretical minimum thickness of the boiler shell for an assumed value of the Yield Point of annealed copper the value of which (as stated in our article) varies depending upon the source of information.

Saying “Leave it to the designer to account for such factors”  could result in them not making sufficient allowance. My point was that using the existing code which allows boilers to be based on existing proven designs would likely be safer than one where someone designing their own may not allow enough. And would therefore not result in a safer boiler that you claim your method will produce.

I have also effectively said a couple of times in this thread that I feel a ploughing engine should have a higher “usage factor” applied than a traction engine let alone a loco so certainly not trying to mislead people about the need to consider the use the boiler will be put to as the other well established disigners did in the past to arrive at the thicknesses shown on the table.

Even worse, and quoting your statement “a new boiler can be based on calculations or based on a similar existing design to the discretion of the inspector” is completely wrong by any measure of proper engineering practice.

I was simply stating the current UK Code, I was not saying if it was wrong or not but it seems to work as we don’t see many failures. And the boiler inspector may not accept all existing designs if he feels they are not well proven based on there experience.

Oh Bill Huges did not design Minnie

 

 

[JasonBModerator @jasonb]

So as not to mislead anyone I have quickly added the thickness in thou that the Evans formula gives as well as the difference between the two.

Seems very little and as use factor take sthe shell well over the minimum is there much point in changing from what has been proven to work in the past?

[JasonBModerator @jasonb]

Now what could be misleading is using a different temperature allowance here (1) to th eone you used in the article (0.8) giving a 20% difference rather than a 4%. Was there an engineering reason for this? As the Model boiler testing setup is very much based around Locos then the Evans method is more likely to be what will be used and quite possible preferred by inspectors and that is why I used it above.

 

[MEinThailandParticipant @meinthailand]

JasonB, I’m responding to you post and quote below the points you raised.

“it would give your new proposed method more credability if calculations were done using the correct numbers.”

Our calculations DO use ‘the correct numbers’ with the caveat that no one can say what is the correct number for the Yield Point of copper, nor the UTS number for that matter.

I assume, therefore that you accept that the numbers used in our calculations are correct but that you are referring to number not used in our calculations, in this case the thickness of available copper materials.

Let me make one point very clear.

The Yield Point Method only goes as far as calculating the minimum safe theoretical thickness for copper boiler shells.

This is exactly the same situation for the ‘well established’ UTS method.

So whatever you say about the thickness of available materials and additional factors such as external forces (eg ploughing engine drums etc) applies equally to both calculation methods.

If you say that the YPM method should not be used because “Saying ‘Leave it to the designer to account for such factors’ could result in them not making sufficient allowance”, then equally the UTS method should also not be used!

There is no logic to your statement.

Next, your “I have also effectively said a couple of times in this thread that I feel a ploughing engine should have a higher “usage factor” applied than a traction engine” I quite agree with you which is why our article says that boiler designers should take such factors into account. I’m glad that we both agree on that point.

As to “Oh Bill Huges did not design Minnie” this is a snide and yet another misleading remark. I quite clearly did not say that Bill Hughes (correct spelling) designed the Minnie.

What I said was in relation to the Allchin engine (by Bill Hughes) “As for Minnie, the designer, WJ Hughes, appears to have recognised that the calculated thickness of 0.06″ (1/16″) is rather fragile for a 1 1/1/2″ scale traction engine and has chosen a sensible thickness of 3/32″)”.

If you can’t understand the context in English I’ll spell it out for you:-

As for the case of the Minnie design, the designer, of the Allchin, WJ Hughes, appears to have recognised that the calculated thickness of 0.06″ (1/16″) is rather fragile for a 1 1/1/2″ scale traction engine and has chosen a sensible thickness of 3/32″)

Above all else let me reiterate that the YPM is a proposed substitution for the old UTS calculation method. Both stop short at the point of establishing a calculated minimum thickness of boiler shells for copper boilers. Anything beyond that point is not a part of the YPM proposal.

[Bob WorsleyParticipant @bobworsley31976]

I like the Haining designs, simple, and get a working model at the end, just like LBSC. That they are not detailed enough now is another problem. I used to hate the ME show results when the great and good criticised a model for some trivial variation to the full size, after the guy spent hundreds of hours making it. For that reason I woud never exhibit a model of mine.

I find I have no interest in actually driving or using a model, loco or engine, what I like is the challenge of “how do I make that part”. I like 2.1/2″ locos, light enough to lift, small enough to now want the fiddy detail, but look good.

[JasonB]

On JasonB Said:

Well all this is better than the usual tea room fare!

[…]

Let’s just hope that the debate can be kept civil

MichaelG.

[JasonBModerator @jasonb]

Yes I’m sorry that I’m a bit dyslexic and make the odd typo and my english and grammar may not be upto that of others

But to spell it out, wrong numbers = using the ID when the calculation should use OD. I am not doubting the formula. The incorrect 6″ was used when the correct figure should have been 6″ + (2 x T)

If you can explain where I have got that wrong I would be most greatful.

As you say the result is based on what you pick the copper figures and safety factors from, as my table comparing Evans and yours shows the differences can be bery small, if a designer were to pick a slightly different figure it may be the results are the same or the old method could even give a thicker boiler even before other factors are added.

[Bob WorsleyParticipant @bobworsley31976]

Apologies for my last post, I missed that there were two more pages of comment.

Yield point of copper. What seems to have escaped discussion is the difference between the yield point for annealed and hard copper, as in the article graphs. This is the same material, copper, and both are valid. If, during testing or use the annealed copper is stressed then it work hardens, and the yield point graph moves from the annealed to the hardened one. This happens automatically and safely. The safety factor used takes this into account.

That no one can say what the yield point stress is for copper seems totally wrong. You can test copper as many times as you want, hard or annealed, and will get similar results. Why? Because if you didn’t then there is something wrong with the processing of the copper, or other material. Repeatability is what is assumed. The copper used in boilers is the C106, this is as near pure as possible for both heat conduction and electrical conduction. The ultimate test seems to use hard copper, the yield test use annealed copper.

If interested Journal of the Franklin Institute, reprinted in Mechanics Magazine, did a whole series of boiler tests in the mid-1800s. There has also been quite a few articles in ME over the decades about similar. That copper thickness in boilers has evolved is also true, the 1906 undertype, pre-Bolton, used a 15swg by 5” boiler, updated to 10swg. Likewise my BB boiler.

Use of FEA does seem to be over the top, just make it thicker. As an example boiler stays, there was some testing done by Martin Evans in the middle of the Princess of Wales series, 3409 11th Jan 1971, that resulted in just one comment. The firebox will bulge between each stay, being flat, but how much? Can also calculate using the theory of continuous girders, doesn’t matter if the load is in going out or downwards like an industrial building.

Reading snippets in many articles and books and it seems very true that copper is a good material to make boilers from. Whilst very flexible on the yield point, the work hardening is very important. But the choice of copper is also important, it takes very little alloying, 1%, to completely change the electrical conductivity, and possibly also strength. There have also been many articles about embrittlement of stainless steel, both from welding and firebox fumes, likewise the craze to use the phosphors based silver solder. And as for using steel in a hobby boiler, umm, rust?

[Bob WorsleyParticipant @bobworsley31976]

Wrong ME for boiler stays article, 3406 4th Dec 1970.

 

[JasonB]

On JasonB Said:

 

To borrow Jason’s list and part of the document in question below .

The UTS Method​
Over the years the thickness of the boiler shell has been determined by multiplying the Ultimate Tensile Strength (UTS) of cold worked (or cold drawn) copper by the Working Pressure (WP) then by an arbitrary large safety factor (SF), typically 8. A Temperature Factor is also sometimes applied. This will be called “The UTS Method”.

The fact is that when copper is heated during the brazing (silver soldering) process the strength of the material is dramatically reduced in the heated zone because of the annealing effect. See Fig 1.

The original UTS figure used to determine the boiler thickness therefore becomes invalid. The yield strength of annealed copper can be as much as one sixth of the original cold rolled copper material.

The calculations below illustrate that boilers having wall thicknesses sized using the UTS method are susceptible to permanent deformation during pressure testing.

However, many model steam boilers have been constructed of copper and brazed (silver soldered) where the UTS method has been used to determine the shell thickness but there are few reported cases of boiler failure or deformation.

The reason that there appears to have been few reports of boiler failures or of permanent deformation may be because
a) in many cases the entire boiler shell is not heated to annealing temperature and the remaining material strength thereby remains largely unaffected
b) the amount of deformation is not noticed or is considered ‘acceptable’ by the tester.

If this is the reason, then it is a rather haphazard and unpredictable design method. In the case, for example, of a boiler barrel rolled from flat plate and having a lap joint the annealed zone could be extensive.

If proper engineering procedures and practices are to be followed it is unacceptable to say that because there have been no failures, then the formulae and assumptions used during the design process must be correct.

From Jason’s list it’s clear that no current boiler built to established plans is anywhere near the thickness’s that either calculation throws out at the end. Therefore no real boilers are affected by any of the dire warnings in bold above.

“but there are few reported cases of boiler failure or deformation”.  this implies there are some and a reference to them should be made.

To take the example of Minnie or a similar size loco boiler the ‘safe’ barrel would be too thin to attach anything to.

BTW when silver soldering boilers of this size the whole thing glows red so is fully annealed.

 

[JasonBModerator @jasonb]

Bob bring sup one of the point sI asked about in my inital reply. It is extreamly unlikely that a new boiler will go straight from pickle to a 2X WP test. At the very least any reasonable builder would want to check it for leaks starting with a low pressure and building up to higher maybe even 2xWP This will as Bob says start to harden the copper and change the graph. As the thicknesses between the YPM and that of Evans are so small then this additional hardening will likely reduce the diference even more to the consistant 3.5% thicker wall it gives

[JasonBModerator @jasonb]

As for this bit

“it would give your new proposed method more credability if calculations were done using the correct numbers.”

Our calculations DO use ‘the correct numbers’ with the caveat that no one can say what is the correct number for the Yield Point of copper, nor the UTS number for that matter.

I assume, therefore that you accept that the numbers used in our calculations are correct but that you are referring to number not used in our calculations, in this case the thickness of available copper materials.

If the numbers were correct then why did you alter them to my suggested ammendments and appologise for the mistake in this post ?

[duncan webster 1Participant @duncanwebster1]

I’ve tried to reverse engineer the Aus code table for traction engine boilers, and it looks as tho’ their acceptable max hoop stress is around 18 MPa, 2610 psi, but I wouldn’t put any money on that.

The UTS of annealed copper is the same as that of work hardened, as the annealed stuff yields plastically it becomes work hardened

The important factor is the INSIDE diameter, that is where the load is applied.

If an FE analysis of a thin shell under internal pressure gives a hoop stress significantly different to the simple theory, I’d go and check the FE because it’s wrong. The stress at the inner surface is very slightly different to the outside, this matters for thick shells (in relation to the diameter). Taking a 127mm OD tube, 2mm thick with an internal pressure of 620kPa (95 psi) the thin shell theory gives hoop stress 19.06 MPa, thick shell theory gives 19.4 MPa at the inner surface, 18.8 MPa outer. We don’t need to get worked up about such small differences.

I think it wise to assume that a silver soldered shell is annealed. According to my little book from the CDA, anything above 300C will do the trick. I’ve always gone to red heat. For a boiler, at least the ends will have been a lot hotter than 300C.

[MEinThailand]

On MEinThailand Said:

Luker, just in case readers are not familiar with your many articles in Model Engineer (now MEW) magazine, can I say that you are a prolific contributor and ardent supporter of new methods and new trends in engineering that can be and are being increasingly adopted by progressive model engineers?

I am therefore disappointed and perplexed why you should chose here to apparently slander the Yield Point Method article when by your own admission you have not even read it!

To quote your post:-

“Unfortunately I have not read the article in question”

“an article that is fundamentally flawed will dissuade builders making their boilers.”

I acknowledge that you are not directly saying that the YPM article is “fundamentally flawed” but it could be inferred by the average reader that that is what you meant.

I would be grateful if you could clarify whether you think that the YPM article is “fundamentally flawed” or not and give your reasons.

I have permission from the Moderator JasonB to give you access to the Yield Point Method article as posted on the ModelEngineeringInThailand.com forum and look forward to your positive comments when you have read and digested the article.

I apologise if my initial post was a little brash, and I should have sourced your article before commenting, so again I do apologise. Having said that; from the OP I had a very good idea of what methods were being employed and unfortunately I could see the potential for a ‘zero deflection boiler test’ to cause ‘confusion’ with already fickle boiler inspectors. I have subsequently read through your article, and Jason has forwarded links to another forum where I skimmed through some of the comments.

The yield point of a material has a defined strain, typically 0.2% but this varies depending on the standards used. Fundamentally; for a zero deflection model you should be using the elastic limit which is very difficult to measure accurately for copper (the YP and UT point scatter is related). The fact is, you will get some deflection when testing an annealed boiler either with the UTS method or a YP method. From practical experience, copper boilers are in a fully annealed state. Even the TIG welded copper boilers I’ve made could be dented by pushing on the shell with my finger (the required interpass temperature is enough to anneal the copper). The silver brazed boilers are worse.

Any new idea or innovation should improve the status quo.  My concern is, that, the boiler code writers would see this method and incorporate it into the code because it makes copper boilers ‘safer’ (not making boilers at all is even safer, but nobody wants that!). And just like that, numerous designs are void and many people cannot complete their builds that have been lingering under a bench, when in fact the supplied materials would have been fine. Then there’s the cost involved (which is significant) and the added heat input requirement (already tricky) for silver soldering etc. etc.

Practically you can test nearly all boilers that were designed (and properly made) using the UTS method at 2XWP, allow the deformation… have some coffee… release the pressure, then call the boiler inspector for the test. The chances of further deformation are small due to the work hardening of the copper.

On a related matter, but from the other forum, you cannot use a ductile failure model for calculating the stresses in a stainless boiler. The mode of failure is not that simple. If you have a look at my stainless boiler designs they seem ridiculously overdesigned, but this is intentional to limit the surface potential for SCC and CCC. I had to use FEM (the element type is critical for accurate results) for these boilers to make sure I had a safe design. You can use austenitic, but not 304, you need to source 316 or higher (low carbon). This limits delta ferrite and improves weldability.

Please don’t stop innovating and contributing!

[Author]

Posts