Boiler Design – issue 4765

[Andy StopfordParticipant @andystopford50521]

The Haining designs are useful as an approximate reference, but you can get copies of the original works drawings from The Museum of English Rural Life and rescale and adapt them to your own requirements. They’re not cheap if you need a load of them, but better value to my mind then buying the Haining ones.

I’m working on a 1/12 scale Z7S, a scale which is somewhat…challenging for such a complicated machine, but its as big as my lathe and indoor workshop allow, and getting the MERL drawings is much better than re-scaling something that’s already been scaled and changed to suit the facilities of the 1980s.

At this moment my 3d printer is making a mock-up of the coiling gear – I was quite unable to translate Haining’s design into anything workable and I don’t have a works drawing for that (there isn’t a full set for the Z7S) – 21st century technology to the rescue here for something that would be very tiresome to iterate multiple test pieces using old-school methods.

I’ll probably use the printer to make patterns for some cast bits too(not strictly necessary to use castings for a model of this size, but I have a little propane fired furnace, so why not give it a go – my limited experiments so far have come out far better than some of the commercial Minnie ones I foolishly bought).

[Bob Worsley]

On Bob Worsley Said:

Just dug out the emails when boiler kit bought, and contacted Neil Tyler about the changes and never got a reply. Reeves saying it is up to the designer, they just supply what is asked for, Neil Tyler says it is due to the supplier of the kit. So I ended up with £670 worth of scrap copper.

 

If you buy anything mail order you can return it for a full refund no questions asked. Thats UK law. There might be a time limit, not sure, ask citizens advice

[JasonBModerator @jasonb]

I have made contact with Alan Brown, one of the authors and invited him to comment. He is a forum member.

[duncan webster 1]

On duncan webster 1 Said:

On Bob Worsley Said:

Just dug out the emails when boiler kit bought, and contacted Neil Tyler about the changes and never got a reply. Reeves saying it is up to the designer, they just supply what is asked for, Neil Tyler says it is due to the supplier of the kit. So I ended up with £670 worth of scrap copper.

 

If you buy anything mail order you can return it for a full refund no questions asked. Thats UK law. There might be a time limit, not sure, ask citizens advice

That limit is normally 14 working days after delivery, so this case seems time-expired.

[duncan webster 1Participant @duncanwebster1]

What diameter and pressure is the BB boiler? The Aus code allows 90 psi for a 5″ diameter 2.5 thick shell. Sorry for the mixed units, peopld of my antiquity had to cope with both

[JasonBModerator @jasonb]

5″ OD x 75psi.

Though I would want to make an allowance for the fact it is a ploughing engine so quite long and unsupported plus that big winding drum slung below will put loads into the barrel if you pull anything. There are no stiffening pads for the drum bracket, though there are for the cylinder and valve bracket

I assume that the original 13swg design was found to be a bit wanting and that is why Haining upped it to 10swg (0.126″)

[Andy StopfordParticipant @andystopford50521]

Doesn’t Haining also suggest using a steel shell round the boiler to take the loads from the drum and front axle?

[JasonBModerator @jasonb]

He does give it as an option in the article but the photos in the article are of a boiler with the brackets soldered direct to the barrel and that is also what is shown on the drawings. Even if you did make it I don’t think you could slide it past the clack valve bushes or fit the smokebox which is still shown to fit the end of the boiler not the steel tube so boiler taking the axle loads

[duncan webster 1Participant @duncanwebster1]

There is a separate table for traction engines where the boiler is subject to external loads, thats the figure I quoted. I wouldn’t have thought you’d be doing a lot of ploughing with a 2″ scale model

[JasonBModerator @jasonb]

Yes but with a ploughing engine typically having a boiler barrel 50% longer than a traction engine I’d want to see some additional thickness even over the thicker TE table.

You have not see Harry Williams pulling loads accross a field with his pair of Kitshons or Ross Bishop running his 3″ under load. generally the models are run with a load rather than an actual plough as rally organizers are not keen to have the grass torn up and it is also hard to find someone smallenough to sit on a 2″ scale balance plough to steer it

https://youtu.be/hAcpujv-_4g?si=t8j70PygSsoB_OD2

[JasonB]

On JasonB Said:

Thank you JasonB for pointing out an error under Example 2, a boiler by KN Harris. The figure 3 in the YPM formula should indeed be 2.

This results in a calculated YPM thickness of 0.112″ (Not 0.122″ as you stated) resulting in the Harris thickness being 77% of YPM not 55% as in the Article.

I do apologise to MEW and our readers for my mistake.

Please note that this in no way detracts from the philosophy nor the conclusions of the article.

A revised section of the calculation referred to is shown below.

 

[MEinThailandParticipant @meinthailand]

Thank you JasonB for pointing out an error under Example 2, a boiler by KN Harris. The figure 3 in the YPM formula should indeed be 2.

This results in a calculated YPM thickness of 0.112″ (Not 0.122″ as you stated) resulting in the Harris thickness being 77% of YPM not 55% as in the Article.

I do apologise to MEW and our readers for my mistake.

Please note that this in no way detracts from the philosophy nor the conclusions of the article.

A revised section of the calculation referred to is shown below.

 

 

[JasonBModerator @jasonb]

Thanks for that and also correcting my Typo, just looked at what I wrote when working it out and it was 0.112.

 

That was based on your 6″ but for the Harris example his 6″ is the Inside diameter in which case 6.173″ should have been used in the calculations, this gives 0.115″. However as this has a similar effect on both methods the percentages stay about the same. (Assuming Evans is using OD as I don’t have his book)

I’m sure those that have posted hera are keen to know why we have not seen boiler failures using the old methods and will therefore need convincing of the need to change the way things are done particularly as you say your results consistantly give a thicker boiler shell which will add to the already hight cost of a copper boiler and the resulting thicker barrel will have a knock on effect on room to get tubes in, reduced water space and reduced grate area. What are your thoughts on this and assuming the flat plates are also annealed what do you propose to calculate their thickness and subsequent stay spacing?

[Charles LamontParticipant @charleslamont71117]

Erm … The pressure is exerted on the inside of the boiler shell.

[Michael GilliganParticipant @michaelgilligan61133]

Yea-But the failure will start on the outside skin, which is in tension

[ just a thought from one uninitiated in the art, but observing from afar ]

Please shoot me down if that’s nonsense.

MichaelG.

[JasonBModerator @jasonb]

Charles, I was making the point that the suggested YP method as in the article used D as the OUTSIDE diameter so should have been altered for the calculation. If however the ID should have been used then that is not what is shown so could that be another error?

[JasonBModerator @jasonb]

Online calculator also used OD of tube.

Interesting if you put the OD of harris’s boiler in and fiddle with the thickness until you get 180 (2 x WP) the number that works is 2.35mm. Now to me that is not far off the 3/32″ that Harris says “will be perfectly satisfactory”  Though the online calculator does not allow for a butt strap.

[JasonB]

On JasonB Said:

5″ OD x 75psi.

 

I assume that the original 13swg design was found to be a bit wanting and that is why Haining upped it to 10swg (0.126″)

I now wonder if the reduced thickness came about from either UK tube stocks or simply the shortage of copper back then. BT switched to aluminium busbars and we could not get some cable at all.

[MEinThailandParticipant @meinthailand]

JasonB, since we prepared the Yield Point Method proposal over 6 months ago we have been busy researching and refining the subject almost on a daily basis including flat plate analysis and FEA modelling.

We have also compared the commonly adopted ‘standards’; UTS method, YPM method and the Australian Code by AMBSC.

We have reached a conclusion that should help to resolve misunderstandings of the differences between these methods and provide clarity for the design of copper boilers going forward.

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.

 

 

[SillyOldDufferModerator @sillyoldduffer]

These formula are both simplified models of reality and as such should be taken with a pinch of salt.

Are they good models?   Certainly not complete – factors like temperature. metal fatigue and thermal conductivity are ignored.  As are the other important constructional features that affect strength: staying, joints, penetrations, and cylinder ends etc.

Both models seem to be based on Barlow’s Burst Formula for Pipes which will get order of magnitude right.  Though Barlow broadly decides the answer, the models aren’t identical: the UTS model factors in Ultimate Tensile Strength, not wrong except the UTS of Copper varies by about 50% – it depends!  Nonetheless, results are similar, but in both cases the maths generates ball-park estimates, not spot-on accuracy!  Anyone care to calculate the error margin in these formula?

No matter!  Both ensure the boiler gets nowhere near burst pressure by applying a Safety Factor.  Remembering my maths is terrible, I decode the formula thus:

• they start by estimating the thickness of material that will just fail at design pressure, then
•  the material thickness is increased by a safety factor, the size of which is left to the designer or the code he is following.

In the article, the YP example has a safety factor thickness x 3 whilst UTS applies thickness x 8.  Quite a difference.   In manufacturing safety factor varies between about x2 for non-critical items, up to about x25.  The safety factor of an airliner is much higher than that of a guided missile because the latter only flies once!

The size of the safety factor makes a marked difference to the calculated shell thickness.  Maybe the debate disappears in a puff of smoke because safety factor is an arbitrary multiplier!

Might put Bob Worsley’s mind at rest.  His kit came with 2.5mm rather than 3mm Copper.  Maybe someone decided to increase the Safety Factor rather than the 2.5mm boiler being dangerously weak. What does the boiler inspector think?

Though both mathematical models produce estimates rather than precision answers, large Safety Factors make them fit for purpose when designing model boilers.  Unfortunately I feel their simplicity makes them less suitable for MEinThailand’s purpose; his desire to push the envelope requires better.

I’m all for pushing the envelope, and suggest 3D-CAD and Finite Element Analysis is the easiest way to tackle it.   Barlow’s Formula was derived from a sharply focussed mix of theory and experimental data: he burst a lot of pipes made of many different metals and then did a regression analysis.   To prove MEinThailand’s idea by applying Barlow and the other engineering formula needed to calculate stuff like tension and compression in stays and joints is hard work, even with spreadsheets.   FEA is more general:  it knows nothing of Barlow, so the operator need not either! Rather objects are triangulated so that the stress in each side of many triangles can be calculated.  Accuracy depends on the number of triangles in the mesh and there can be millions of ’em.  FEA’s mesh copes automatically with stays, joints, ends, shapes and complications of all kinds.   Main problem seems to be getting the 3D CAD model to represent reality adequately and then to understand the results.  Also hard work, too much for me!

 

Dave

 

 

 

 

[Michael GilliganParticipant @michaelgilligan61133]

Just as a rider to Dave’s dissertation

I’ve said it before, and I’ll say it again

Finite Element Analysis is, by definition, almost always an approximation of an infinitely detailed reality.

MichaelG.

[Howard LewisParticipant @howardlewis46836]

Purely as a disinterested spectator, what are the formulae used by full size boiler makers?

Do they use similar formulae, or something different?

Failure of a full scale locomotive or traction engine boiler will be far more lethal and damaging than failure of a model; not that i am minimising the danger of models! Even at relatively low pressures, the kinetic and heat energy in a boiler poses a major danger.

A friend lost the end of a couple of fingers when a small pressure vessel exploded.

I have a photo of a boiler failure of a post war, US articulated locomotive, boiler.

It leaves no doubt as to the capacity to kill and damage. Crew killed, boiler wrenched from loco frames, track torn up!

Howard

[JasonBModerator @jasonb]

Alan

I have swapped a few emails and PMs with a few others who do this sort of thing on a daily basis, They may post here but I’ll not repeat what they said about what has been shown so far as a lot is over my head but suffice to say there are flaws.

If you do decide to publish further findings then I’m sure a lot more would be convinved if you could show some actual test boilers, nothing much just a tube with endcaps that has been built to either method and show the results after pressure testing. Failing that show some of the boilers that have been made to the YS method and have passed local inspection.

Howard. As far as I know the YS method is not an accepted industry method on full size stuff.

[noel shelleyParticipant @noelshelley55608]

For those who want to know what happens when it ALL goes wrong see “Historic Boiler Explosions” by McEwan. I fall back on the fact that there so few incidents that what is being done is not far off right. Hair splitting or excessively high safety factors prove nothing. Numerous full size accidents were the result of human error resulting in crown sheet failure – no amount of calculations Etc could have stopped what happened.  Noel

[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.

Your motivation is of course laudable and article interesting, but can I please take you back to my original post. The article states: “The premise of the new method is that when subjected to the maximum expected pressure during hydraulic testing, the boiler will not deform permanently.”

Could you please explain and justify your premise, rather that taking it, as you appear to do, as axiomatic? This is a challenge, not a criticism.

[Author]

Posts