Boiler thickness and pressure

[SteambuffParticipant @steambuff]

Looking at the drawings for the loco's I have or am building …

3.5" gauge CONWAY – 4" Dia : 13 gauge Tube

3.5" gauge LILLA – 4" Dia : 13 gauge Tube

3.5" gauge DARJELLING – 4 3/4" Dia : 1/8" Tube

So 16 Gauge at an even larger diameter than above seems way too thin for me.

Dave

[nigel jones 5Participant @nigeljones5]

Hi Julian. I can assure you that I am far from bonkers! The fact remains that there is a published and accepted formula which if applied to the letter would allow the use of the gauge I referred to. As a scientist I can either accept the formula or reject it. It is either right or wrong and the fact that no one has published a boiler to this specification is completely immaterial to the fact. I have no intention of building said boiler without thorough understanding, but as yet I have not seen a single argument which in any way proves the formula to be wrong. Indeed if we had never departed from accepted standards we would still be building wooden battleships and car bodies would be made of 1/4" steel plate. I openly welcome constructive input and criticism but without supporting evidence it is merely an opinion and they don't stand up to scientific scrutiny. I still fail to understand upon what grounds an inspector could fail the boiler other than he didn't much like the look of it. Why should it be rejected if it satisfies published and well used formula?

[Steven VineParticipant @stevenvine79904]

If there is 'any doubt' or conflict, it seems prudent to err well on the side of caution when dealing with pressure vessels. I'm in the midst of planning a boiler and my main concern is not causing harm to any nearby onlookers, so the shell will be 3.25mm, silver soldered. The cost of materials does not really figure in my calculations.

Substituting 2x test pressure for Wp into the formula seems a good idea. I read this other thread earlier today ….. **LINK**

Steve

[JasonBModerator @jasonb]

Using the 2x pressure dose not have such a large bearing on the outcome as you will be working with a cold boiler so no temperature allowance needed in the calculation so one will to some extent cancel out the other.

Julian do you know what Maths was used by Don Young eg what stress value and what safety factor, it would be interesting to actually be able to do the calculation using Fizzy's diameter and working pressure and see what thickness it throws out. It's these two factors which seem to vary between the various people – Harris, Haining, tubal Cain, evans and presumably Young.

Also why should the formula vary between a loco boiler and that of say a marine or vertical boiler if the fact Harris not having built a loco one is of significance? The only one that I can see may need to be thicker is a traction engine where the boiler is also a stresses chassis

J

[FMESParticipant @fmes]

Having been testing boilers since the early eighties you do get accustomed to the various sizes and gauges of materials used.

I was trying to find a document from our old insurance company of some years ago that 'recommended' a factor of safety of '8'.

Unfortunately I couldn't find it (time has passed) but I always used that figure for any pressure vessels that did not have a current design available.

This boiller in question is similar in size to others and as previously mentioned would expect to see a minimum of 10swg for its construction.

The other point worth thinking about is that the heavier the boiler construction the better the tractive adhesion is likely to be, hence removing some of the need for added lead weights.

And to answer Jason directly, I seem to recall a clause in the Boiler Test Manual (Southern Federation) quite clearly stating that:an inspector may,if he is not satisfied that a boiler and/or its associated equipment and fittings are fit for service, refuse to undertake an examination and/or test. The reason for refusal shall be given in writing to the owner.

[JasonBModerator @jasonb]

Thanks Lofty, but what would make the boiler inspector feel the boiler was not fit for service, surely it must be more than just "the old boys designed them that way" yours if its not the same so it does not get a ticket. If the person presenting the design had done the figures to show it was fit for purpose in their view then I would assume you would show your own figures when you give the written reason?

What if the new boiler presented to you for testing was made by a professional and had passed a 2x test would you still turn it away?

Also you say you would use a factor of safety of 8, that is the same factor Fizzy used in his opening post and got 16g material, what formula do you use with this factor to arrive at double the thickness?

J

Edited By JasonB on 24/10/2014 07:57:21

[JasonBModerator @jasonb]

Fizzy, you still have some unread PMs from a few days ago, please read them

J

[Mark CParticipant @markc]

This is an interesting discussion and as such, I decided to sketch up a simple "boiler". It is drawn to the tube and materials mentioned in the first post and I have used the stock material "copper" from the materials selection in Solidworks with a pressure of 70 psi and a safety factor of 8 as indicated. The "boiler has a flat end with a 1/4 inch fillet between the end and the tube and when I run a simple (cosmos express) stress analysis on the model you get it failing as shown in red! The analysis does not take account of temperature but I understand that this is done by using the SF of 8?

If someone wants to give me some more detail on the shape and thickness of the parts then I might see about sketching something more accurate.

Mark

Edited By Mark C on 24/10/2014 10:49:38

[S.D.L.Participant @s-d-l]

Posted by fizzy on 23/10/2014 21:55:00:

Hi Julian. I can assure you that I am far from bonkers! The fact remains that there is a published and accepted formula which if applied to the letter would allow the use of the gauge I referred to. As a scientist I can either accept the formula or reject it. It is either right or wrong and the fact that no one has published a boiler to this specification is completely immaterial to the fact. I have no intention of building said boiler without thorough understanding, but as yet I have not seen a single argument which in any way proves the formula to be wrong. Indeed if we had never departed from accepted standards we would still be building wooden battleships and car bodies would be made of 1/4" steel plate. I openly welcome constructive input and criticism but without supporting evidence it is merely an opinion and they don't stand up to scientific scrutiny. I still fail to understand upon what grounds an inspector could fail the boiler other than he didn't much like the look of it. Why should it be rejected if it satisfies published and well used formula?

Hi Fizzy

The answer that I posted earlier was based on the Martin Evans formula. I have been rooting around in my workshop and found a guidance note titled.NORTHERN ASSOCIATION OF MODEL ENGINEERS BOILER INSPECTORS SEMINAR. Boiler Design Materials Calculations. In this they recoment that the Martin Evans formula of

P=( D*F*W.P)/(T.S.*R*C*T*2) be used

Now if I was a Boiler inspector at a club in the northern federation that is what I would use until the guidelines were changed.

On a different note, I have never used a strength of 1 on welded joint design even with coded welders hence my use of .9. When you were coded did the welds still have to be x rayed on butt welds and dye pen on fillets?

Steve

[JasonBModerator @jasonb]

Mark C Very interesting analysis, at this stage Fizzy has only mentioned the barrel thickness so thats all we can go on at the moment. But does seem to show that a thinner barrel is in theory possible.

It is not a surprize that the end failed first as they are no form of stays which are a must for any boiler.

SDL would it be possible to put some figure sinto that formula particularly the stress value that you used as that seems to be one of the big variables.

I have a couple of interesting things to add to this post but will have to wait until I've finished work

[Anonymous]

I have been looking at a similar calculation recently to work out what wall thickness I need for the copper pipe from the water pump to boiler clack on my traction engine. I had a quick look at the equations in the Harris book, but dismissed them. It seems logical to start with basic equations to calculate the hoop stresses at an appropriate working pressure irrespective of material. Then allowances can be made for temperature if needed. Finally values for the yield stress, and safety factors, can be used to see if the calculated hoop stress is acceptable for a given pipe diameter and wall thickness.

The Copper Development Association quote a 0.2% yield stress for C106 coppper in the range of 50-340N/mm². The lower value is equivalent to 7252ksi, which is what I am using for my calculations.

Regards,

Andrew

[julian atkinsParticipant @julianatkins58923]

in his book 'model locomotive boilers' martin evans provides the formula used by most on p.28,

and on p.29 gives a worked example for a 4" dia copper locomotive boiler where the thickness works out as 90 thou or in other words 3/32" or 13 swg. (these days you would use 2.5mm).

club boiler inspectors will know these pages intimately.

hence my comment that fizzy's proposal to use 16swg for a 5.625" dia boiler is bonkers.

(my copy of the above book is the first edition 1969)

you then have to consider what seamless drawn tube is available commercially – since the above book was written many of the sizes of barrel shown on drawings are no longer available.

cheers,

julian

[David JuppParticipant @davidjupp51506]

Posted by Andrew Johnston on 24/10/2014 12:09:32:

 

The Copper Development Association quote a 0.2% yield stress for C106 coppper in the range of 50-340N/mm². The lower value is equivalent to 7252ksi, which is what I am using for my calculations.

 

 

I think that should read psi NOT ksi (factor of 1000 different).

Edited By David Jupp on 24/10/2014 12:21:45

[Anonymous]

Posted by David Jupp on 24/10/2014 12:21:14:

I think that should read psi NOT ksi (factor of 1000 different).

Ooops, indeed it should – Andrew

[S.D.L.Participant @s-d-l]

Posted by JasonB on 24/10/2014 11:26:54:

SDL would it be possible to put some figure sinto that formula particularly the stress value that you used as that seems to be one of the big variables.

Figures I used are shown below from my spreadsheet based on Martin Evans book which are the same as the Northern Association guidelines I mentioned earlier.The .9 join allowance is what I would allow, drawn tube is 1.

this gives .091" or 2.32mm

 

Steve

Barrel thichkness

Source

Model locomotive and marine boilers

By Martin Evans

ISBN 1 85761 138 1 P28

Where

P Plate thicnes in inches

D OD of Barrel in inches

 

F Factor Of Saftey 8

 

WP Working pressure PSI

 

S Ultimate tensile strength cu 25000 psi cs 60000 posi ss 70000 psi

 

R Riveting allowance single rivet = 0.5 double rivet 0.7 silver solder=0.8

 

C Corrosion allowance cu 0 ss 0 cs 0.5

 

T Temperature allowance <100psi = 0.8 100-150psi = 0.7

 

 

I used as below

 

 

	P	=			0.091	"
when	D	5.875	" OD
	F	8
	WP	70	PSI
	S	25000	PSI
	R	0.9	Joint Factor
	T	0.8	Temperature Factor

 

Edited By S.D.L. on 24/10/2014 13:26:06

[JasonBModerator @jasonb]

Thanks SDL. Seems that the Evans based figures are not what the professional boiler makers use, see the folowing.

[JasonBModerator @jasonb]

Although I don’t want anybody to rush out and make a boiler from a bit of copper foil and I’m not knocking any of our club boiler inspectors who do a very good job keeping the hobby affordable some may not like the following but it is based on fact.

There seems to be an opinion that some of these old designs that new ones are having to be based on were a bit over the top as far as materials were concerned and the other side of the argument has not yet come up with any figures to back up why these thick boilers are still what we should be using apart from they have been like that for ages when boilers were put together with a paraffin blowlamp and pile of coke.

Over the last few months I have been party to a friends alterations to a published boiler design, mainly to bring it upto date with current requirements eg separate feed to pressure gauge but also to get a working pressure of 120psi which is the point of this post.

The original design which is by a published author with at least a dozen model designs published most of which have been in ME at some time. It is a 5” OD boiler x 10swg in copper and silver soldered, drawn tube for the barrel and with a working pressure of 70psi.

Now my friend does not have the facilities to take on the boiler build so a professional builder (member of the assoc of prof boiler makers) was asked to come up with a new design that met the 120psi requirement which is a 70% increase in the original designs working pressure.

So what’s on the drawing for this new boiler?

3mm rolled and soldered copper barrel. So not only down from the 3.25mm (10swg) but also a joint right down it which from the calculations looks to be about 0.8 as the joint allowance, the higher temperature would also have been figured in to go along with the higher working pressure so all in all the old design was about twice as thick as it needed to be.

Now there is also the fact that this was for a traction engine, as mentioned last night by someone this makes the barrel a structural chassis so make an allowance for that. Now there is also the cyclic effect of a big compound cylinder stuck on top so again make an allowance for that in the calculations. Oh and did I mention it’s a ploughing engine with a longer barrel than most traction engines not to mention a big old casting an about 70m of cable slung below that will weigh in at around 12kg. And best not even think about the loads when it’s got a scale plough hooked to it.

So a boiler sitting comfortably between a pair of loco frames should have a very easy life. This seems to also be borne out by the stress analysis posted today where the 16swg barrel did not fail, only the unstayed end.

SDL's figures would give a thickness of 0.070" for the original 70psi design pressure and seamless tube

Where this new boiler has had changes is in the staying and flat plates, something that Fizzy mentioned was where the stresses were in his home experiments. I won’t go into too much detail about the changes but the main one is to do away with girder stays which do appear on these old hand me down designs and go for radial rod stays between the top of the firebox and crown of the boiler. Much like the Australian code requires

Well something to think about, we have moved on from riveted and caulked boilers and ones using sifbronze so why not move on from these heavily over engineered designs that people seem to want all boilers to be like. You moan enough about the cost of copper well if you are using twice as much as you need to who is to blame.

That’s enough for now I’ll hide behind the sofa for a while but may have a few more tip bits to add. Hopefully this will make for civilised discussion and get people looking at boiler designs in a new light rather than just following whats already been done

J

Edited By JasonB on 24/10/2014 14:42:10

[Mark CParticipant @markc]

Jason,

The FEA I posted, if you only consider the tube section, is little more than an "easy" hoop stress calculation as mentioned by Andrew. The information was just to give some indication as to what might be happening in a pictorial form, it should not be used for any justification of design etc.

Mark

[Anonymous]

The safety factor of 8 does seem rather high to me? However, in at least some of the calculations it appears to be used to account for, at least in part, the difference between ultimate tensile stress and yield strength. So presumably if one uses yield strength one can use a lower safety factor, whatever that might be?

Regards,

Andrew

[Mark CParticipant @markc]

Andrew,

I would think 8x factor also takes account of the reduced strength due to temperature etc. I do not know what the reduction is for copper but mild steel is reduced to 40% at 400 degrees C. I am certain the early posts mentioned it took account of temperature?

Mark

[JasonBModerator @jasonb]

From SDL's posting the temperature is taken account of by a separate allowance and the safety factor remains the same at 8.

J

[David JuppParticipant @davidjupp51506]

I'm not familiar with boiler codes, only some unfired vessel codes.

Over the last few decades, required design factors in the codes have tended to reduce somewhat, reflecting improved metallurgy, application of quality systems in manufacture, the use of more sophisticated stress analysis etc.

Sometimes the lower design factors are allowed only in very limited circumstances, but there has been a general reduction in required design factor.

A code with higher design factor typically uses more material, but requires less/simpler analysis, maybe less testing, and perhaps requires pressure test to a larger multiple of MWP.

The saving in materials from a lower design factor code may be offset by costs of extra testing, additional stress analysis, fatigue life evaluation, and 3rd party inspection costs.

Both approaches have merit – and both should result in equipment that is fit for purpose. I feel that some of the contrasting opinions here reflect this. It is important not to try to 'cherry pick' aspects from both approaches – that would lead to major problems.

Edited By David Jupp on 24/10/2014 15:46:30

[S.D.L.Participant @s-d-l]

Posted by JasonB on 24/10/2014 15:24:04:

From SDL's posting the temperature is taken account of by a separate allowance and the safety factor remains the same at 8.

J

The martin Evans book that I took this from has good worked examples and certainly the tube thickness is what is in the Northern association guidelines. Note these design guidelines are separate from the testing book. This should not stifle developments but few boiler inspectors are going to have the experience to review FEA but this should not stop people adding blind bushes for fixings of door runners etc

Please spare a thought for your Boiler inspectors, many do a thankless job, at the end of the day its them that has to sign off and pass things. At our club for many years the boiler inspectors thought they were covered by the club insurance, It was only when I raised a question in writing during the review before the last code change that it was made clear that the club needed separate indemnity insurance for the boiler inspectors otherwise there assets are on the line. as you could imagine we took this out promptly. This insurance is shown on the southern fed form but historically the implications had not been understood.

Steve

[CrabtreeengineerParticipant @crabtreeengineer]

Hi guys, I know I have come this thread a bit late, but heres my two cents worth! Most burst calculations are based around Barlows equation. If I calculated this I would use UTS & apply my temperature correction on this value before calculating further. I would take a nominal ambient temperature of around 24 degrees C & calculate my temperature offset from that. Then plug that into Barlows equation. I note that some of the equations quoted seem to incorporate a temperature factor into the entire equation. This may well be one reason there is such large descrepancies. Regards R

[Author]

Posts