Boiler Design – issue 4765

[Charles LamontParticipant @charleslamont71117]

I found the boiler design article interesting. But I am unconvinced. The authors start from the premise that there should be no permanent deformation of a copper boiler subjected to its hydraulic test pressure. Why not? I don’t accept this as axiomatic.

[Paul LousickParticipant @paullousick59116]

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.

[JasonBModerator @jasonb]

It would have been nice to see some actual test pieces and compare them to the theory.

Although the authors say at what pressure the barrel will start to deform permanently they do not say by how much, if only a thou or two is that going to be a problem or will it end up looking like a balloon? Paul, is there an allowable deformation in diameter in the Aussie code or must it be zero?

With modern methods it would also have been good to see the two methods shown with FEA

It also makes no allowance for the fact that a boiler is likely to have had a few tests done by the builder at hme often starting at les sthan 2x WP which will start to harden the copper.

The loco boiler calcs are quite similar only being 4 thou difference in thickness but the Harris boiler is a big difference 0.091″ Harris to 0.168″ yet many have been made over the years and I’ve not heard of them going barrel shaped.

Putting the Harris figures from p31 into the spreadsheet that several of the current commercial boiler makers use then it is still over allowable figures so is there a need to go 50% thicker with the barrel material? Paul what would the Aussie code require for a 6″ boiler at 90psi WP?

Later in the book Harris shows his actual design for a 6″ vertical boiler which has 1/8″ thickness barrel (could be where the article got the 3mm from?) that is almost double the allowable factor of safety figures

[JasonBModerator @jasonb]

Ops I think I spot a cock up or two.

For the Harris example they give the shell thickness twice as per the other example but have the first as 3mm and the second as 0.0865″ should that 3mm be 13swg in a similar way to the first examplethat gives SWG and then decimal thickness?

However the main thing that looks wrong to me is that for the Harris example the 90psi working pressure has been multiplied by a factor of 3 not 2 for a test of twice working pressure. This would give a result of a wall thickness of 0.122″ rather than their result of 0.168″

Think it is time for a letter to the mag unless the learned gentlemen are reading and would care to comment?

 

[JasonBModerator @jasonb]

[David JuppParticipant @davidjupp51506]

I didn’t see the article in question, but I offer a little background from general pressure vessel codes/standards.

Codes and standards vary, but usually include maximum working stress values as a proportion of both ultimate stress and of yield stress – for example a code might allow maximum working stress to be the lower of      1/2 x Ultimate Tensile Strength   OR    2/3 x Yield Strength  of the material   (both at working temperature). Joint factors and perhaps other adjustments get applied, plus corrosion/erosion allowances if applicable.

Testing (typically at ambient temperature) will be to some multiple of maximum working pressure, but also subject to not exceeding (perhaps 95% of) the yield stress of the construction material at test temperature.

Here we are considering general yielding – it is accepted that there may be some very localised yielding in the vicinity of stress concentrations.  There are additional calculations to check that such localised yielding will lead to a beneficial redistribution of stresses, rather than leave a situation where there will be future localised yielding in service (and ultimately a failure).

It is common to check (e.g. using strain gauges) that there was no permanent deformation during pressure test.

There are special cases (autofrettaged thick walled pressure vessels) where some yielding at test is allowed for, and is even beneficial – but these are well outside the scope under consideration here.

 

[Paul Lousick]

On Paul Lousick Said:

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.

A small amount of permanet movement in a copper boiler when first pressurised is not necessarily a sign of weakness.

I know the article is only talking about hoop stress in the cylindrical shell, but someone is going to misinterpret this as applying to, for example, flat stayed surfaces as well.

(I also have a B.Sc.Eng.)

Where do I get copies of the AMBSC codes?

[Bob WorsleyParticipant @bobworsley31976]

As Jason comments, I didn’t understand the working either and agree with what he says.

Likewise if the boiler ‘gives’ during a test, so what, no one can measure the odd thou, if it becomes balloon shaped then quite a different matter. There have been many articles in ME over the years about boilers, and by the very absence of explosions then what is designed and built seems to work. That copper work hardens is also well known, so using annealed strength rather than as drawn doesn’t seem to be sensible.

I was surprised to get a boiler kit for a 2″ BB traction engine that used 12swg/2.5mm barrel, I would have thought that 10swg/3mm was really called for, as mentioned in the article.

 

[JasonBModerator @jasonb]

That does seem thin for a ploughing engine with the added stresses it may see due to length and possibly pulling something. Have you checked that it will be approved by however will be testing it as that is quite a deviation from the accepted design?

[JasonBModerator @jasonb]

Oh Dear, Harris uses inside diameter on page 31 but they have applied the 6″ as if OD in their calculations assuming they are using OD as shown in fig 2.

 

I don’t have a problem with the suggested method but with errors it is going to be hard to convince people to change. Even more so when the corrections are made and there is only a few thou difference in wall thickness between the two methods.

Am I expecting too much that with their results showing such a difference in thickness and no reported failures that they may have gone back and checked the calculations, particularly as it was a joint effort at least one should check the other?

[Bob WorsleyParticipant @bobworsley31976]

The BB boiler is as designed in the ME article. Still wondering about that.

 

[JasonBModerator @jasonb]

Interesting, the BB1 drawings that I have show the shell as 10swg and the same for all the other plates.

ME as you say 12g

Countryman’s Steam book looks to have a correction as the 10SWG is typed on but the 5″ and rest of the details are had written. The text also mentions a change to 10swg as well as the type of crown stays used.

[Bob WorsleyParticipant @bobworsley31976]

Ah, looking back to when bought from Reeves, August 2020, I see that the kit sent didn’t match the drawings and I queried the thinner copper, also that the barrel and firebox were separate so some sort of joining band needed. Got no admission of a fault since it was to the (old?) published design, so still have it all, hence my wondering since not started making it.

[Bob WorsleyParticipant @bobworsley31976]

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.

 

[LukerParticipant @luker]

Unfortunately I have not read the article in question (I’m not a subscriber) but I can give some insight into copper boiler design. You will get deformation when hydraulically testing a copper boiler because it’s in an annealed state. Even when TIG welding; the preheat and interpass temperature will result in an annealed state. During the test the material deforms and the stresses are relaxed, the material also goes through isotropic hardening, resulting in a stronger boiler. I actually simulated one of my boilers using a non-linear hardening material model and the strength improvement is significant (this was presented at an international FEM conference, due to the complexities of the coupled physics and linked model states).

Personally I’m not a fan of copper boilers, for various reasons, but an article that is fundamentally flawed will dissuade builders making their boilers. If the design has been made and tested (i.e. proven design); then I can’t see how anyone can question subsequent builds. I won’t comment on the modern ME codes, simply because I don’t think they are written by qualified builders!

[Paul Lousick]

On Paul Lousick Said:

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.

No deformation under load is not possible with any engineering material that I know of.  Permanent deformation is another matter and would indicate plastic deformation has taken place as a result of exceeding the yield stress.  With a material like annealed copper where the yield point is very poorly defined things get more nuanced.

I assisted in translating a FEM analysis on a Crampton boiler (published in ME some years ago).  That indicated some localised plastic deformation in firebox corners and stays would occur on a well proven published design.  I agree with others that would be of no consequence.

What interests me is how such deformations would be measured during a hydraulic test (under Aussie rules?)?  A picture of the set up would be interesting.

Martin

[Martin Johnson 1]

On Martin Johnson 1 Said:

 

What interests me is how such deformations would be measured during a hydraulic test (under Aussie rules?)?  A picture of the set up would be interesting.

Martin

Measuring for permanent set would be before and after, unpressurised, not during.

[duncan webster 1Participant @duncanwebster1]

I know I keep droning on about this, but the Australian standard is extremly clear and is referenced back to national standards.  The article in MEW was not clear, in places was wrong. Why reinvent the wheel.

[Robert Atkinson 2Participant @robertatkinson2]

Plastic deformation during a hydraulic test can be detected by monitoring the volume of fluid in and out during pressure cycling. If less comes out than was put in there was permanant deformation. When testing gas cylinders they tend to immerse it in liquid and measure the displacement caused by external movement. Bothe these make allowance for small amounts of trapped air that can be hard to eliminate for simple internal volume measurement with liquid.

Robert.

[duncan webster 1]

On duncan webster 1 Said:

I know I keep droning on about this, but the Australian standard is extremly clear and is referenced back to national standards.  The article in MEW was not clear, in places was wrong. Why reinvent the wheel.

Where do I get a copy?

[JasonBModerator @jasonb]

See bottom of page here or you can get them here

Out of interest would a boiler built to the Australian code be readily tested and certified here in the UK by club inspectors even if it differed from the accepted heritage design?

Of the couple of copper boilers I know of built in Australia they had to alter them to the local code so I assume it does not work the other way round and that  UK designs are not accaptable unless the construction falls within the rules.

[Charles LamontParticipant @charleslamont71117]

Thanks, Jason

[JasonB]

On JasonB Said:

Interesting, the BB1 drawings that I have show the shell as 10swg and the same for all the other plates.

ME as you say 12g

Countryman’s Steam book looks to have a correction as the 10SWG is typed on but the 5″ and rest of the details are had written. The text also mentions a change to 10swg as well as the type of crown stays used.

The drawing does indeed have a strip of paper covering up the original size with 10swg etc, but turn the page and everything reverts to 12swg and 13 swg firebox. I’m not surprised Reeves 2000 supplied it wrong.

[Bob WorsleyParticipant @bobworsley31976]

Doesn’t really help me, still supplied with 12swg/2.5mm and Reeves said ok, and Neil Tyler never replied to the query, which I think is really not on.

I bought the BB as a part complete model, and the seller had had enough of the miserable quality castings from Neil Tyler, having to weld additional aluminium onto the balance plough castings isn’t on either. So it might be that the old Haining traction engine models are now not worth looking at? Shame, some nice engines, always fancied a Z7!

Along with the non-existant service from Brunell with the 16hp ploughing engine I didn’t receive, is the hobby dying? The quality of the castings from50+ years ago is excellent.

 

 

[JasonBModerator @jasonb]

The Haining designs were OK in their day but thing shave moved on.

It is so much easier to find reference material on the net that would never have been available back in the day unles syou had acces sto an engien to measure so the budding 21st centry builder will see the short cumings.

They were also buildable with what the average model engineer had available which would likely have been a Myford with a vertical slide. Now anyone taking on a traction engine is likely to have a mill and things like laser cutting are only a few mouse clicks away.

I would also say that there is less demand for the 2″ and smaller engines these days which at one time would have been considere dlarge but look at any TE rally and people are building bigger. The 2″ suits those who are interested in traction engiens but not really so interested in spending the weekend in a field.

Many of these old designs use the same patterns that castings were done from back in the day so that should not affect quality too much unless the patterns are too worn or damaged. A lot will be down to the foundry if one can be found these days. The Superba castings that I have ( could be talked into parting with) are reasonable but I don’t know their exact date. Hainings patterns were not true scale, lack detail and the same are used for more than one engine which although Fowlers did to some extent is not what would be needed to make a truer model. So can you really say they are better?? From the photos I see on Reeves Facebook page they castings they are currently getting look quite good.

On the other hand going down the exact replica route can end up with an engine that is not practical to run, that very nice 1.5″ BB1 made by the Dutch guy being a good example as it became a static model by the time he was done.

The hobby seems to be dying at least on this forum. Elsewhere I see plenty of models being made but the way they are being made has changed as more people seem to be using newer methods to produce the models they want be that making even simple parts from a piece of bar or block rather than paying several times as much for little more than a cast rectangle or cylinder that may have casting flaws to using CNC, 3D printing, cutting from solid, fabricating,etc.

Bob I did send you a PM a couple of days ago.

[Author]

Posts