To Flange or not

[runner42]

A colleague has been conducting finite element analysis on the possibility of producing copper boilers without flanging the plates. The results show that it is possible without reducing the integrity or strength of the boiler. The assumptions are that the joining is by use of silver solder and that the thickness of plate is slightly increased for any given operating pressure, than is currently specified in prevailing boiler codes. The slight increase in plate thickness would more than compensate for the need to flange the plates. I mentioned that this proposal would not get much traction amongst the  diehards who have traditionally used flange plates, particularly in the UK where there is a long tradition of model boiler construction, but apparently it was from the UK that this was first proposed.

This is an initial study and perhaps more investigation may need to be undertaken before such changes occur in boiler codes, but before that it may be appropriate to solicit members views on this. Not so much to challenge the results but to explore the practicalities of constructing boilers without flanged plates. I see that mechanical joining the flanged plates to the adjacent copper sheet prior to silver soldering is easier than would be with an unflanged plate.

Brian

[uuu]

The welded boilers that I've seen have not used flanges. This might be part of the reason they're cheaper than the soldered ones.

I had always thought that it was the larger wetted are that made flanging superior. Boilers with an unflanged throatplate to wrapper join were always sniffed at. But the joint can clearly be made to work in that location, so why not everywhere else?

It might be interesting if the study could examine the performance of the joins around a first-failure point. If you get a small hole, does the seam unzip, or hold firm?

Wilf

[RGR 60130]

This seems like an ideal opportunity to record a few of the things I was taught about pressure vessel design which might be of interest and perhaps thought provoking.

The concept of building a boiler without flanges will no doubt alarm a few people but once you look at the science it's not as crazy as you might think. For a thin wall vessel (Plate thickness up to 5% or 10% diameter- definitions vary), the stresses that want to split the boiler barrel down its length are twice those that want to blow the end out. In practical terms this means that the stresses on, for example, the front tube plate join are only half those on the longitudinal seam of a fabricated barrel. The flange therefore is not as important as you may initially think. Things like sharp corners which are stress raisers and joint integrity come into play, however the 2:1 ratio still remains. I read an article (in one of the magazines?) where someone did lots of experiments to force a boiler to fail and the results were in line with the theory though I don't think the chap was actually aware of the theory.

With stresses in mind the issue of boiler penetrations should also be considered. Man-way doors, generally elliptical in shape should be placed with the short axis on the longitudinal line of the boiler. The meaty ring you usually find surrounding a man-way isn't just there to provide a sealing face for the door, it's also there to replace the strength of the material removed when the aperture was cut. Having said that, I have seen photos of traction engines with the man-way placed the other way round.

Other penetrations should be considered similarly. Bushes for safety valves etc; in addition to providing extra threads also replace lost material. Dome bushes are probably the biggest hole in most model boilers and should be robust.

The success or failure of flange-less joints will rely primarily on joint integrity and insurers may well have concerns about how competent a man in a shed is at consistently producing the joints. A flanged joint may perhaps be seen to give a bit more leeway to the amateur constructor.

Other views will no doubt prevail.

Reg

[chiptim]

This is an interesting topic.

I believe that the joint strength is dependent upon the gap dimension at the time of soldering. A flanged plate would likely leave a more uneven surface compared to a cut plate such that the joint gap dimension deviates more from the ideal?

But then we need to analyse whether a weaker joint (in places) over a distance say 3 - 5 times the plate thickness due to the flange is better than a thinner and potentially stronger (with more uniform strength) joint over just the plate thickness?

I struggle to imagine how I'd have held my boiler together to solder it without the flanges and the copper screws I made?

Tim.

[ettingtonliam]

Wasn't Mr Austen Walton much derided for designing the boiler for Twin Sisters with no flanges? Maybe he was way ahead of his time?

[jo479]

An interesting point is that, what drawings are tig welded boilers being made to, I asked the question to a firm making tig welded boilers the last time at Doncaster, the boiler concerned was Lbsc's Pansy, the answer was Lbsc's drawings, to which I replied have you had them redrawn without flanges or are you just ignoring them, you are supposed to make a boiler to an approved drawing,everybody was suddenly very busy. I've asked the same question to our insurance company, you don't get any answers off anybody. As a Club boiler inspector I'm concerned, serious comments very welcome.

[JonL]

So is the primary difference that a flanged boiler uses a filling material and a non-flanged boiler relies on a weld which means the material is effectively "one piece"?

[uuu]

Aug 24, 2021 20:35:11 GMT jo479 said:

An interesting point is that, what drawings are tig welded boilers being made to, I asked the question to a firm making tig welded boilers the last time at Doncaster, the boiler concerned was Lbsc's Pansy, the answer was Lbsc's drawings, to which I replied have you had them redrawn without flanges or are you just ignoring them, you are supposed to make a boiler to an approved drawing,everybody was suddenly very busy. I've asked the same question to our insurance company, you don't get any answers off anybody. As a Club boiler inspector I'm concerned, serious comments very welcome.

Most commercial boiler makers don't follow the drawings exactly - they use their knowledge and experience to tweak here and there. Bushes instead of threaded holes. A slightly thicker gauge of copper. An improved stay layout. A reinforcing strip behind the barrel to firebox wrapper joint. We agree to these changes because of the knowledge and experience of the maker. I can't see you'd want to fail any boiler presented with these tweaks. I put the flangeless welded joint in the same category. If the commercial maker has relevant competence and experience to justify the change, then it's OK.

Wilf

[chiptim]

Aug 24, 2021 21:35:41 GMT JonL said:

So is the primary difference that a flanged boiler uses a filling material and a non-flanged boiler relies on a weld which means the material is effectively "one piece"?

Hi John, well the conversation has now included Tig welded boilers but Brian's original post was flangless boilers that were silver soldered.

[jo479]

Both are using flangeless plates, I would prefer flanged plates, silver soldered

[timb]

I would have thought it comes down to shear and tensional forces. In a flanged plate the shear force is applied over a larger area - the wigth o the flange say 1/4 inch, without the flange the shear is over a much smaller area - the thickness of the plate say 1/16 inch. Welded joints use the same metal to fuse the parts together so if done correctly (hence coded welder required) the resulting joint will be at least as strong and the material around it.

Having said that I have seen copper tear at a siver soldered joint where the copper was still soldered, ie the tear was as a result of the copper itself mechanically breaking, not the silver solder joint failing.

LBSC used to make boiler throat plates as butt joints to the barrel with a 'good fillet of solder' between the two. I do not recall any of his boilers failing but this practise is definitely frowned upon now.

Just my thoughts.

Tim

[jackrae]

My predecimal pennies worth

It is my understanding that amateur manufactured boilers are generally acceptable for approval and use (post testing) on the basis that they are manufactured in accordance with an approved (historical) design. It may be fair to suggest that most amateur builders either do not possess the knowledge (or the inclination) to conduct their own detailed design and stress calculations including suitable factors of safety. If I wished to manufacture a boiler to my own (non-standard) design I would be required to perform and produce full design calculations to justify my proposal and present it to a suitable authority (club boiler inspector) for approval prior to manufacture. The question that follows is 'does the club boiler inspector have the knowledge (as against "experience") to understand, critique and verify the design proposal and relevant calculations'.

Might I suggest that deviation from an approved flanged design to a flangeless one will require detailed stress calculations as per Brian's colleague until such time as a flangeless design is deemed as an "approved" design without the need for further calculation.

[runner42]

Sept 1, 2021 17:53:07 GMT jackrae said:

My predecimal pennies worth

It is my understanding that amateur manufactured boilers are generally acceptable for approval and use (post testing) on the basis that they are manufactured in accordance with an approved (historical) design. It may be fair to suggest that most amateur builders either do not possess the knowledge (or the inclination) to conduct their own detailed design and stress calculations including suitable factors of safety. If I wished to manufacture a boiler to my own (non-standard) design I would be required to perform and produce full design calculations to justify my proposal and present it to a suitable authority (club boiler inspector) for approval prior to manufacture. The question that follows is 'does the club boiler inspector have the knowledge (as against "experience") to understand, critique and verify the design proposal and relevant calculations'.

Might I suggest that deviation from an approved flanged design to a flangeless one will require detailed stress calculations as per Brian's colleague until such time as a flangeless design is deemed as an "approved" design without the need for further calculation.

Your response is predicated on the assertion that previous historical designs had stress calculations including suitable factors of safety conducted, but I don't think these were conducted routinely in the manner you indicated. I would imagine that initially they were designed on the basis of precedents, thickness of copper plate versus working pressure became a precedent. So designers used these data a "bit more of the same". Latterly physical properties such as the maximum allowable stress in annealed copper was introduced in the design process. Boiler codes prescribe certain criteria for all elements of the boiler design but does not require detailed design and stress calculations to be conducted. 

The only parameter of a design is being changed in the proposal is to omit the flange but increase the thickness of the end plate to compensate for the absence of the flange.

Brian

[jackrae]

Sept 2, 2021 0:18:20 GMT runner42 said:

Sept 1, 2021 17:53:07 GMT jackrae said:

My predecimal pennies worth


The only parameter of a design is being changed in the proposal is to omit the flange but increase the thickness of the end plate to compensate for the absence of the flange.

Brian

In a flanged build the joint is subject to both tensile and shear stresses. The stresses per unit area are, relatively speaking, very low and will tolerate some degree of lack of penetration of the solder in those joints that cannot be visually inspected. In a flangeless joint the stresses per unit area are, relatively speaking, very high and will be less tolerant to penetration failure on joints that cannot be visually inspected.

I therefore suggest that the change of this "only parameter" is of critical importance and will demand a superior degree of care in manufacture (both mating face fit and soldering) than the flanged variety.

[JonL]

The quality of manufacture is what I suspect is a lot more difficult to regulate in an amateur made boiler than a professionally made boiler where one would hope hidden issues with quality of construction would be less prevalent due to the number of boilers made.

It suspect boiler regulations are so stringent probably because if they demand 150 Percent overkill or whatever, on material thickness, joint thickness or whatever, that the times when there is a hidden flaw means that the risk is reduced as you are still within the margin of safety.

[runner42]

A Boiler Inspector specifies the stages of manufacturing assembly that requires a stage inspection. This is usually at the wrapper to end plate assembly and the criterion for acceptance is to visually determine that wicking through has occurred. This will be the same for flanged and non flanged boilers so if we accept that a silver soldered joint is stronger than the copper it is joining then the weakest point is in the copper end plate or wrapper. In a flanged plate this would be in the bend established by forming the flange. In a non flanged boiler this stress point is absent. The argument rests with how much copper contact is required to establish a good joint. The finite element analysis results show that only a slight increase in thickness is required to establish comparable strength joints.

Brian

[coniston]

If it is of any relevance Don Young specified 1/4" plate gusset pieces in the top corners of the throatplate to fill the holes between barrel and throatplate. This was in 1972! A boiler made to this 'published' design should be accepted by a club BI, although of course the BI has the right to refuse any boiler for inspection.

Just to show that non flanged plates, or at least parts of plates, are not new.

Chris

[runner42]

Like this? This is how my Black 5 boiler is constructed. However 1/4" inserts is slightly over sized to the issue being presented. In say a 3 mm thick plate which is flanged it would only need to be say 4 mm unflanged.

Brian

inserts by Brian Leach, on Flickr