Boiler Design – issue 4765

[lezsmith]

On lezsmith Said:

 

This thread is however specific to small stainless-steel boilers.

 

Have to disagree with you there.

Your article and subsequent discussions relate to COPPER boilers. The thread has only recentyl drifted towards Stainless.

 

Well hopefully your small Stainless boiler will be OK now that you have now been told the correct grade of Stainless to use. Strength does not just come from the material but also use of the correct weld design & prep, welding methods as well as sound design.

[JasonBModerator @jasonb]

Oh and your Case is to a redesign by JULIUS De Wall not Julian. A man known for shall we say “inventive” boiler design.

may help others find the design if they know the correct designer.

[Paul KempParticipant @paulkemp46892]

Lez,  I thought this was getting silly as this “thread” is about copper boilers and the YPM article first raised by the OP some 10 pages ago.  Jason beat me to it with the same impression it seems.  It also wasn’t clear to me from my quick read of your post we had jumped to <3 bar litres.  I did follow your link but it didn’t take me to part 2 of the code so I went to part 1 which doesn’t have a 3.18, maybe I should have realised then!  No problem moving the discussion to steel or stainless piddlers (less than 3 bar litres) but that would merit a new thread altogether in my opinion.

Without wishing to encourage this discussion in this thread, from my knowledge of inspectors in the club I am a member of I would speculate they would not be terribly receptive to testing your boiler in the circumstances you describe.  There are certain reservations among them on Vol 2 of the code and at the very least I imagine they would still require material certs and welder competence evidence for a stainless example as well as checking it was in compliance with a published design and requiring a pressure gauge to be fitted.  As you know club level inspectors (as commercial ones for that matter) are not obligated to test any boiler and can simply decline to conduct an inspection and test.  The “federation” are perfectly at liberty to comment on the possible under their code but they are unable to guarantee that if you present your boiler to a member club their inspector(s) will test it.

Paul.

[JasonBModerator @jasonb]

Paul, just to make it clear for anyone who may still be following along. Part 2 does not allow for any boilers with steel or stainless steel to be tested under that part of the code. So your inspectors would be right in not testing a small stainless boiler. Part 2 does allow brass though, can’t have those Mamods with out a cert!!.

I would think that Les’s boiler would fall in the 3 to 500 bar litre range so again if in stainless would not be accepted for TESTING under the CLUB code Part 1 either. Don’t know why he went looking at part 2.

 

[lezsmithParticipant @lezsmith]

Hi Paul & Jason,

FYI the Traction engine was not designed by Julius De Wall, he simply converted the original design by Rudy Kouhoupt to Metric, his plans are freely available at https://modelengineeringwebsite.com/Julian_TE_2.html. (please note their are some issues with the conversion I have sent a list of the issues to Julius).

Julius De Wall named the traction engine “Julian” hence I refer to is as the Julian traction engine.

Peter Squire (Sectary of FMES) has put me InTouch with their safety officer and Boiler inspector so I can get my UK regulations and insurance questions answered.

Jason is correct, my boiler is 6.15 Bar-litre, sorry my link was to part 1 by mistake, however as Jason points out the SS limitations are the same.

Both Paul and Jason, are correct about this thread being about Copper boilers, therefore we should drop any further references to Steel or Stainless Steel.

[JasonBModerator @jasonb]

Les, as I said “your Case is to a redesign by JULIUS” no need to inform me of it’s origin.

Also not just a simple conversion to metric. He doubled it up, and rounded the numbers to the nearest whole mm in many cases both rounding up and rounding down.

Yes lets leave this Traction engine to another thread if you wish to start one.

[lezsmithParticipant @lezsmith]

Hi Jason.
The URL was for other readers if they are following the thread, I remember you and shared emails about Julius 🙂

I did not realize he doubled it up; I do not have Rudy’s original design. Although the one in the museum looks very similar in size to what I’m building.

But I agree let’s leave that one for another thread.

[Phill SpowartParticipant @phillspowart84010]

Given the much publicised problems of AIs lying and talking rubbish, I’d be careful using them for, well, anything, but especially this. Bear in mind the training data is scraped off the internet, including internet forums, and we all know how much rubbish gets spouted on those…

(Grok in particular has had some quite alarming responses. Google at your own risk!)

[lezsmithParticipant @lezsmith]

Hi Phil,

I think you posted to the wrong forum, this is about copper boilers.

 

[JasonBModerator @jasonb]

Looks like the right thread to me and posted as a response to your earler post

[Phill SpowartParticipant @phillspowart84010]

Not sure what happened there, it was in reply to someone feeding queries about the Yield Point method into an AI earlier in this thread. I pressed reply to that post, don’t know why it doesn’t make that clear as published.

[Julie AnnParticipant @julieann]

Phill: When you press ‘Reply’ the system just appends the post to the end of the thread, not directly after the post on which ‘Reply’ was pressed.

However I had no trouble understanding where your reply fitted in, and that it was relevant. Not for me to say why Les missed the connection.

Julie

[Julie Ann]

On Julie Ann Said:

Phill: When you press ‘Reply’ the system just appends the post to the end of the thread, not directly after the post on which ‘Reply’ was pressed.

However I had no trouble understanding where your reply fitted in, and that it was relevant. Not for me to say why Les missed the connection.

Julie

If you want the bit you’re replying to use the ‘quote’ option, like what I did just now

[lezsmithParticipant @lezsmith]

Given that this has now turned into discrediting “AI”, let me just add.
Phill, basically you are saying:
“The internet lies, I know this is true because I read it on the Internet”.

As someone that has spent most of his life developing AI models for a USA federal agency, I can state “AI” can not lie unlike “HI”.

 

[duncan webster 1Participant @duncanwebster1]

If by lying we mean deliberately telling an untruth, then AI probably doesn’t lie, but there are plenty of reported instances of it making things up as it goes along. One particular instance an investigator made up some nonsense sayings and asked AI to explain, it came back with fanciful explanations rather than just saying ‘don’t be ridiculous’

 

[Charles LamontParticipant @charleslamont71117]

(Off topic)

Les, you are way behind on that one. Just one of a plethora of of credible and alarming reports:

https://www.psychologytoday.com/gb/blog/tech-happy-life/202505/the-great-ai-deception-has-already-begun

 

 

[MEinThailandParticipant @meinthailand]

I’ve done a lot of engineering work using the AI bot Copilot. It makes plenty of mistakes. Sometimes it cannot even do simple arithmetic.

Just one example, I used Copilot to develop Excel spreadsheets to calculate flat plate and stay thicknesses in copper model boilers. This demands an iterative solution which Copilot struggled with. I pointed out to Copilot that Copilot’s approach would generate circular references in Excel and indeed it did. Copilot’s response was to ask me to look for these circular references errors and then, notify Copilot then it would find a work-around, which was done.

But a human would have drawn a flowchart first before coding the spreadsheet. Copilot didn’t do that and was relying on it’s ‘memory’ of the equations to iron out the errors. But it couldn’t figure out the flow of the formulas it had given me. Copilot and I guess all AI bots can’t think.

Often, when I point out an error in Copilot’s responses it replied with “Grate Catch”. Ridiculous.

Yes, Copilot is a great tool that can be used to help solve engineering problems and is a great tool for model engineers to use. But you’ve got to know what you’re doing.

In effect you’ve got to have pretty good idea of what the answer should be before you start.

 

[Charles Lamont]

 

On Charles Lamont Said:

(Off topic)

Les, you are way behind on that one. Just one of a plethora of of credible and alarming reports:

https://www.psychologytoday.com/gb/blog/tech-happy-life/202505/the-great-ai-deception-has-already-begun

 

 

Hi Charles,

If you like and believe that one you should read some of his other blogs and books. e.g.
https://www.psychologytoday.com/us/blog/tech-happy-life/202203/why-its-good-that-you-arent-the-you-that-you-think-you-are

 

[cogdobblerParticipant @cogdobbler]

Many studies have found AI to have error rates between 60 per cent to 94 per cent (the latter for Grok). I would be reluctant to rely on  it for anything important.

[cogdobbler]

On cogdobbler Said:

Many studies have found AI to have error rates between 60 per cent to 94 per cent (the latter for Grok). I would be reluctant to rely on  it for anything important.

Sources, please.

[MEinThailand]

On MEinThailand Said:

On cogdobbler Said:

Many studies have found AI to have error rates between 60 per cent to 94 per cent (the latter for Grok). I would be reluctant to rely on  it for anything important.

Sources, please.

Just one among many: https://www.eweek.com/news/ai-chatbot-citation-problem/

And ironically, here is what AI has to say about the appalling accuracy of AI: https://www.google.com/search?q=studies+finding+high+error+rate+in+AI+Grok&client=ms-android-samsung-rvo1&sca_esv=bdc65ffe85f8e301&source=android-home&sxsrf=AE3TifOPLh9J7PSfPO-2TKVVBhUMNc0j1A%3A1754121504345&source=hp&ei=IMWNaOu5Etjl2roPzPD7qAo&oq=studies+finding+high+error+rate+in+AI+Grok&gs_lp=EhFtb2JpbGUtZ3dzLXdpei1ocCIqc3R1ZGllcyBmaW5kaW5nIGhpZ2ggZXJyb3IgcmF0ZSBpbiBBSSBHcm9rMgUQIRigAUiGqgJQtytYzpwCcAN4AJABAJgB5wOgAcNeqgEIMi0zNC44LjK4AQPIAQD4AQGYAi-gAopnqAIPwgIHECMYJxjqAsICDRAjGPAFGCcYyQIY6gLCAgsQABiABBiRAhiKBcICDhAuGIAEGLEDGNEDGMcBwgILEAAYgAQYsQMYgwHCAhQQLhiABBixAxjRAxiDARjHARiKBcICCxAuGIAEGLEDGIMBwgIKECMY8AUYJxjJAsICBBAjGCfCAgoQABiABBhDGIoFwgINEAAYgAQYsQMYQxiKBcICEBAuGIAEGNEDGEMYxwEYigXCAhMQLhiABBixAxjRAxhDGMcBGIoFwgINEAAYgAQYRhj5ARijBcICBRAAGIAEwgILEC4YgAQYxwEYrwHCAhAQABiABBhGGPkBGIwF2AEBwgIGEAAYFhgewgILEAAYgAQYhgMYigXCAggQABgFGA0YHsICBhAAGA0YHsICCBAAGBYYChgewgIFECEYnwXCAgQQIRgVwgIHECEYoAEYCpgDTfEFzKT_mssrJPe6BgQIARgTkgcKMy4wLjMyLjkuM6AHtNgBsgcIMi0zMi45LjO4B5lmwgcLMi01LjE2LjIyLjTIB6kI&sclient=mobile-gws-wiz-hp

[Michael GilliganParticipant @michaelgilligan61133]

This seems to be a concise summary [devoid of internet hysteria]

https://www.niemanlab.org/2025/03/ai-search-engines-fail-to-produce-accurate-citations-in-over-60-of-tests-according-to-new-tow-center-study/

MichaelG.

[JasonBModerator @jasonb]

Well if proof were needed that AI does not get things wrong, have a look at the latest GROK of S**t it has come up with posted on MEinT hialand’s forum.

How anyone can post something containing the statement “e.g., 6” diameter ≈ 0.34 liters” as a way to discredit Mark De Barr’s reply is hard to understand. Even if the 6″ diameter were used to work out an area without the missing length there is no way it could even approximately equal a volume. And you don’t need to be an engineer or have letters after your name to spot that GWOK has got it wrong 😉

Even AI itself s\ys the statement is incorrect, or should I not believe that either😂😂😂

[cogdobblerParticipant @cogdobbler]

1. 

[JasonBModerator @jasonb]

Another classic mistake in the latest GROK

This is another incorrect statement from GROK

 

“BS EN 12953-1 applies to boilers >2 liters, pressures >0.5 bar, and temperatures >110°C, making it irrelevant for model boilers “

 

Why is it not applicable?

Model boilers are often greater than 2lts capacity. Model boilers are almost certainly going to be over 0.5 bar and the steam temperature likely to be over 110degC.

If the two gentlemen can’t even read the Grok Review and see that there are errors then what is the point of them going on about how qualified they are if they can’t even spot howlers like that but blindly post what comes up in the belief that it is correct.

Though Mark de Barr does give BS 12953 as an example of Industrial Boiler Design so Grok should not really be picking up whether it is applicable to model boilers or not.

However, Grok’s conclusion does raise the question of what Grok defines as a “model” boiler ? Maybe as it thinks the boiler will be so small and have less than 7psi applied it is thinking of non functioning models such as a boiler on an electrically powered “toy” steam engine’s boiler eg. Hornby OO model railways.

[Author]

Posts