If you register, you can do a lot more. And become an active part of our growing community. You'll have access to hidden forums, and enjoy the ability of replying and starting conversations.

Boiler design & construction ex Patriot thread.

Discussion in 'Locomotive M.I.C.' started by Steve, Jan 22, 2016.

  1. W.Williams

    W.Williams Well-Known Member

    Joined:
    Dec 20, 2015
    Messages:
    1,582
    Likes Received:
    1,439
    Occupation:
    Mechanical Engineer
    Location:
    Aberdeenshire
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    Iv made a start on some hand calcs. Copper does have a much higher expansion coefficient, so I may look first at steel/steel design. Also helps keep costs down.

    What kind of temperature would one expect to see on the crown and on the immediate water interface around the box? 150C...???

    What I'm noticing already is that what you have going on in the boiler is a battle of hydrostatic v thermal deflections. in some places they cancel each other out and work favorably together, in others they double up and create issues. I think this is where the clever aspects of FEA can be brought to bear.
     
  2. Steve

    Steve Resident of Nat Pres Friend

    Joined:
    Oct 7, 2006
    Messages:
    11,910
    Likes Received:
    10,061
    Occupation:
    Gentleman of leisure, nowadays
    Location:
    Near Leeds
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    I don't have any data for boiler plate temperatures so I can't give any help here. Better men than me could probably produce a theoretical calculation based on a firebox gas temperature of 1200 degC water temperature at saturation temperature for the pressure and outer temperature not much different, due to being lagged (but only above frame level.)
     
  3. 8126

    8126 Member

    Joined:
    Mar 17, 2014
    Messages:
    823
    Likes Received:
    962
    Gender:
    Male
    I don't presume to be a better man than @Steve , because I had a bit of a think about the dynamics of gas flow and conductive/convective heat transfer in a firebox, and part of my brain ran away screaming. However, I did wonder how much you'd get with radiant heat transfer, and disappeared up a complete blind alley where the assumptions gave me a 40hp Terrier tank. They may be LBSC, but they aren't that bad. :D

    So instead I decided to start with the power output and go the other way, with some numbers for the SAR class 26 (Phil Girdlestone's book Camels and Cadillacs, about the SAR Class 25 family, has some very good stuff on the problems of designing big boilers). It's all going to be rounded in metric dimensions. Here are our basic figures for the Class 26:

    Max drawbar power =4000hp, approx 3000kW,
    If we assume the cylinders are 10% thermally efficient (about right I think), boiler power needs to be 30000kW. I seem to remember that boilers were usually over 90% efficient, so I've ignored that. I've made a completely unfounded guess that about a third of the boiler power proportionately is transferred in the firebox, so...
    Power transfer in firebox Q = 10000kW = 1 x 10^8 W
    Total firebox heating surface A = 30m^2
    Wrapper thickness guesstimate s = 1 x 10^-2 m
    Coefficient of thermal conductivity for steel k = 43 W/(m.K)

    Our equation for the necessary temperature difference to get a given rate of heat transfer over the heating surface area, given the wrapper material and thickness is:

    dT = Qs/(kA)

    Therefore:
    dT = 77.5 K

    Assuming the water to be an infinite heat sink at 225psi, the boiling temperature is 200 degrees C, so our temperature on the inside of the wrapper is 277.5 degrees C. Call it 300, because it certainly deserves no more precision than that. It's just an average over the full firebox based on shaky assumptions and the numbers can be tweaked if any of my assumptions are grossly out.
     
    Martin Perry and W.Williams like this.
  4. Steve

    Steve Resident of Nat Pres Friend

    Joined:
    Oct 7, 2006
    Messages:
    11,910
    Likes Received:
    10,061
    Occupation:
    Gentleman of leisure, nowadays
    Location:
    Near Leeds
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    In the absence of anything else, this is the best we've got and is probably not too far off the mark. The firebox plates aren't getting into the dull red spectrum so are certainly less than 400 deg C, even on the fire side. If the firebox plates are made too thick there is a tendency for them to burn away because they overheat and plate thickness is something of a compromise. In the UK we are used to relatively thick copper and tend to go for 1/2" plates with steel whereas non-UK practice tends to favour 3/8" or thereabouts. A limiting factor is the strength of screw threads in thin plate but this shouldn't be a problem with welded stays.
     
    Last edited: Jan 28, 2016
  5. 8126

    8126 Member

    Joined:
    Mar 17, 2014
    Messages:
    823
    Likes Received:
    962
    Gender:
    Male
    What I didn't mention was the effect of copper's significantly better thermal conductivity. For pure copper k = 400 W/(m.K), so near ten times better than steel, although I would guess that's ultra pure stuff and even some trace impurities will knock it back; it doesn't take much alloying to hit the electrical conductivity. Even aluminium with k=200 is five times better than steel (while obviously completely unsuited to this application). Therefore with a copper box you might expect the temperature difference to be more like 10 degrees C, although the trade off with poorer strength at temperature and greater thermal expansion may mean the difference is less significant than it looks. It certainly won't affect the rate of energy transfer through the wrapper very much, when you consider the gas temperature and significant radiant heat transfer.

    Heavy scale deposits on the water side probably put the wrapper temperature up by more than any effect of the heat transfer across it.
     
    Steve likes this.
  6. W.Williams

    W.Williams Well-Known Member

    Joined:
    Dec 20, 2015
    Messages:
    1,582
    Likes Received:
    1,439
    Occupation:
    Mechanical Engineer
    Location:
    Aberdeenshire
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    Say we have 1200C on the inside and 300C on the outside of the wrapper and we assume that creates for the purpose of calculation a mean temperature of 900C, then I am getting 21.1mm, just under an inch for a 2m length plate, unconstrained of course. However, this dl all has to be taken up somewhere. Bending and compression in the stays. Steve, pass the oar please, I need to do some back-rowing...

    I admit I didn't think strains would be that high...it gets even worse for copper at 28.2mm

    dl = L0 α (t1- t0)

    where

    dl = change in length (m, inches)

    L0= initial length (m, inches) 2m here

    α = linear expansion coefficient (m/moC, in/inoF) 12x10^-6 for steel and 16x10^-6 for copper

    t0= initial temperature(oC, oF) 20degC

    t1= final temperature(oC, oF) 900degC

    It just struck me how much havoc this must play with the stresses in the winter during periods of prolonged cold sitting...These machines really are best kept, "on the boil" at all times, as the reverse cycling just exacerbates the issues.


    Yes, I guess this is why washouts are so important. I have been wondering if the use of copper was something that was more prolific to achieve what may be marginal gains at a time when the stuff wasn't nearly as expensive as it is now.
     
    Last edited: Jan 28, 2016
    Steve and Martin Perry like this.
  7. JJG Koopmans

    JJG Koopmans Member

    Joined:
    Nov 12, 2014
    Messages:
    382
    Likes Received:
    474
    Gender:
    Male
    Heritage Railway Volunteer:
    No I do not currently volunteer
    Sorry to interrupt, but it makes little sense to me to re-invent these data. Both Phillipson and Johnson treat these questions in their locomotive design books. The American Lawford Fry published a lot of data and was one of the most serious calculators of his days. In the U.K. Diamond did some work. So these boiler data need some homework and publication reading!
    Kind regards
    Jos Koopmans
     
    fergusmacg and Sheff like this.
  8. Steve

    Steve Resident of Nat Pres Friend

    Joined:
    Oct 7, 2006
    Messages:
    11,910
    Likes Received:
    10,061
    Occupation:
    Gentleman of leisure, nowadays
    Location:
    Near Leeds
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    I can't comment on the other sources but I am unaware of Phillipson considering the question of plate temperatures and expansion in any detail.
     
  9. W.Williams

    W.Williams Well-Known Member

    Joined:
    Dec 20, 2015
    Messages:
    1,582
    Likes Received:
    1,439
    Occupation:
    Mechanical Engineer
    Location:
    Aberdeenshire
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    May one please be a touch more specific on titles as I could find only one of those on Amazon.

    Incidentally, http://www.steamindex.com/jile/jile.htm cites IMechE as having a lot of info in the virtual Library. Time to put this professional registration bit to good use...
     
  10. Steve

    Steve Resident of Nat Pres Friend

    Joined:
    Oct 7, 2006
    Messages:
    11,910
    Likes Received:
    10,061
    Occupation:
    Gentleman of leisure, nowadays
    Location:
    Near Leeds
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    'Steam Locomotive Design data & Formulae' by E A Phillipson. My copy pre-dates ISBN but is available as a reprint from Camden
    http://www.camdenmin.co.uk/collecti...team-locomotive-design-data-and-formulae-1936.
    A bargain at £12.95 when my copy was originally £3 in 1936. One of my bibles.

    You might also find 'Locomotive boilers for the 21st Century' by Alan Haigh to be of interest.
    http://www.camdenmin.co.uk/collecti...comotive-boilers-for-the-twenty-first-century
     
  11. K14

    K14 Member

    Joined:
    Nov 6, 2011
    Messages:
    403
    Likes Received:
    298
    Location:
    Catford
    Heritage Railway Volunteer:
    No I do not currently volunteer
    8126 likes this.
  12. W.Williams

    W.Williams Well-Known Member

    Joined:
    Dec 20, 2015
    Messages:
    1,582
    Likes Received:
    1,439
    Occupation:
    Mechanical Engineer
    Location:
    Aberdeenshire
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    Thanks gents. That adds some more wight to 2016's reading list. :)
     
  13. JJG Koopmans

    JJG Koopmans Member

    Joined:
    Nov 12, 2014
    Messages:
    382
    Likes Received:
    474
    Gender:
    Male
    Heritage Railway Volunteer:
    No I do not currently volunteer
    True, but I was commenting on the earlier question of the temperatures in general. As for the plate temperatures, I was taught 50 years ago that this is linear between the outside and inside temperature. Expansion is something that FEM will supply.
    I am wondering whether it would be an idea to look for a boiler that has been tested and calculated in the past and perform a
    FEM analysis for comparison.
    Kind regards
    Jos Koopmans
     
  14. W.Williams

    W.Williams Well-Known Member

    Joined:
    Dec 20, 2015
    Messages:
    1,582
    Likes Received:
    1,439
    Occupation:
    Mechanical Engineer
    Location:
    Aberdeenshire
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    Do you mean to attain some strain numbers from a physical test to compare with the finite element analysis?
     
  15. JJG Koopmans

    JJG Koopmans Member

    Joined:
    Nov 12, 2014
    Messages:
    382
    Likes Received:
    474
    Gender:
    Male
    Heritage Railway Volunteer:
    No I do not currently volunteer
    Yes, FEM is a well proven method but it needs convincing evidence especially in the field of locomotive boilers.
    Kind regards
    Jos Koopmans
     
  16. W.Williams

    W.Williams Well-Known Member

    Joined:
    Dec 20, 2015
    Messages:
    1,582
    Likes Received:
    1,439
    Occupation:
    Mechanical Engineer
    Location:
    Aberdeenshire
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    Indeed. One approach will be to take cold and hot dimensions for a boiler undergoing a steam test and start from there. Get some strain gauge rosettes on the go and see what kind of numbers the FEA should be hitting.

    The expecations is of course that fully welded boilers for the same overall plate thickness dims are going to expand and contract more due to the reduced rigidity from lack of doubler plates and overlaps.
     
  17. 8126

    8126 Member

    Joined:
    Mar 17, 2014
    Messages:
    823
    Likes Received:
    962
    Gender:
    Male
    I haven't covered the recommended homework yet, but I'm fairly sure your assumption of a 900 deg C average is flawed, because, as Steve pointed out, the plates don't get red hot and these certainly would be. They would also have undergone a phase transformation to austenite and have lost most of their useful tensile strength.
     
  18. W.Williams

    W.Williams Well-Known Member

    Joined:
    Dec 20, 2015
    Messages:
    1,582
    Likes Received:
    1,439
    Occupation:
    Mechanical Engineer
    Location:
    Aberdeenshire
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    I agree. As above, I'm going to go through some more literature before making a start on some serious numbers.
     
  19. estwdjhn

    estwdjhn Member

    Joined:
    Jan 24, 2013
    Messages:
    324
    Likes Received:
    637
    Occupation:
    Boilermaker
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    I'm fairly sure that the design specs for some of the new build stuff I've worked on (all fairly small - narrow gauge or traction engine size) had some hard numbers on the maximum working temperature as part of their PED submissions.
    I've a feeling something like 330deg c was the relevant number, but can't recall what it was exactly.

    Most boilers are massively over-engineered in most areas when new/freshly overhauled - this gives a half decent working life before they need heavy repairs. It is probably as well that this is the case - over time I've opened up a good few boilers either in still "in ticket" or just out of ticket to find things that give one the screaming heebie jeebies - epic levels of plate wastage, headless rivets, literally rows of broken stays... if they weren't so over-engineered, bad things would happen much sooner.
     
  20. 8126

    8126 Member

    Joined:
    Mar 17, 2014
    Messages:
    823
    Likes Received:
    962
    Gender:
    Male
    Thanks, that's interesting and suggests that my finger in the air estimate of 300 deg C wasn't too bad. What sort of working pressures are these boilers at?

    I must admit that if I heard the locos on any given railway were frequently turning out to have whole rows of broken stays I might give it a wide berth, I thought routine inspections with a hammer were supposed to catch that sort of thing before it got out of hand.
     

Share This Page