Steam Engine Design

[Martin Johnson 1Participant @martinjohnson1]

Hello all,

Back in the '80s Jim Ewins published some proposals for a set of engine factors to help in designing rail locomotives. I have been looking at how these apply to road locomotives and in the course of that have been doing some research of my own, starting from Jim's articles in Engineering in Miniature in 1985.

Does anybody know of discussions, updates, adherents, opponents or any work that has been done in that vein in the intervening 30 years?

I am fine with the Engine factor Ee. Also accept the tube factor Kt. But from a theoretical viewpoint, I have big problems with the Boiler factor Eb and as a consequence the Overall factor Eo.

Many thanks,

Martin

[Martin Johnson 1Participant @martinjohnson1]

Plea for help

[AjohnwParticipant @ajohnw51620]

There are some older books about on engine "physics" in general. You might find some on the internet archive.

One I read some where or the other offered a mathematical proof that an external combustion engine of any type can't exceed 50% efficiency in a theoretical sense which probably means they don't even achieve that in practice. It also covered internal combustion engines of various types.

No doubt written by an academician who are unfortunately inclined to prove what they want to prove.

John

–

[julian atkinsParticipant @julianatkins58923]

hi martin,

john baguley has produced Jim Ewins' formlae on a spreadsheet.

you have to be a bit careful with Jim's formulae. it was primarily used by Jim to show that the old Greenly boiler and cylinder proportions were wrong. however, the formulae doesnt work in lots of cases. an example is Don Young's No.1 Railmotor boiler which shows up very badly on Jim's formulae, yet it is one of the best free steaming boilers i have encountered. the formulae also doesnt work with some narrow gauge wide but short fireboxes such as the Hunslet Quarry type, despite again Don Young's Hunslet boiler design also being another very free steaming excellent boiler design.

as road locomotive boilers are often different in proportion to miniature steam locos i am not surprised that Jim's formulae doesnt work.

the C.M Keiller formula for tube diameter to length is widely accepted as being essential, and is used by Jim. Don Young was very keen on a free gas flow as a percentage of grate size. Jim found by extensive experiments that most heat was transferred in the firebox, and that little if any heat was transferred via the final 2/3rds of the tube length.

there are of course lots of other factors and design considerations.

cheers,

julian

[Ady1Participant @ady1]

Jim Ewens was quoted by Martin Evans quite a few times and did a build series in the early editions of EIM

His work doesn't appear to be consolidated anywhere that I can find and I also have a 1966 reference to boilers

5 October 1979 ME 3618

Train Resistance
In the August issue of Railway Magazine there was a
very interesting article by O. S. Nock, doyen of rail-
way enthusiasts, on Train Resistance and Horse-
power. It seems to me that I have seen very little
indeed on this subject as applied to our small
locomotives and trains in Model Engineer; all I can
recall is the measurements of performance done on
the S.M. & E.E. test bed some 25 years ago and a
report by Jim Ewins on Model Boiler Performance in
1966 which also went into some detail about horse-
power. .

 

Edited By Ady1 on 01/01/2016 11:44:01

[Neil WyattModerator @neilwyatt]

For anyone interested in the maths of steam engine efficiency see:

en.wikipedia.org/wiki/Carnot_cycle

Bear in mind that this only sets an 'upper limit' on engine efficiency and says nothing about boiler efficiency, just that the maximum theoretical efficiency depends on the difference between the input and output temperatures. This is intuitive – for a steam engine where the drop in steam temperature clearly relates to the energy taken out of it.

It also explains the basic reason why superheating (that raises the input temperature) and condensors (that lower it) both raise efficiency.

Neil

[duncan webster 1Participant @duncanwebster1]

I'm glad someone else has difficulty with Ewins boiler factor, at least it's not just me. However, I have to take issue with Neil on superheating. Most of the heat transfer in a boiler is used up evaporating the water, the heat transferred at higher temperature to superheat the steam is quite small, and has has little effect on potential efficiency. The real reason why superheat improves efficiency is that it massively reduces condensation. Once you have enough superheat to stop condensation there is little thermodynamic gain from more. This all depends on what cut off you are aiming for, shorter cut off gives more temperature drop to the steam, so you need to start with more superheat. Thus with lower boiler pressure on our models than full size, we run at longer cutoff, so less superheat needed.

Compounding reduced the pressure drop, and hence temperature drop, in each cylinder, so reduced condensation. Once superheat came into vogue, compounding of locomotives went out. Even if it gives better indicated efficiency, all that extra gubbins whirling round absorbs extra power.

Edited By duncan webster on 01/01/2016 13:03:06

[JasonBModerator @jasonb]

If you are looking into Road steam then have a read through some of Julia Old's and Ross Bishop's (Suctionhose) threads over on Traction Talk Forum. Likely to be more relevant and based on higher pressures due to steel boilers rather than lower pressure copper ones usually found on rail locos.

[John BaguleyParticipant @johnbaguley78655]

Hello Martin,

As Julian says, I made a study of the Jim Ewins formulae and converted them to easy to use spreadsheets.

**LINK**

Also, as Julian says, the formulae are not the be all and end all and they don't 'fit' a lot of published designs, not that that itself means the formulae are wrong. It could well be the designs that are wrong. Jim really just came up with his formulae to try and establish design criteria based on some published designs, some of which worked well, and some not so well.

After much thought on the subject I think the critical areas to concentrate on are the engine factor Ee and the free gas area of the tubes versus the area of the grate. Personally, I now don't think the Keiller factor Kt is of any importance in 'our' sizes. It's pretty much accepted that only the first few inches of the tubes contribute much to the heating effect as the flow is laminar rather than turbulent in such small tubes so the length is pretty immaterial within reason. That also means the boiler factor Eb is not really valid and I think can be ignored.

So, my latest thoughts are that you should design the grate area to match the cylinders (steam consumption) and then try and get enough tubes in the boiler to match that grate area i.e. 12 – 15%. Also, fit a properly designed valve gear to make the best use of the steam produced. Most published designs have pretty poor valve gears. Oh, and use as much superheat as you can get to increase efficiency!

John

[Martin Johnson 1Participant @martinjohnson1]

Thanks to all who took the trouble to reply.

The basics of Carnot cycle efficiency limits were drummed into my alcohol fuelled student brain some 40 years ago, but they are of scant help when optimising boiler flue layouts.

Ady1, thanks for the reference to ME 3618 – I shall look it up.

Julian and John, you need to get together and come to an agreement on the Kt factor! It seems to give a figure that will use most of the heat in the flue, but I think it is very empirical.

As I said, I have no problems with the Ee factor, which is all about balancing heat input to steam demand from the cylinders. One term that it does not include is engine speed, presumably because most 3.5 / 5 / 7.25” railways trundle round at approximately the same speed for safety. The other term it really needs is “Grate Loading” – i.e. how many kg/hour/m2 of coal can you put through a grate before you get clinker. I am not convinced about consequences if you take it too low. Grate Loading must be a relevant design parameter that is a function of the chosen fuel – clearly it would not be a limiting factor with gas firing, for example. I have expanded the formula to take account of these parameters for road engine use and it is quite a simple theoretical derivation.

Duncan, I am glad I am not the only one who has problems with Eb (I also agree that superheating is a grossly misunderstood science). Julian, your comments about the rail motor are exactly pertinent to the flaw in the Eb factor – as well as it not being dimensionless. John, I have come to the same conclusion as you that the critical boiler factor is to get the ratio of tube bank area to grate area about right. I can see that a short tube stack might lose some efficiency up the chimney, but can see no other consequence, assuming blast nozzles are adjusted to match the reduced flow resistance.

As a result of all this, I am working on a program to calculate the heat transfer and flow resistance in a bank of tubes from first principles. I am using the summary of heat exchange correlations as reported here:

**LINK**

So far, I have proven that flow in model (and many full size) tube banks will be laminar, and have got the kernel of the spreadsheet working fine. I hope to either prove or disprove some of the assertions about “all the heat transfer taking place in the first n% of length”. Test results in my 1929 Locomotive Engineer’s Pocket book tend to disprove the assertion, and I hope to calibrate the mathematical model against these.

JasonB, I am already in contact with Julia and SuctionHose on TT forum. My alter ego there is tenor. I have a build thread there for the project that is prompting all this at:

**LINK**

or at:

**LINK**

Not surprisingly, this is a stumpy little boiler with a very short tube stack – but how many of what size? And if they are very short, can I use a smokebox superheater coil to soak up some of the waste heat (as Fowlers did) or should I go to flue superheaters? Time and a lot of sums will tell.

I was working on an article about the Fowler design and build which prompted this line of thought. Then when you start to marshal your thoughts, you find they are not very clear in the first place. Hence the plea for any other work. Seems I might now have two articles to write………………

It all keeps the grey matter going.

Martin

[Anonymous]

Posted by Martin Johnson 1 on 02/01/2016 10:00:16:

JasonB, I am already in contact with Julia and SuctionHose on TT forum. My alter ego there is tenor.

I thought that there was a certain familiar "tenor" to the original post……………

Andrew (oly2brf5 on TT)

[Martin Johnson 1Participant @martinjohnson1]

Just a brief update. I still don't believe in the Ewins boiler factor. Nor, having investigated it's roots, do I believe in the Keiller tube factor.

What I have done is to produce a spreadsheet that calculates from first principles the heat transfer in a boiler.

The firebox segment uses the well known Stefan Boltzman radiant heat transfer equation.

The tube bank section uses accepted correlations of convective heat transfer for the laminar or turbulent flow regimes as appropriate. For a summary of these see this website or degree standard heat transfer texts:

http://web2.clarkson.edu/projects/subramanian/ch302/notes/Convective%20Heat%20Transfer%201.pdf

The whole lot needs a correlation of flue gas properties (from my steam tables) against temperature to provide constants for the equations.

The final key to the problem is knowing the mass flow through the boiler, which has prompted a review of a lot of IMLEC results, amongst others. This seems to be a key aspect that nobody has picked up on before!

I have achieved a good correlation with test results from full size locomotives, but test results from models are as rare as hen's teeth. I am particularly pleased with the agreement of heat transfer Vs. tube length to full size tests.

The program also calculates pressure drop of the flue gas through the boiler (quite easy as part of the heat transfer calcs.). It seems that flow resistance in the tube bank is a relatively small part of the whole – entry and exit losses dominating – so that the basis for Ewins' formulation including tube length as a factor in pressure drop has no basis in fact.

I have confirmed that tube bank flow in "full size" practice is normally turbulent, while in 5" gauge sizes the flow is laminar. Once we get to larger boilers (such as my 4" Burrell TE) it all rather depends.

I am intending to develop the program to look at the whole ashpan to chimney circuit, including blast pipe performance. The reasoning is that any fool can make a good heat exchanger with enough pressure drop. The trick is to design one that does not need lead to a performance loss in the engine by needing an excessively small blast pipe orifice. There is a balance to be struck, but it is not as simple as Ewins suggested.

Hope to write it all up for one of the comics in due course.

Martin

[Anonymous]

Martin: Very interesting post; it's great to see some real engineering analysis for a change. I look forward to seeing the results in print.

I'm assuming that the heat transfer is better with turbulent flow? Certainly for heatsinks in electronics the heat transfer is much better with turbulent flow. Of course the heat flow is the other way round with a heatsink, but I'm not sure that is significant.

Presumably for smaller boilers, with smaller tubes, the Reynolds number is too low for turbulent flow? Is there any merit in considering a rougher surface in the tube to try and trigger turbulent flow?

On sailplane wings the holy grail is a laminar flow boundary layer across the whole wing chord. In reality this is impractical, and it has been found better to trigger turbulent flow at some point rather than let the boundary layer transition to turbulent flow when it feels like it. To trigger the turbulent boundary layer some sailplanes have plastic zig-zag tape, about 1mm thick, on the top surface of the wing at about 75% of chord. In some ways it's a bloody nuisance as it is very sharp and leads to cut hands when rigging. Strangely some owners get fussy when you bleed over their sailplane!

Andrew

[Martin Johnson 1Participant @martinjohnson1]

Hi Andrew,

Yes, the turbulent heat transfer is better due to the thinner boundary layer. Back in the day, there were ridged flue tubes available for that very reason of tripping up the boundary layer; whether or not it would be practical in miniature, I have severe doubts – it never really caught on in full size.

Fully understand you cutting yourself on plastic boundary layer trippers. Wifey used to work in a wind tunnel (very fast military stuff), and BL trippers in miniature are little sticky up razors – used to cut her hands to ribbons.

Also thanks to Duncan Webster who has P.M'd me with some useful leads.

My present task is trying to set up a calculation routine for evaluating superheaters (both smokebox coil and flue type). The numeric complexity is "challenging".

Martin

[Anonymous]

Posted by Martin Johnson 1 on 31/01/2016 20:30:42:

Yes, the turbulent heat transfer is better due to the thinner boundary layer. Back in the day, there were ridged flue tubes available for that very reason of tripping up the boundary layer; whether or not it would be practical in miniature, I have severe doubts – it never really caught on in full size.

Fully understand you cutting yourself on plastic boundary layer trippers. Wifey used to work in a wind tunnel (very fast military stuff), and BL trippers in miniature are little sticky up razors – used to cut her hands to ribbons.

There can't be many places with those sort of wind tunnel facilities. I'd hazard a wild guess at the 3ft or 8ft supersonic wind tunnels at RAE Bedford?

It sounds like the idea of ridged tubes to trip the boundary layer to turbulent flow is like my idea (mooted on TT) of a plate in the steam take off to increase the steam dryness. A nice idea but didn't really provide any substantial benefit in practise. I was thinking more in terms of something like sandpaper at the start of the flue tube to trip the boundary layer. One of the problems with the early laminar flow aerofoils was that they were very sensitive to disturbances on the surface. So dead bugs or rain would be enough to destroy the laminar flow, and the glide performance. At one time there was a craze for bug wipers that could be wound out along the wing to remove said bugs and then wound back into storage.

Andrew

[Author]

Posts