Locomotive Performance and Tractive Effort Discussion

Nice picture of the Chapelon Pacific in Mulhouse Museum!

The rebuilding history of the Paris-Orleans Pacifics was more complex than you suggest above, with much larger numbers involved. The attached link gives a useful summary of the French Pacific types, outlining initial construction, subsequent rebuildings and inter-regional transfers.

https://en.wikipedia.org/wiki/List_of_French_'Pacific'_steam_locomotives

As well as the rebuilding of the Paris-Orleans 3500-class and 4500-class Pacifics to Chapelon's designs, many of the Pacific types on the other French railways, including the PLM, were also modernised to incorporate Chapelon ideas. But a point of note is that the French Railways were winding-down the production of new passenger engines (both express and suburban) in the 1930s, so the new batch of 28 that you mention (built for the Nord to the Chapelon design) were among the very few new-build Chapelon types, alongside several hundred rebuilds that he either designed or influenced. After the formation of SNCF in 1938, the lack of passenger engine new-build became even more pronounced, although large numbers of mixed-traffic 2-8-2s were constructed.

https://en.wikipedia.org/wiki/List_of_SNCF_classes#SNCF

I fully agree with you that Chapelon and Gresley cannot be directly compared, as their roles were so different. However, some of Chapelon's designs were not just prototypes, but were built on a substantial scale and saw widespread service for many years. This applies particularly to the 1930s rebuilds of the Paris-Orleans Pacifics, which were directly his work (Although those designs had to be endorsed by the CME of the Paris-Orleans Railway, who under more usual circumstances would have been given the credit). Chapelon went on to have a global influence, which cannot be measured, but was clearly visible in the rebuildings on the other French Railways and in the subsequent SNCF 141P and 241P designs.

 

One of the most beautiful looking locomotives I know about!

 

Thanks a lot for all your information and pictures about Chapelon`s masterpieces RAB3L. Much appreciated!

Knut

 

I'm not aware of any locomotive that is fitted with a Lemprex exhaust.
There is a sketch of a Lemprex exhaust in the "paper" (package of suggestions) that Porta sent to the A1 group in the early design stage. It was a very complex design, including reservoirs to smooth out the pulsing flows. I think I can follow what Porta was thinking, but to put it politely, I think he might have had problems with it. I certainly can't see it having any advantages over any other established design.

Porta did some work developing the 2'6" gauge Mitsubishi 2-10-2 on the Rio Turbio railway in the far south of Argentine.
The height of the exhaust system is an important dimension, and the locomotive had the luxury of some space within the loading gauge for a reasonably tall chimney above the top of the smokebox.
One of Porta's earlier designs was for a "Kylpor". There is not much written down about these but photographs seem to indicate that it looks like a Kylchap within the smokebox, but with a tall conical chimney above the smokebox (which would improve the smokebox vacuum). I understand that these were reasonably successful.

Porta then moved on to the "Lempor". These are well documented with Porta producing a paper describing how to design one for any particular locomotive. One of the key features of the Lempor is "de Laval" convergent-divergent nozzles. In his book "Red Devil" David Wardale explains the operation of these, probably better than I can, but during the "chuffs" the nozzle generates supersonic flow, which does the majority of the draughting work, and between the chuffs the flow becomes subsonic, and present a lower back-pressure to the pistons. Overall this can be designed to reduce the back pressure on the pistons, for the same draughting work, compared with plain nozzles.

The "chuff" occurs when the piston is close to the dead-centre position, and back pressure is less important.
The Lempor also has a long mixing chamber and conical diffuser.

There are many designs of industrial vacuum ejectors which look very similar to Lempor.

There are also some locomotives that have been fitted with convergent-divergent nozzles, but with conventional chimneys, and these seem to have been reasonably successful. One of the more notable of these is the "Niagara" 4-8-4 on the New York Central, which has the Selkirk front end, and is possibly where Porta got the idea.

One of my concerns about the Lempor is that it has a fairly small opening foe the combustion gases to enter the exhaust.
I can't really see any fundamental reason why Lempor convergent-divergent nozzles couldn't be used in conjunction with Kylchap cowls, to overcome this perceived limitation.
Then, if you are really keen and have the space available you could have a Kylpor with Lempor nozzles.

 

Nigel Day has had experience designing exhaust systems. He used the Lempor on the WLLR locomotives and others and I believe developed further on based on his experiences - he might well be in possession of more information concerning the Lemprex type. I would suggest having a quick read through of the sections devoted to the exhaust systems fitted to 19D 2644 and to 3450 to be found in The Red Devil. Then have a look at the drawing to be found on page 152 which is an SAR drawing of the Double Lempor as designed for 3450 but also includes modifications made while "tuning-up". The nozzles sit just at the entrance to the bellmouth and this is a double ejector, you might perceive the area for the combustion gasses to enter the exhaust to be "too small" but this is not the case. The cowls are from earlier thinking but the Lempor does not need them and if you were wanting to use them where would you install them bearing in mind the importance of the mixing chamber, diffuser and system height?

 

The Lempor undoubtably does work. But in my opinion, an exhaust system based on Lempor nozzles and Kylchap cowls may work better. Someone who knows more about these things than I do told me that the bottom half of a Lempor is easier to make (than a Kylchap), and the top half of a Kylchap is easier to make (than a Lempor). So assuming you can get the proportions right, what's not to like?

 

In the above it is stated that supersonic exhaust flow always exists. This is only true in the case of the Red Devil which was redesigned for higher outputs and needed
a high exhaust pressure also for the GPSS steam under the firebox. Sofar I have never seen any proof of supersonic flow measurements.
As for Porta, he designed the Lempor parallel to the theory as established in modern textbooks on fluid dynamics, so, like in modern science, his approach is outdated.
As for cowls, already Prof. Goss found around 1900 that they dit not work as hoped for. Imho plain nozzles and a long diffuser chimney give the best and low-cost results.
Kind regards,
Jos

 

The "Selkirk" tests on the New York Central showed that the supersonic flow starts with a pressure at the blastpipe tip of around 11 psig. I agree that supersonic flow doesn't "always" exist, but the argument for the convergent-divergent nozzles is that the flow goes supersonic just during the initial exhaust release.
I would very much like to see some measurements using modern transducers and data recording to see what is actually happening in terms of mean and peak pressure.
For the Kylchap cowls, I have two theories, one is that they reduce turbulance in the smokebox, and the other is that they reduce the effect of the jets in close proximity to one another merging together. I saw a calculation (I forget where) that the momentum of the combustion gases entering a Lempor deflect the steam jets towards the axis by 11 degrees.

 

Were the "Selkirk" tests not with a high-powered locomotive with conventional exhaust?
A changeover to a modern multiple exhaust for the same draft would quite probably not generate supersonic flow due to the larger exhaust area.
Kind regards
Jos

 

The Selkirk Front end was applied to the 4-8-4 Niagara on the New York Central. This is a conventional single chimney, but with a divergent nozzle. There is a 10.25" gauge scale model on the Stapleford Miniature Railway, and the exhaust system has been faithfully modelled from the original.

 

Tragic that none of the real Niagaras were preserved (but that`s another disqussion of course)

 

Most if not all members of the forum will be aware of the phenomenon of spark arrestors and their impact on the steaming of a locomotive. These devices frequently had a negative effect on the steaming of a locomotive to the point that shed staff/crews removed them in order that the work expected of the engines could be effectively carried out. They increased the boiler gas flow resistance to the point where the exhaust system was incapable of providing adequate vacuum to, shall we say, drive the boiler.
Draught losses occur through the ashpan, firegrate and fire, they occur through the firebox, through the tubes and through the smokebox. We are aware of rail tours getting into difficulties and having delays because of blocked or obstructed spark arrestors.
When Class 25NC 3450 was rebuilt into what became Class 26 it was understood that the application of GPCS would increase draught loss through the firebox and this loss turned out to be far more than predicted by the design calculations but at least they tried to predict it. The system used/uses a lower primary air flow but the draught loss through grate and firebed was found to be essentially double that to be found on the 25NC brought about by the restricted grate air area and the use of a thicker fire. The energy required for the secondary air streams and the turbulent mixing of these high velocity streams with the rising gas from the firebed was an even larger factor. There were also increased losses through the tubes, though this was a small factor, against this they did manage to reduce loss through the spark arrestor.
The exhaust system has to efficiently provide enough energy to overcome these losses. It should also have reserve capacity to overcome vacuum leaks (smokebox doors don't always fit and smokeboxes can have holes in them).
Do you need a high power output locomotive? If you do you can simply keep on building larger and larger engines with poor power to weight ratios until you reach the limits of your infrastructure. But what then? Use two locomotives or three or even four or more? That is a great deal of metal, machinery and manpower not being put to good use. I am sure that your customers will appreciate the costs, as for the shareholders .......
You need a high power to weight ratio and having high exhaust pressure should not be a part of the equation. Neither should poor combustion system design, one which can launch 50% of the fuel fed to it uncombusted into the air. Losses through leakage are another matter.
The inclusion of cowls in the Kylchap system is something which has caused debate for years but the system has been in use for many years, works well and is a part of our heritage. If they offer a slight improvement in some areas that is all well and good but newer systems do not use them. This is the steam locomotive we are dealing with here and any improvement is a little step forward and a long journey is made of many steps.
The Niagara class achieved 6,600 ihp on test for a locomotive weight in excess of 210 tons. 242A1 could achieve 5,500 ihp and weighed a little over 145 and a half tons.

 

Thank you for an informative post. But your question about need (highlighted) can only be answered meaningfully where you look at the entire system, and how costs mount up through that system. As well as material, labour and fuel costs, arising from double (and more) heading, you also have other costs associated with both development/procurement of locomotives, and ensuring sufficient locomotives are available to fulfil the required traffic. There is also a time factor for the use of capital, where the payback may be too extended to be of real value to the operator.

These practical considerations may outweigh the genuine gains to be made from the use of more advanced technology, and make a decision to stay "as is" a rational commercial decision despite being sub-optimal in engineering terms.

 

Yet it is interesting that in Britain at least, attempts to produce higher power locos (i.e. actually built, such as the Gresley P1; or else designed but not built, such as the Maunsell 4-8-0 goods engine) foundered because the rest of the system couldn't effectively utilise the higher power: loops weren't long enough, wagon drawgear wasn't strong enough.

I'd suggest that in Britain, locomotive power wasn't really a limiting factor. Where we were hamstrung was the continued reliance on low capacity, unbraked goods wagons. No amount of research into optimum locomotive draughting was going to fix that problem.

Tom

 

It's interesting to apply this logic to electrification. The power levels and speeds rise to the economic limit of the available technology.

 

Indeed - but electrification (and dieselisation) brought in other factors to the payback at the time of implementation, for example utilisation.

 

You can read more about the Niagara class here:

https://nycshs.files.wordpress.com/2015/01/the-niagara-story.pdf

http://anyflip.com/qswoc/nguz

Knut

 

Thank you for a thoughtful post. As others have noted ‘Discount Cash Flow’ considerations will apply
aligned with the risk of redundancy through technical innovation ( less likely with steam technology
than more modern power ).

I note the comparison between Niagara and 242A1. Effectively the Niagara would have been 36 tons lighter
if designed as per the 242A1. (i.e. 174 tons ) In the context of a 1000 ton train a not inconsiderable saving
of 3.6%. However the unknown ( to me ); was the structural integrity of the 242A1 the equal of the
Niagara. (frames etc. ) . The Niagara, AFAIK, were very intensely rostered, a level never attempted in
France ?

Michael Rowe

 

Comparison of Niagara Class and 242A1:



from here: https://static1.squarespace.com/sta...d122b821d/1441150783767/#+DOMS-1_Chapelon.pdf

242A1 had almost half the grate area of the Niagara!

 

From a clean sheet of paper, that gives 242A1 a clear advantage. What interests me more here is whether that advantage for 242A1 was practically usable, or whether other factors meant that the compromises in the Niagara design were viable.

On a tangent, I recall my surprise at a friend talking about the car he'd hired when in Australia a few years ago. It was the Holden equivalent of the Vauxhall Omega, but with from memory a 3.5l engine where the UK equivalent would have been 2-2.5l. That seemed very inefficient, but his reply was interesting - it was much less highly stressed so worked well within it's capacity, and the fuel economy was actually little different.