very fine analysis [quote user="sgriggs"]:
Here is my top 5:
N&W J (Huge boiler, clever and unorthodox approach to running gear design)
ATSF 3765 class (Perhaps the only other 4-8-4 that could match the N&W J's physical specifications and peak output)
NYC S1a (The 4-8-4 optimized for very constrained Northeast loading gauge and NYC track pans)
UP FEF2 (Best of the UP 4-8-4's by virtue of its higher performing Type E superheater)
Milwaukee Road S2 (Seldom-mentioned type, fully qualified based on boiler & running gear specifications)
[/quote above] Thanks for putting the matter straight as possible, in my opinion.
A similar comparison can be with the various 2-10-4's, with the AT&SF tops, and the PRR-C&O next, but C&O before PRR if you value the booster, then Missabi, and then T&P, but all great locomotives. And then we can look at 4-6-6-4s. Were the last of the UP\s the best? New threads?
sgriggsIn terms of boiler evaporative capacity, direct heating surface (firebox walls and circulators) is approximately 6 times as effective as indirect heating surface (tubes and flues) on a lb evaporation per sq foot basis. The 3765 would be a more capable machine. I suspect the Santa Fe made the change because they felt the performance gain afforded by the additional siphons in the 3765 class did not justify the additional maintenance.
This is precisely the sort of answer I like to see.
There is a bit more to radiant uptake in syphon or circulator area that might not have been fully appreciated by some steam designers: the 'available' heat that follows Stefan-Boltzmann fourth-power-of-temperature absorption depends to an extent on luminous flame (which in both coal and oil firing means actual bright carbon in the plume). Problem with lots of uptake area configured to optimize internal water circulation, as with both syphons and tube circulators, is that they are cold as well as black as night to the plume, and with the TOF in a typical plume 9almost wholly at subatmospheric pressure combined with relatively high mass flow) once the glowing carbon quenches it may not be able to reach high luminosity before it hits the rear tube sheet or enters the tubes and 'goes out' to form soot. Conversely there may be so much heat in the plume in certain regions that some forms of circulator could encounter DNB leading to high spot overheating associated with mechanical stress from thermal distortion.
You may recall that the 3760 class had a boiler design as late as 1947 that involved double chamber circulators (I don't remember if they were syphons and think they were only tubes). That boiler was not retained, and although documentation, at least that I have seen, is not very specific about the reason for that, the removal of the revised boiler despite its nominal great steam-generation advantages speaks to fairly dramatic practical 'issues'.
sgriggsThe Type E superheater was a more advanced design than the Type A, and nearly all late steam designs used the E rather than the A. Although I can't put my hands on it at the moment, I have seen references stating the type E produced higher superheat than type A. Higher superheat translates to more efficient steam use in the cylinders and more power.
That the FEF3 returned to the type A configuration (considering the way UP intended to use these locomotives) may give you a guide to the issues with the type E on large engines run for relatively long distance at high speed. The E design had higher heat transfer, but remember that the Superheater Company formulae for proportioning superheater area in that era did not include superheater dampers, so at high speed the superheat even at 'efficient' mass flow could run crazily high, beyond what any contemporary cylinder oil could handle without starting to coke.
Meanwhile, of course, the useful enthalpy for expansion per degree of superheat is not all that great; by far the most important contribution of Schmidt superheating is reduction of various kinds of condensation during the important parts of expansive working. While it can be nice to have a few extra "free" BTU/joules from heat otherwise 'wasted' in the exhaust, once you've accounted for nucleate condensation up to working cutoff you start encountering reversible volume problems that complicate compression effects AND excessive exhaust-tract steam volumes in the wrong ways if you have too much superheat in the steam.
Part of the reason why jacketing is much more effective than 'more' superheat in the inlet steam can be seen in some of the physics associated with wall condensation. Only something like .007" of wall metal actually cycles between superheated and exhaust steam, and while it's theoretically practical to 'really' overheat the metal so it cycles up and then stays above where nucleate condensation starts in the expanding steam at any given point in the effective stroke, in my opinion it's easier and better to heat up the cylinder itself (or at least the part of it adjacent to the bore and heads) so that the actual cycling removes less superheat from the steam.
Overmod sgriggs In terms of boiler evaporative capacity, direct heating surface (firebox walls and circulators) is approximately 6 times as effective as indirect heating surface (tubes and flues) on a lb evaporation per sq foot basis. The 3765 would be a more capable machine. I suspect the Santa Fe made the change because they felt the performance gain afforded by the additional siphons in the 3765 class did not justify the additional maintenance. This is precisely the sort of answer I like to see. There is a bit more to radiant uptake in syphon or circulator area that might not have been fully appreciated by some steam designers: the 'available' heat that follows Stefan-Boltzmann fourth-power-of-temperature absorption depends to an extent on luminous flame (which in both coal and oil firing means actual bright carbon in the plume). Problem with lots of uptake area configured to optimize internal water circulation, as with both syphons and tube circulators, is that they are cold as well as black as night to the plume, and with the TOF in a typical plume 9almost wholly at subatmospheric pressure combined with relatively high mass flow) once the glowing carbon quenches it may not be able to reach high luminosity before it hits the rear tube sheet or enters the tubes and 'goes out' to form soot. Conversely there may be so much heat in the plume in certain regions that some forms of circulator could encounter DNB leading to high spot overheating associated with mechanical stress from thermal distortion. You may recall that the 3760 class had a boiler design as late as 1947 that involved double chamber circulators (I don't remember if they were syphons and think they were only tubes). That boiler was not retained, and although documentation, at least that I have seen, is not very specific about the reason for that, the removal of the revised boiler despite its nominal great steam-generation advantages speaks to fairly dramatic practical 'issues'.
sgriggs In terms of boiler evaporative capacity, direct heating surface (firebox walls and circulators) is approximately 6 times as effective as indirect heating surface (tubes and flues) on a lb evaporation per sq foot basis. The 3765 would be a more capable machine. I suspect the Santa Fe made the change because they felt the performance gain afforded by the additional siphons in the 3765 class did not justify the additional maintenance.
Overmod,
You'll have to forgive me, but can you define the acronyms 'TOF' and 'DNB'?
Scott
TOF = "time of flight". In this context how long it takes a particle of carbon carried in the combustion plume to get from some point along its trajectory (where it reached a level of incandescence where heat transfer to the radiant section and hence the boiler water becomes net positive) to the point where it is generally accepted to leave the radiant section (ideally, where convective heat transfer becomes more effective than radiant).
DNB = 'departure from nucleate boiling', where heat transfer across the sheet or flue becomes so great as to cause film boiling. Steam is a good insulator, so for high heat transfer rates this can cause runaway heating of the underlying plate to produce Leidenfrost/Eisenhoffer effect even under pressure. Two of the things a LaMont waterwall firebox is supposed to do are keep the water speed very fast past the waterside metal and then arrange for very quick steam separation at the end of the circulation path, ideally giving much better radiant heat transfer out of the gas and into the steam.
DNB='really really bad ju-ju' in any water cooled reactor... Design criteria is that nowhere in the reactor will heat flux exceed 1/3 of the DNB limit.
Ah, understand terminology now. Everything I have read and understand suggests the only way you are likely to see critical heat flux on the water side of a steam locomotive heat transfer surface is if that surface is not adequately covered by water (i.e. running with low water over the crown sheet). In that case, there is not enough water depth to carry away steam bubbles that form at the surface via buoyancy forces. In the case of a circulator or siphon, the surfaces are nearly all vertical or inclined, and because of their position in the boiler, they should always be well below the water level (if the circulators or siphons are not submerged, the crown sheet is completely uncovered and the circulators are the least of your problems!).
I did some calculations which suggested to me that a steam locomotive water side heat transfer surface should never approach the critical heat flux point.
I've tried to attach the classic boiling curve for water. I realize this curve is for 1 atm. I have searched for curves at higher pressures and was never able to find anything. The Y axis shows heat flux in W/m^2, which is convenient, because Ralph P. Johnson's book "The Steam Locomotive" provides typical evaporation rates for direct heating surfaces such as firebox side sheets, crown sheets, circulators, etc.
Johnson gave evaporation rates of 55, 80, and 125 lbs water/sq ft/hr for direct heating surfaces. The 55 number came from Professor Goss' Coatesville tests in 1912, and represents a very conservative rating, while the 80 value represented what could be expected from a modern locomotive at a reasonable firing rate. The 125 lb/sq ft/hr rate is applicable for very high capacity firing (on the order of 150-200 lbs coal/sq ft/hr firing rate).
You can use the latent heat of vaporization of water to convert the Johnson evaporation rates to heat flux values in W/m^2, which can then be compared with the boiling curve. Below is a table with the results of these calculations:
What I infer from this is that even at direct heating surface evaporation rates consistent with 200 lb/sq ft/hr firing rates, the heat flux is considerably under the critical heat flux point. Therefore, the only way I can see approaching the critical heat flux point would be if the water side surface was uncovered.
Scott Griggs
Louisville, KY
In my opinion, the effect of imposed pressure will limit, rather than exacerbate, formation of DNB film on affected surfaces, so in a sense there isn't much objective reason to plot the associated curves ... or to design boilers that can flirt closer to DNB in any component in normal operation.
I doubt that any 'conventional' Stephenson locomotive boiler approaches DNB in its usual components, and my analysis would be little different from yours. What I was referring to would be a fairly limited set of areas in large boilers equipped with multiple chamber syphons, specifically like those ATSF had diagrams for in 1947, and then quietly replaced. Since circulation patterns in such things are cheerfully drawn by the Nicholson people, but the actual water may not always follow those patterns as expected (perhaps in the presence of acceleration/braking or vibration, for some period of time) there may be a mismatch of where the fireside plume is producing high heat transfer at the inner waterside face and what the actual circulation at that point is providing.
That is not, of course, to say that it actually, or even repeatedly, would provide either short-term or runaway DNB at that point. Only that it can, and if it does the results are highly likely to be catastrophic in the cab.
I am almost completely certain that the 'detonation' of C&O Allegheny 1642 was due to the effect of low water complicated by the action of syphons doing just what Nicholson advertised as a 'safety feature'. At least part of the crown and perhaps some portion of the upper sheets was substantially exposed (leading to rapid plate heating) while the Nicholsons cheerfully pumped gouts of water over the exposed plate in nearly as perfect the wrong kind of random pattern as could be devised to quench the plate repeatedly to failure. Or perhaps more appropriately stated, quenched parts of the plate while leaving stochastically determined adjacent parts in full Leidenfrost/Eisenhoffer isolation even under boiler pressure.
sgriggs charlie hebdo sgriggs UP FEF2 (Best of the UP 4-8-4's by virtue of its higher performing Type E superheater) Milwaukee Road S2 (Seldom-mentioned type, fully qualified based on boiler & running gear specifications) Along the lines of Overmod's question: Why UP FEF2 rather than FEF3? Why MILW S2 instead of S3? UP FEF2 was built with Type E superheater, rather than Type A as built on FEF1 and FEF3. The Type E superheater was a more advanced design than the Type A, and nearly all late steam designs used the E rather than the A. Although I can't put my hands on it at the moment, I have seen references stating the type E produced higher superheat than type A. Higher superheat translates to more efficient steam use in the cylinders and more power. If you compare the specifications of the Milwaukee S2 and S3, you will see that the S2 has a much larger boiler, a higher operating pressure (285psi vs 250psi), the same size cylinders, and the same size drivers. The S2's were thoroughly modern engines with cast steel locomotive beds and roller bearings. Scott Griggs Louisville, KY
charlie hebdo sgriggs UP FEF2 (Best of the UP 4-8-4's by virtue of its higher performing Type E superheater) Milwaukee Road S2 (Seldom-mentioned type, fully qualified based on boiler & running gear specifications) Along the lines of Overmod's question: Why UP FEF2 rather than FEF3? Why MILW S2 instead of S3?
sgriggs UP FEF2 (Best of the UP 4-8-4's by virtue of its higher performing Type E superheater) Milwaukee Road S2 (Seldom-mentioned type, fully qualified based on boiler & running gear specifications)
Along the lines of Overmod's question:
Why UP FEF2 rather than FEF3?
Why MILW S2 instead of S3?
UP FEF2 was built with Type E superheater, rather than Type A as built on FEF1 and FEF3. The Type E superheater was a more advanced design than the Type A, and nearly all late steam designs used the E rather than the A. Although I can't put my hands on it at the moment, I have seen references stating the type E produced higher superheat than type A. Higher superheat translates to more efficient steam use in the cylinders and more power.
If you compare the specifications of the Milwaukee S2 and S3, you will see that the S2 has a much larger boiler, a higher operating pressure (285psi vs 250psi), the same size cylinders, and the same size drivers. The S2's were thoroughly modern engines with cast steel locomotive beds and roller bearings.
I interpreted "Top 5 4-8-4s" to mean my personal favorites, not necessarily the best. The FEF3 and S3 made my list because they both still exist and I have seen them run.
Our community is FREE to join. To participate you must either login or register for an account.