No doubt Ralph Johnson (the Baldwin guy) knew what his simple formulas were worth.
Overmod If "power" refers to horsepower, and it would be more or less stupid to pretend to gin up another 'definition', the documented Q2 maximum horsepower (in the PRR records) beats the Lima maximum Allegheny horsepower (as reported by Hirsimaki)
If "power" refers to horsepower, and it would be more or less stupid to pretend to gin up another 'definition', the documented Q2 maximum horsepower (in the PRR records) beats the Lima maximum Allegheny horsepower (as reported by Hirsimaki)
I thought that the Q2 Max HP of 7987 (right?) was IHP on the test stand, and the Alleghenys 7498 (right?) was measured at the drawbar over the road. IHP is taken at the cylinders so is going to be higher. I know that the book "The American Steam Locomotives: Design and development 1880-1960" by William Withuhn quotes the Q2 as having 6,645 drawbar HP, and conversely the C&O didn't ever get a IHP rating because they didn't have the facilites to do it. So comparing DBHP to DBHP the Allegheny has more... right? You seem versed in the deep lore of the Q2 and may know things that I've missed.
Overmod Not to take away from what timz just said: what 'characteristics of the engine' were you thinking of using to prepare the initial performance-curve estimates?
Not to take away from what timz just said: what 'characteristics of the engine' were you thinking of using to prepare the initial performance-curve estimates?
I wasn't thinking of using any characteristics of the engine because I wasn't thinking of preparing any performance curve estimates. the UP says that they used Balwin calculations to prepare drawbar pull/speed and HP curves and then said how their testing conformed with that or didn't. I suspect that the baldwin calculations are the ones from 'The Steam Locomotive, It's Theory, Operation, and Economics' which was written by baldwins Chief Mechanical Engineer.
IHP using something like the PLAN formula is, in my experience, often even more optimistic than Italian-car speedometers when you actually compare it to dynamometer data. Add to that the lack of a proper method to calculate direct steam mass flow together with effective piston thrust in the 1940s (it is interestingly difficult even today) so you have to fall back on flowmeter data from the injector and feedwater heater -- and these can lag actual steam consumption in the engine(s) by a significant margin.
One of the interesting Q2 details on test was the observation at high cyclic of what they thought were actual inertial effects in the steam passing down the long admission pipes to the rear engine. I would not have predicted you would observe this, but it was apparently severe enough on the test plant (!) to produce enough surge that there was expressed concern over the safety of continuing at that power level. You are getting into undiscovered country for steam designers using old formulae to approximate things like 300psi steam with in dampered superheater elements at cutoff longer than 40%.
There have actually been criticisms of the PRR designers for furnishing "excessive dead space" in the ports and passages of both engines, the claim supposedly being that there was excessive expansion after exhaust cutoff leading to flow disruption on subsequent admission. (While I have not seen a formal discussion of the port design, I suspect it was part of admission 'steam streamlining' to reduce perceived wiredrawing and similar effects at high cyclic/high mass flow.)
This is where the other 'they were fools' observation -- that the locomotive suffered excessive compression under the same extreme conditions -- comes in for discussion. In the '40s there was something of a 'fad' to adopt Okadee blowoff valves to limit compression, but once you vented the excess 'pressure' you would find it falling below admission pressure by the time the long-lap long-travel piston valve made it back to open the admission steam edge for the return stroke... into that relatively large port and passage volume... and your steam would waste some heat energy expanding to fill the dead space, just as it would on older locomotives. What you can do to solve this is to relieve the excess into an insulated reservoir connecting with the cylinder head, and then modulating the valve to allow the trapped steam back during the reverse of the valve so that the effective pressure from the port edge back into the cylinder volume was just falling past admission chest pressure on the 'other side' of the valve steam edge as it cracked open for admission. As cyclic increases, the interval for admission gets shorter and shorter even as cutoff gets shorter per stroke, but if the port is effectively 'pressurized' up close to 300psi, even if saturated you get better thrust on the piston 'early' and can use some of what may be excess superheat in the actual admitted charge to give mixing for low wall and nucleate condensation during expansive working.
It is much easier to work with things like this empirically than try to 'calculate' them, particularly without nondeterministic math. And you need actual road or plant testing, again with 'maximizing train-hauling effectiveness while minimizing cost' as the criterion, to have the empirically-derived results make sense to management...
Conductor_Carlyou could make a calculated drawbar pull over speed curve relatively easy ... based on the characteristics of the engine.
Have you by chance seen Farrington's book The Santa Fe's Big Three? It has DBHP curves for a 4-8-2 and a 4-8-4 that couldn't be calculated "based on the characteristics of the engine". Or based on anything else. Only way to get those curves was with a dynamometer car. (According to the curve, the 4-8-2 had 1200 dbhp at 65 mph.)
If "power" refers to horsepower, and it would be more or less stupid to pretend to gin up another 'definition', the documented Q2 maximum horsepower (in the PRR records) beats the Lima maximum Allegheny horsepower (as reported by Hirsimaki et al.)
In any case (and this is also substantiated in the PRR motive-power records at the Hagley) the argument is relatively nugatory, because the water rate of a practical locomotive using any kind of modern water treatment for its boiler life becomes impossible to support above about 8000hp -- that was the specific factor that caused the V1 project to be shut down even while there was still a perceived 'wartime' need for it.
Depends what you mean by "settle". Assuming "powerful" refers to power and not tractive effort, then the 2-6+6-6 is the leading candidate, but there isn't enough documentation to nail anything down.
Is there even enough surviving documentation to settle the whole "which steam locomotive was the most powerful?" debate?
I have had to be very, very careful around Big Jim in particular regarding the "6300hp" for the earliest A -- you will notice not only was that by far the highest 'outlying' figure for any A, even the 'perfected' ones running the precision-scheduled heavy trains to Kenova, we never saw it duplicated for the record in the comparative testing against the F7s (where supposedly the EMD guys tried to cheat and N&W cranked up the safeties to 315psi, etc. etc. etc.)
Could you have refined a locomotive with the A's controlling dimensions to produce 6300hp? Quite possibly. But I really don't see it for the engines as built, and that is not meant as a slam at N&W or its people.
Hirsimaki documented the rejoicing (with, to me, clear overtones of gloating) when the Allegheny test data came back 'beating' the A's figure.
Of course you also have to explain why PRR's test data came in even higher than the 'best' for the Allegheny, with two fewer drivers, no hinge, and truly ghastly nominal dead space in the tracts. And the predicted 8000hp for the V1 using a virtually Q2-sized boiler's steam more efficiently was if anything conservative based on the numbers...
You design the engine around IHP. If you want to tinker around with some of the numbers, particularly the 'allowance' you make for AWP in service relative to MAWP as safetied, you may certainly do so, and you can then confirm the estimated figure with actual indicator data as you run around (level, upgrade, downgrade, or whatever) to determine that the steam is going through the cylinders as expected.
THEN you start testing to see whether the power is being expressed at the driver rims (if you have a fancy Prony-braked test plant) or at the drawbar (if you care about what matters to actually using a locomotive to try to make money in the railroad business). And make no mistake, what matters is not what the horsepower 'number' or bragging rights about it might be, but how much you can effectively hang on the drawbar to work most effectively over the road to make revenue.
That is why, in the T1 testing, the 'machine friction' part of the efficiency is not emphasized more (as I recall it's nominally just under 10 percent, which I consider phenomenal for a reciprocating steam locomotive). What you care about isn't what the indicator tells you you 'ought' to be getting -- what you care about is what the engine produces. And from there, how you either improve what you've got, or how you operate it.
We might also mention, as a cautionary tale, Fry's book about locomotive boiler performance from 1922. This was effectively 'unobtainium' to actually read until comparatively recently, but it was famous among other things for utilizing a lolog function for practical heat uptake from combustion gas traversing the tubes and flues. Imagine my horror when, on finally getting to read it, most of his 'calculations' involve empirical factors scaled to 1922 practice, making none of his formulae particularly worthwhile in predicting performance from, say, the double Belpaire for Townsend's 4-8-6 project.
Incidentally, while I continue finding the original volume containing discussion of the initial PRR test plant at the St. Louis Fair circa 1904, here is something that Carl might care to read as representative of Test Plant analysis in 1924:
https://books.googleusercontent.com/books/content?req=AKW5QacoyG9VtpMSuTJHxUhbDqFdSZLTibqiPBlUuxx-WUS7G9Kaw7yJOBjjf68BsaQEo2ZtKBI2UZESW0EeGvDPoeKHVnji6srAQLQ0fAN58UQTZTdZmdZZO58ksj5x-Pv9Rz3oeijIwbKERLtNzwjn7dvgNV8nAz3-9fHCW9SiaAiZHuceP9DtQP4V2gcxzQ4TcxDC8RIlg9F71MTnX1_twKdb2Uyq00Ik-fdO8fl5870qYDOhwXJ1QH4DlLIYeswOqohHFmYYMuyEyanX-rCHoT7Ozg2j14SK3Xp9dTp14p4FLVxYcco
timz What calculations? You make it sound like they thought they could calculate the power the engine was supposed to produce. No doubt they knew better. They could try to predict its power, but there wasn't any "supposed to" involved.
What calculations?
You make it sound like they thought they could calculate the power the engine was supposed to produce. No doubt they knew better. They could try to predict its power, but there wasn't any "supposed to" involved.
you could make a calculated drawbar pull over speed curve relatively easy (and very easily now that we have spreadsheets that can do the gruntwork of calculating values in 1MPH increments) based on the characteristics of the engine. So they did just that, and the drawbar pull over speed curve that they have includes a dashed line for the calculated curve. That statement that it overperforms the calc over 30 MPH and underperforms the calc between 15 and 30 MPH is one of the conclusions of the test report (and yes, I did goof, its between 15 and 30, not 20 and 30 as I said in the quote).
And with regards to the N&W and documentation, they have heaps and heaps and heaps of documentation on all kinds of things... just not those tests from 1936.
Conductor_Carltesting showed that it actually underperformed the calculations at between 20 and 30 MPH
Conductor_CarlI full stop don't believe the A's stated HP of 6300
N&W's unlikeliest claim was that an A pulled 7500 tons at 64 mph on the level. (They originally said "comparatively level", but later the adverb fell thru the cracks.) For all we know the A did produce 6300 dbhp, but for all we know it didn't pull 7500 tons at 64 mph. N&W said it did, which tells us: when N&W says something, we need documentation, same as with anyone else. (Don't hope to get it, tho; documentation isn't something Americans do.)
Overmod Data taken on the PRR test plant are effectively 'on the level' so no grade compensation is observed. When calculating train factor the service weight of the engine and tender, with fuel and water anticipated for the ruling part of a grade, would simply be added as a factor in the Davis formula. I think Tim Zukas may know how many railroads did their horsepower calculations based only on drawbar pull at speed vs. calculating a dbhp number (which to me always smelled faintly of an effort to secure bragging rights for the biggest number, like the reported Lima celebration when the Allegheny surpassed the N&W's (itself a bit suspicious as never repeated) horsepower for the A.
Data taken on the PRR test plant are effectively 'on the level' so no grade compensation is observed. When calculating train factor the service weight of the engine and tender, with fuel and water anticipated for the ruling part of a grade, would simply be added as a factor in the Davis formula. I think Tim Zukas may know how many railroads did their horsepower calculations based only on drawbar pull at speed vs. calculating a dbhp number (which to me always smelled faintly of an effort to secure bragging rights for the biggest number, like the reported Lima celebration when the Allegheny surpassed the N&W's (itself a bit suspicious as never repeated) horsepower for the A.
Sure, test data can allow you to spike the ball, but confirming your calculated values and confirming that what you bought can do what you want it to do is important.
Anyway, in the case of Big Boy I feel like the testing cuts both ways. the testing showed that it actually underperformed the calculations at between 20 and 30 MPH but overperformed at all speeds above 30, which is good information for the railroad to have. On the other hand, putting the corrected drawbar horsepower seems more like a effort to secure bragging rights than just using the drawbar HP from the data you got due to how the corrected HP will be larger than the measured, and due to how the correction has to be calculated.
And if we want to crack open a different can of worms, I full stop don't believe the A's stated HP of 6300. due to how the only documentation is a article in railway age, that does not describe tonnage, does not describe grade, and isn't dated. Add to that that these tests don't seem to be availible on the N&WHS site and the fact that later drawbar/speed curves for the 300 PSI A's are lower, and I think those tests in 1936 were botched at a minimum and fabricated at a max.
There are extensive PRR records, for the Q2 in particular, with charts of drawbar pull vs. speed that could easily be converted to dbhp. (I have posted a scan of the 'comparison' graph generated for the V1 turbine in a number of past threads here; the original is a lovely thing in multiple colors at the Hagley Museum in Delaware -- this has a number of highly interesting locomotives and, for example, clearly shows where the Q2 booster was cut out after starting...)
sure enough, when I jammed in the big boys weight to the grade resistance calculation (595 tons times the grade times 20) the number it spat out ( 6780 pounds) was just 300 pounds over the difference between the measured and compensated pull. Add in the fact that the train had probably lost some water and coal at this point and I think that you are correct in that this is compensating for the engine weight on the grade.
Now I need to get my mitts on other railroads dynamometer reports and see if they do the same thing. What I am looking for is a like for like comparison, and I dont know if anyone else used this compensation factor for their drawbar HP.
Rather than use 'horsepower', some railroads measured 'drawbar pull at speed' (which is easy to plot with a typical dynamometer or even an instrumented drawbar). This is a different thing from 'indicated horsepower' or IHP, but it accurately describes how much 'revenue train' a locomotive will pull under given conditions. (I presume you understand the Davis formula and its use...)
BUT
Horsepower measured at the drawbar when the train is working upgrade is only partially representative of cylinder output, because the weight of the engine and tender is also being 'moved upgrade'. Obviously if the track were level, that much more weight could be pulled, and therefore could be 'added' in the Davis formula to get an 'equivalent' drawbar-pull correction.
One of the reasons 'brake locomotives' or dynamic-braking arrangements were used for some testing is that they can simulate increased trailing load, as if for the 'ruling' part of a grade, without having to backtranslate for the locomotive weight.
What if they're testing it on a 5% upgrade? At 40 mph, its drawbar pull would be zero. Not too informative.
But if we know the engine and tender totalled 600 tons, then we know the engine was working hard enough to produce a drawbar pull of 60000 lb, on level track. That's informative.
Was wondering if some clever people could set me straight on something.
I just got a copy of "Big Boy" by William W. Kratville. In this book there is a table documenting the testing of locomotive 4016 hauling a Dynamometer car (borrowed from the Santa Fe) and 70 additional cars weighing a total of 3883 tons. While the book does not have a trace of the dynamometer car readings, it does state the milepost, the speed, the actual drawbar pull, the grade, the curve at that milepost, and any acceleration. Coupled with excellent grade profiles of the Wyoming division, it paints a very detailed picture.
And then it does something that I just dont get. It adds a column for corrected drawbar pull (corrected to level tangent track) and then bases the horsepower off of that. This leads to the statement that at milepost 958, at a speed of 41.1 MPH on a uphill grade of .57 percent with no curve or acceleration, the Big Boy produced 6290 HP with a corrected drawbar pull of 57400 pounds derived from an actual pull of 51000 pounds.
So my question here is, why is the actual pull not what is used?
If my knowledge of Dyanamometer cars is correct, they do nothing to compute the curve or grade or any of those effects, they just measure the pull. Any impact of changes in grade or changes in curve basically have to be inferenced by comparing the grade profile with what Dynamometer trace shows. For example (because I have them on hand) dynamometer traces of the Allegheny on the Alleghany subdivision will show how the drawbar pull increases and the speed decreases as the engine leaves flat ground and starts to head up the .57% grade, when you plot horsepower from this it will start relatively low (as the train moves on the level) it will spike as the train starts to head upgrade increasing the pull but before the speed loss is really noted, and then it will return to another equilibrium as the train settles at its new speed and pull. At no point on these traces is the pull corrected for the grade.
The way that the pull is corrected for the Big Boy seems like they are stacking the deck. As a train goes from level track to a gradient the drawbar pull is going to increase due to gravity, but then it seems the 'compensated' pull calculates what the equivilant drawbar pull would be on flat ground. Basically, it seems that if the pull was 10,000 pounds on the level, and then it increased to 11,000 pounds as the train started uphill, the calc the UP (or William Kratford) did would then say that the train pulling 11,000 pounds up a hill is the same as 12,000 on flat ground and then calculate HP off of that. It looks like this calculation adds to the measured pull when it should probably do the opposite.
So am I missing something, or should the actual measured HP of the Big Boy be 51,000 pounds (the actual measured pull) at 41.1 MPH/375, for a total of 5,590 HP?
Our community is FREE to join. To participate you must either login or register for an account.