Armstrong presents the graph below specifying the resistance (friction) of train cars both filled and empty vs. speed.
does anyone know the typical rolling resistance (lb/ton) of a steam locomotive?
greg - Philadelphia & Reading / Reading
Greg,
Like rolling stock - I'm sure there would be a difference between locomotives outfitted with roller bearings and those that were not. I'm also wondering if there is a correlation between that and tractive power.
Tom
https://tstage9.wixsite.com/nyc-modeling
Time...It marches on...without ever turning around to see if anyone is even keeping in step.
i believe tractive effort at the wheels (or drawbar) is the difference between the force resulting from the cylinders and the resistance/friction of the locomotive that I'm asking about.
The rolling resistance of a steam locomotive was extremely dependent on the friction imparted by the many bearings on the many moving parts. Old-style friction or journal bearings imparted significant friction, especially when the mechanism was at rest, and there was no oil wedge present on the internal bearing surfaces.
Timken roller bearings were a vast improvement in the internal fricition situation. Check out the Timken 1111, the "Four Aces," a 4-8-4 heavy freight loco sponsored by Timken and a consortium of other parts suppliers, as a demonstration platform for the significant advantages inherent in Timken roller bearings. Built by Alco in 1930, and operated on a variety of mainline railroads, first on NYC, then on 13 more. Timken put much publicity effort into the locomotive, and there were numerous photos taken of the static locomotive being pulled by 3 or 4 young ladies, demonstrating the minimial friction effort needed to get the locomotive (417,500 lbs working weight) rolling from being at rest.
Check out https://en.wikipedia.org/wiki/Timken_1111 . There are also loads of photos of the "Four Aces" being pulled by young ladies available on the 'net, typically taken when the loco was passed on to another railroad for their turn at trying it out.
Today, the vast superiority of roller bearings over friction bearings has been well-established, to the point where roller bearings have become ubitquitous on almost all railroad applications.
What many call 'friction' bearings are known in the engineering world as 'simple' bearings. They were stickier than the more modern roller bearings found on most steamers built after about 1939-ish, and if not supplied as-built, they were retrofitted. Interestingly, I read somewhere that FEF 844's roller bearings were removed at some point...dunno why, or if it is true.
For any Whyte configuration, the heavier the weight on the drivers, the greater the deformation at the tire surfaces, meaning that much more rolling resistance. Just as a pneumatic tire has more rolling resistance as its internal pressure is reduced, it applies as well to steel on steel, just a lot less, so it is minor. A person attempting to shove such a locomotive by hand might not agree.
The post above mentions the Timken 4-8-4 which must have impressed not only those who witnessed the staged PR occasion, but also people wanting new and more efficient steam locomotives. I believe the only two 4-8-4 class constructed by the manufacturers with roller bearings were the Niagara and the Class J. However, many others were retrofitted subsequently when the results were known.
selectorthe heavier the weight on the drivers, the greater the deformation at the tire surfaces, meaning that much more rolling resistance. Just as a pneumatic tire has more rolling resistance as its internal pressure is reduced, it applies as well to steel on steel, just a lot less, so it is minor.
engineering class says it's just the weight, the area doesn't matter.
selectorInterestingly, I read somewhere that FEF 844's roller bearings were removed at some point
If weight on the bearing is all that really matters, is it that "simple" bearings required frequent lubrication
Armstrong's chart says 2 lb/ton for a 125 ton loaded car, 250 lbs. The Timken 1111 is 208 tons. could 3 people provide 400 lbs of force.
Looking at the chart used the notes say it uses the Davis Formula. A Google search brought up this basic form for the formula:
R = A + BV + CDV^2
R = resistance
A = Rolling resistance component independent of train speed
B = Coefficient of rolling resistance based on train speed
C = Streamlining coefficient
D = Aerodynamic coefficient, often combined as part of C.
V = velocity.
So the coefficients would be based on test data and manufacturing data based on the various conditions the train will be running in. The guide I found recommends modeling journal bearings using the A coefficient. Since this is based on Newton's formulas, I can only assume that it would be possible to model the change of resistance at different speeds via application of differential equations.
Looking at more details through the document I found there were several recommended methods to calculate the estimated coefficients via emperical units. It is worth noting the equation while it increases resistance for locomotives, it also neglects the actions of the running gear or traction motors; only focusing on the friction in the driving axle itself. So really the model doesn't care if its a steam or diesel or not; as long as the drive axle friction is comparable which is were the difference between simple and roller bearings would come in.
Anyways, here is a link to the document I was reading this from, probably teh best way to figure out how to answer your quesiton would be trying to get some data and calculating it based on the equations in here: file:///C:/Users/u0718580/Downloads/Train_Resistance_v3.pdf
Interesting —
Regards, Ed
Thanks Ed
i think the value of 25 lb/ton of the weight on the drivers is the value i'm looking for. Not sure how much it varies with speed. Also saw an aerodynamic drag coefficient of 0.002.
i read through other pages, but will look them over again.
gregci read through other pages, but will look them over again.
OK, glad it was of some use. I sent you a PM regarding drafting of locomotive smokeboxes as you were previously searching for. Maybe it is also of some use.
based on pg 20 of ALCO manual Ed posted, it looks like a formula the matches that data is
R = 25 * (ton + 1) + 0.24 * mph^2
gregc selector the heavier the weight on the drivers, the greater the deformation at the tire surfaces, meaning that much more rolling resistance. Just as a pneumatic tire has more rolling resistance as its internal pressure is reduced, it applies as well to steel on steel, just a lot less, so it is minor. engineering class says it's just the weight, the area doesn't matter. selector Interestingly, I read somewhere that FEF 844's roller bearings were removed at some point If weight on the bearing is all that really matters, is it that "simple" bearings required frequent lubrication Armstrong's chart says 2 lb/ton for a 125 ton loaded car, 250 lbs. The Timken 1111 is 208 tons. could 3 people provide 400 lbs of force.
selector the heavier the weight on the drivers, the greater the deformation at the tire surfaces, meaning that much more rolling resistance. Just as a pneumatic tire has more rolling resistance as its internal pressure is reduced, it applies as well to steel on steel, just a lot less, so it is minor.
selector Interestingly, I read somewhere that FEF 844's roller bearings were removed at some point
If only weight mattered, then the deformation on pneumatic tires wouldn't either...which is clearly not the case. As I explained, the deformation renders a flattened nether contact point with margins of a size demarking a defined area of resistance. Steel is no different. The greater the weight, the greater the area causing the resistance. It happens that, comparing the two types of wheels, steel-on-steel offers about ten times less resistance than a pneumatic automobile tire on pavement or on concrete. That's because the deformation is less.
I honestly don't know what the real difference would amount to...don't have the skills or experience to calculate it as a relatable example, and maybe the difference would only amount to 50 more horsepower, or less, that the locomotive would have to apply at the rails to overcome its increased rolling resistance due to its own weight...whatever that increased weight was. My point was to make it a real factor, minor though it would be.
So, the rolling resistance on a train, per truck, is much less than that of a road truck weighing the same at the tire bearing points....I mean on the ground, not over the axle bearings. That's why a mere 4400 hp can lift and move, say, 2000 tons at 20 mph before aerodynamic forces come into play in a big way and begin to offer their own contributions to resistance.
About the three (women) generating 400 pounds of tractive effort to move the Timken 4-8-4: Cars have simple bearings (back then, at least) while the locomotive had roller bearings. Additionally, the locomotive had larger diameter wheels, so less rolling resistance just on that account. And, if the women were wearing appropriate footwear, I believe three of them could generate 400 pounds. I'm not convinced they needed to. Probably more like 250-300 pounds, but who knows.
selectorIf only weight mattered, then the deformation on pneumatic tires wouldn't either...which is clearly not the case.
it's a stretch to compare the mechanics of friction between two hard materials and a partially deflated rubber tire and an asphalt road.
selectordon't have the skills or experience to calculate it as a relatable example
standard friction equation and friction and friction coefficients
selectorCars have simple bearings (back then, at least) while the locomotive had roller bearings.
i haven't found any info describing the difference in friction between simple and roller bearings.
as I said earlier, maybe the main advantage of roller bearings is they don't require constant lubrication. On a car, main, rod and cam bearings are lubricated with circulating engine oil. Wheel and drive shaft bearings use roller bearing that are greased periodically.
and yes, presumably the friction is just related to the weight on the bearing regardless of if it is simpler or roller. However, roller bearings are made of steel (0.16) and I think one of the other bearing surfaces is bronze (0.16) ???.
Post edited for quality assurance.
i agree. There are a lot of frictional surfaces on a steam locomotive. It's not clear how they compare: drivers, siderods, valve gear, cylinders, trucks. But base on the Timken experiment, it sounds like the dominant source is the drivers where the forces are greatest.
i think the value of ~25 lb/ton (adhesive weight) for a steam locomotive being significantly more than the ~2 lb/ton for a filled freight car shouldn't be surprising.
I remember the weight, not area thing from college physics. But that referred to sliding resistance. A deformed tire has to rotate, and the more deformed it is, the more of a "corner" it presents to the motion.
So if you locked up the locomotive wheels and rageed it down the track, then teh area/weight thing applies. But when rolling, it is more complex.
EnzoampsBut that referred to sliding resistance.
consider that if the wheel moved exactly its circumference across the rail, there would be no friction. If there is no friction, there is no force pushing the loco or train.
so the wheel does actually slide across the rail. It doesn't travel exacty it's circumference.
gregcas I said earlier, maybe the main advantage of roller bearings is they don't require constant lubrication.
Roller bearings do require constant lubrication - just not pressurized (hydrodynamic) lubrication. In the case of oil-lubricated Timken bearings, that involves an oil bath kept some percentage of the way 'up' the bearing; on grease-lubricated (sealed) bearings following AAR M-942 it involves grease 'packed' into them as in automotive roller bearings (on the axles or stubs for the wheels, as you note)
On a car, main, rod and cam bearings are lubricated with circulating engine oil.
This substitutes gallery pressure for the hydrodynamic force creating the 'wedge' in a typical plain railroad bearing. You may be familiar with the great nominal advantages in longevity that result when the various hydrodynamic plain-bearing surfaces in, say, a diesel engine are 'prelubed' by having the main oil supply pressurized before the engine is cranked around enough to start.
Note that roller bearings have been tried in areas of automotive engines where pressure-lubricated mains are normally used. Almost all these attempts were failures, sooner or later, because a hydrodynamic bearing has a large bearing area in a comparatively soft and self-forming interface, whereas the rollers are a series of near line contacts. The forces bearing on a crankshaft are NOT kind to line contact, especially between two hard-metal pieces.
Yes, there have been attempts to use pressure oiling on steam locomotives, including via a combination of hardlines, hoses, and cross-drilling of main pins, driver centers, and perhaps axles. The 'catch' is that Multirol-style needle bearings (which are not true rolling-element bearings with cages) with good grease lube do most all the job a full hydrodynamic bearing would, and without the problems with leaks and seal losses that an exposed pressure-lubricated setup would provide.
See also the Hennessy lubricator and its premise of action.
... roller bearings are made of steel and I think one of the other bearing surfaces is bronze???
Oh no. You might be able to make a bearing that way, but you'd have a plethora of potential problems for no real advantage (other than making one race or the other at least semi-sacrificial to save surface finish on the other -- and likely if you had grit contamination of the tribology you'd wind up making the softer surface into a more-or-less-effective lap that cuts down even very hard facing surface...)
Where the bronze is, is in a thin layer or sleeve around the entire outer race of the bearing -- this being particularly critical (and carefully machined) in the rod eyes of Timken lightweight alloy rods. (Do not believe sources that tell you the Timken rod eyes are 'brass' -- I have no idea how that wicked falsehood got established.) This gives the necessary compliance to use roller bearings with critical clearances of a couple thou (and damaged by distortion if hard-contact strained beyond that) with the necessary clearances for suspension action and lateral motion. Where the lateral motion is required, the bronze is machined with a spherical inside and outside profile, and the rods move sideways as a ball-and-socket joint.
I hear you crying 'but that isn't even remotely the amount of compliance necessary. How could these rods ever work?' Chapelon wondered the same thing, and came to the rather grudging but Holmesian conclusion that all the rest of the accommodation was done with rod flex or (reversible... we hope!) buckling!
Irrational post. Sorry.
Overmod gregc ... roller bearings are made of steel and I think one of the other bearing surfaces is bronze??? Oh no. You might be able to make a bearing that way,
gregc ... roller bearings are made of steel and I think one of the other bearing surfaces is bronze???
Oh no. You might be able to make a bearing that way,
i think what I was referring to is that the coef of friction between steel-steel and steel-bronze is (surprisingly) the same (0.16-0.6). Of course you make the easily replacable components out of the softer metal.
it seems to me that simple brearing friction is equal to the (max?) tractive effort (adhesive weight * ~0.25). Is this true?
i would think it would be much less for a roller bearing. Bearing Tips reports a coef of friction for roller bearing of <= 0.002.
is there any info on the impact of roller bearing on locomotive resistance?
gregcit seems to me that simple bearing friction is equal to the (max?) tractive effort (adhesive weight * ~0.25). Is this true?
No. The whole point of a hydrodynamic bearing (pressure-lubricated or not) is that once the 'wedge' is established, the rotating 'friction' of the journal surface drags around enough oil to force the surface asperities completely apart. At that point the resistance drops to the shear force of a relatively thin layer of oil molecules in mutual contact at whatever the ambient pressure is -- that pressure being related of course to the average of adhesive weight measured normal to the load (so highest near the crown of the bearing, and (without looking at the rubber bible or any notes) a relatively negligible factor in the effective viscosity.
Hence the observation (not very happily affirmed by the Timken people if you push them a little) that the drag of a good hydrodynamic bearing once started is comparable to that of a roller bearing. That's an average, of course, and there are all sorts of little things that foul up a pad-lubricated hydrodynamic bearing, including road shock that compresses the lube film until mutually-opposed asperities start to meet. The 'real' reason for adopting M-942 grease-lubricated 'package' roller bearings for interchange service is longevity and maintenance, not 'reduction of rolling friction' or some other roller-freight hoo-roar.
I would think it would be much less for a roller bearing.
That is most definitely true for starting friction, especially in a pad (or waste) lubricated bearing, where if the car has been sitting you might have metal-on-metal contact for at least the portion of initial journal rotation that fully 'wets' the crown brass. I think the Milwaukee Road did some careful comparative testing, on passenger equipment, of plain vs. roller bearings under different environmental conditions, with the conclusion that there was little practical difference in bearing frictional resistance above a very low speed, in the 2 to 5mph range. (At any temperature where the oil film 'wedge' can actually establish itself -- and remember there is a risk of scuffing grease-lubricated roller bearings at low environmental temperatures, too...)
I'd expect starting resistance of a pressure-lubricated bearing to be much less ... but the cost per car to provide it ridiculously high, with maintenance on top of that.
Bearing Tips reports a coef of friction for roller bearing of <= 0.002.[/quote]
I expect there to be slight differences between the Timken tapered roller bearing and the SKF barrel-shaped (alignment-more-intolerant) bearing -- you will probably have to dig further than you have an interest to find this. The TL;DR takeaway is as noted in some references: Timken bearings are better on engine and trailing truck bearings, SKF better on driver axles. All are slight because you have nothing but line contacts between very hard surfaces (see the Winans friction wheel at the dawn of the railroad age for a different version of this) and the actual deformation of the rollers or their instantaneous bearing point on the races is slight even in comparison with the wheel tread metal against the railhead. (Again, I'm sure that somewhere out there are high-speed photographs that show the exact form and amount of deformation under load; I can't help you find them.)
Is there any info on the impact of roller bearing on locomotive resistance?
Of course there will be no dramatic 'pull' tests of locomotives with plain bearings -- you'd have to get the girls to pull fast enough to levitate the bearings on the oil film and they can't develop enough horsepower to make the speed against the initial (considerable!) resistance. Were you to 'spot' them a couple of mph and neutralize the piston and valve ring drag they might be able to keep going for some time ... but below the critical speed sustaining the oil film, the asperities will start to meet and they'll drag to a halt.
I do not know (but Ed Dickens might) whether the grease lubrication on 844's rods (which was explicitly designed, and properly too, to allow 100+mph speeds reliably without the expense and fragility of the Timken roller-bearing rod solution) reduces friction at very low speeds. These are only weight-bearing with respect to the rod weight, and the thrust at starting acts almost normal to any motion between bearing surface and journal, so I do expect very low (if not Timken low) friction from that source. I have read, I think in the Red Devil book, the difference between Walschaerts and valve gear completely implemented with pin joints (e.g. Baker); the latter of course famously lends itself to being completely equipped with Multirols and hence possessed of low ongoing maintenance requirement.
Note that a Multirol (or other needle bearing) shouldn't really be thought of as a rolling-element bearing, as there is no cage to keep the rolling elements off each other and therefore considerable friction at twice rotational speed between adjacent needles. It's basically a method of providing a hard surface that self-renews grease on the journal and race surfaces with elements that can rotate rather than scuff when under load. A partial analogy is those machining tools with circular inserts that offer a great number of potential new 'cutting surfaces' with slight rotation of the insert in the tool.
For some reason I think there is some test data on NYC Hudson 5344 when it got its roller bearings and then a little later its roller rods. The difference is measurable, but it is critical to note what the locomotive is doing when the measurement is taken. A big point about driver roller bearings is that they give 360-degree support, so high piston thrust isn't walking the driver axle against the relatively weak, and poorly-lubricated, sides or shoulders of the driver brasses. There were and are methods of plain box construction that get around some of this, but it's just easier to put roller bearings in a proper box and be shut of the worries.
That is, if you know what you're doing. Just putting "roller bearings" on an engine can lead to disaster quick if you don't. You can look on RyPN and similar sources for a great tale of woe involving a certain 2-8-0 whose "caretakers" decided that small commercial roller bearings would be a Great Improvement over those evil old 'friction bearings' that an obsolete era had provided on their little jewel. I won't spoil the fun by recounting any of the shenanigans here.
do you have quantitative info (i.e. values)?
Best I can tell you is that they exist -- all my notes that might contain actual ranges of numbers are in storage.
If you contact Joe Burgard, he has PRR test plant reports that specifically include measurement of machine friction (and if I recall correctly, the methodology used to determine it) It would be short work from there to extract the pure frictional contributions. (If you have a brake-dynamometer test plant, it is quite easy to measure nearly-pure bearing friction; just knock out the crosshead keys and measure the driving power. The mechanical lubricator in a T1 operates off the crosshead, so you get rid of the significant cylinder friction effects, and moving the engine slightly can give you effective driving power for the truck axles (although you might, tediously, have to do each one in turn for clearance reasons). The roller rig for 5550 will specifically be spec'd and equipped to provide this functionality for testing.)
See if NYCSHS has a link (or reference material) to the 5344 material.
There was some controversy about a decade ago whether the Timken company still had some of the raw data from the Four Aces 'project'. It would do no harm to ask them. They have certainly given enthusiastic support to the T1 Trust.
Typically steel on steel dry static friction coefficient 0.8 drops to 0.4 when sliding is initiated - and steel on steel lubricated static friction coefficient 0.16 drops to 0.04 when sliding is initiated.
looks like a factor of ~10
max tractive effort / 10, adhesive weight / 40 or ~50 lb/ton. in the ballpark for 25 lb/ton previously mentioned.
Posted by mistake.
thanks Ed
looks like friction(?) reduced between 25 and 60%
Hope it helps you out.
gmpullmangregcis there any info on the impact of roller bearing on locomotive resistance? Roller-bearing_fig7 by Edmund, on Flickr
gregcis there any info on the impact of roller bearing on locomotive resistance?
Roller-bearing_fig7 by Edmund, on Flickr
I call foul, because this has nothing to do with what gregc is concerned about (resistance, not augment) and comparatively little to do with roller bearings (the lion's share of the actual improvement being in the lightweight rods and inboard location only incidentally facilitated by the thin eyes and reduced 'end liner' requirement allowed by the rod roller bearings. (The bearings themselves were, and are, recognized as increasing weight and therefore both inertial and offset-couple (albeit comparatively slight) augment contribution.)
Although you can bet Timken, which I suspect designed the whole High Dynamic Steel rod project around selling more bearings, isn't going to brag up that this is so.
Take a look (it's covered in Kratville's Mighty 800 book) at the UP developments in the mid-Thirties for low-augment high-speed rod construction that does not include roller bearings. It would be interesting to see the FEF (which is conspicuous by its absence in this supposedly objective graph) plotted in comparison with at least the two converted 'legacy' engines (the 5344 and the indicated K4s) to see where the actual advantage of the rollers themselves in reducing augment might lie.
Meanwhile, in case this is not familiar, the two Union Pacific locomotives mentioned are the ones that were converted (and a bit later streamline shrouded) to be protect power for the Streamliners. This is the 'whole package' of augment reduction including angle balancing, new disc mains, and I believe some suspension changes and snubbing; to attribute the dynamic reductions solely to Timken (as appears to be the clear intent of semantics here) is more than a little specious.
Overmod(resistance, not augment)
perhaps you can explain the difference ... in layman's terms.
My engineer told me that when the Y6 came along with its roller bearings, they had to learn to run the train all over again! And, that they did. No slide rule, no graphs, no high math needed. Just "the seat of their pants"!
.