Why are rails symetrical?

[Murphy Siding]

     I can understand why the base would be symetrical.  This year the rail might be in a corner sweeping right, next year it might be in a corner sweeping left.

     Since the flange only sits on the inside of the rail,  isn't the metal on the outside just along for the ride?

     I also don't quite get why some rail is made taller than others.

[Modelcar]

......I wonder, wouldn't the width of the rail top also give it strength as the web on a steel beam does...

[Murphy Siding]

     I thought about that,  but it seems like that type of strength was mostly in the verticle axis loading(?)

[mudchicken]

// Eli Whitney just rolled over in his grave....

[Frank Rothmann]

I'm not entirely positive, but I would think that a symmetrical cross-section would reduce the tendancy for the rail to distort under high or low temperatures - in other words if the rail were to shorten under cold temperaturs, or expand under high summer temps, the symmetrical rail would NOT tend to distort.

The taller rails used on main line track is better able to handle the weight and stresses of heavy, high speed traffic.

Frank, in Anderson, SC

[jrbernier]

  Many times rail is swapped to the opposite side when the gauge face wears out.    This evens out the wear and the railhead   The height if the rail is afuction of the total weight of the rail - 90# rail is not as tall as 132# rail.  There are 'compromise' joint bars just for connecting different weights of rails.

Jim

[Murphy Siding]

mudchicken

// Eli Whitney just rolled over in his grave....

     OK.  I just went and did a bunch of reading about Eli Whitney,  trying to figure out why cotten gins were relevent to rails.  Then,  I came upon the issue of interchangeable parts and figured out what you were trying to tell me.  jrbernier, in the post above did the same.

Thanks to both for the replies

[Paul_D_North_Jr]

(As edited a little bit over the last 10 minutes or so)

How long do you have to read what I could post about this, let alone what's already been written and published by others ? 

Short version:  If the piece(s) of rail are ever turned or even CWR is ever 'transposed' from being the left rail to being the right rail (or vice-versa) - both of which are commonly done to extract the maximum possible service life from the rail by using up all of the available wearing surface, depth, and width in its head - then the other side becomes the 'inside' or 'gage face' and gets worn instead.

In the meantime, the metal on the outside is a part of the 'beam strength' (= resistance to bending like a thin board) of the rail.  That's really governed by how much metal is how far away - both top and bottom - from the horizontal middle of the rail (technically, the "Neutral Axis").  Mathematically, it's approximately the cross-section area of the head x that distance, squared - so the outside of the head goes to 'area' of the head' part. 

The reason the rail is taller mainly goes to the 'distance, squared' part, although it also has to be a little taller to get more head depth and wearing surface and area in there.  For example, look at the rail sections on this webpage:  http://www.akrailroad.com/rail-sections  Notice that the "Mom. Inertia" - which is an easy measure of the strength of the rail against bending - of the 100-lb. A.R.E.A. Rail is 49.00 In.^4, but for the 136-lb. A.R.E.A. Rail it is 94.20 In.^4.  So, for only 36 % for steel, the strength has almost doubled ! (actually, 1.92 times as much more).  That's mainly because the 136 lb. is much taller - 7-5/16" = 7.313" vs. 6" for the 100 lb. - though that's only 22% more, which demonstrates the 'distance, squared' effect on increasing the comparative strengths, too.  Let's check this quickly and roughly: 1.36 times as much metal x 1.22 times taller x 1.22 times taller = 2.02 times stronger - that's pretty durned close considering all the shortcuts I took with that calculation !  

Quentin / modelcar has a point about the outer portion of the head contributing to its strength.  The rail is just as strong vertically regardless of whether the outer portion of the head is in close or farther out, like the base flange.  However, the outer portion of the head being in close contributes very little to the horizontal strength of the rail - and the outer tips of the base flanges are thin as compared to the head - so there's not much metal there, which is why it's easier to bend rail horizontally than vertically.  But there's enough metal in the outer extremities of the rail - the head and both base flanges - to give it some resistance to torsion, though not a huge amount - most of that comes from the comparatively stocky web and its fillets into the bottom of the head and sloped top of the base flanges.     

And Murphy, go look at the I-joists - by whatever brand name - in your lumberyard, if they're carried there - although they come in a range of widths, you'll see that they all look like a rail with a narrow head, except that they have a narrower base, too - not as wide as a standard I-beam or Wide-Flange beam.  I-joists too come in taller = stronger versions.  More to the point, the I-joists can replace the old standard 2 x 10 or 2 x 12 joists - even though their middles = webs are a lot thinner and weaker material, typ.1/2" OSB - because the heads and base are good sawn lumber or even a stronger glued laminate of some kind; the strongest ones use essentially a 2 x 4 laid sideways, the equivalent of a '4 by' joist, which would be astronomically expensive these days.  It's what's up in the head and down in the base that counts the most, not so much in the middle; and the location or shape of the head and base doesn't matter too much for the overall strength of the rail or I-joist, though those shapes do affect the other properties of the 'section'. 

What you didn't ask is about the inconsistency or mismatch between the standard taper on the car wheels of 1 in 20 (I think), as opposed to the top radius of the head of the rail - 10" for the 136 lb. RE section rail, per this diagram: http://www.akrailroad.com/Images/136_lb_rail.gif  That often wears a groove into the car's wheels early in its life, which can also cause a 'false flange' situation where the outer portion of the car wheel hangs down 1/4" or so below the tread, and hence does bear on the outer head of the rail, and also can cause some 'tracking' or wheel/ rail interaction problems, such as truck hunting, failing to follow the lined route of a switch, etc.  Some have suggested from time to time that the head of the rail also be beveled at 1:20 plane to one side as the primary wearing side - and CF&I even made special "Hi-Cant" tie plates for a while in an attempt to mimic that effect in curves - but that has gone nowhere, as far as I know.

- Paul North. 

[Anonymous]

The rail head needs some amount of horizontal area on top in order to just handle the vertical loading of the wheel.  So it is not just the flange side that needs to withstand the wear.  But that is an interesting question because the rail rolling manufacturing process would not fundamentally require a symmetrical rail cross section.  I would not be surprised if there are actual proposals for asymmetrical rail either already made or will be made in the future.  But I would not predict the reasons for those proposals.   In the future, it might have something to do with changing to new wheels.  

[tree68]

Somewhere in the depths of my memory I recall hearing that the flange is actually rarely in contact with the rail head.  A properly contoured wheelset will actually self-center between the rails.

[Paul_D_North_Jr]

See the 'girder rail' sections for street trackage, such as the 128-lb. and 149-lb. Bethlehem Steel sections at -

http://www.wirthrail.com/en/product_girder.htm 

 http://www.interfacejournal.com/features/10-07/girder_rail/1.html  and

 http://www.trolleyville.com/tv/school/lesson2_2/index.htm 

- which are about as asymmetrical as you could get, but are also hugely stronger because of all that additional metal up top.  But that also makes those rail sections very difficult to bend horizontally as well - for the sharper curves usually encountered in street and light rail vehicle trackage, they usually have to be pre-curved in a fabrication shop. 

- Paul North.  

[mudchicken]

Crane rail, like what Paul is talking about (not the heavy flange stuff), was giant pain in the patootie to deal with. Even with the flangeway, light cars had a really bad habit of wandering off on their own down the street. Between that stuff and submarine switches with single/double tongue switches - a roadmaster's day could get real ugly...Streetcar people can keep 'em....

[CShaveRR]

Two sides, two uses.  Maybe now Eli can quiet back down.  Perhaps he needs a shot of cotton gin to help that process.

[Paul_D_North_Jr]

Frank Rothmann

  I'm not entirely positive, but I would think that a symmetrical cross-section would reduce the tendancy for the rail to distort under high or low temperatures - in other words if the rail were to shorten under cold temperaturs, or expand under high summer temps, the symmetrical rail would NOT tend to distort.  [snip]

  The daily and seasonal temperature 'swings' are slow enough that the entire rail section is heated pretty much uniformly - and since it's spiked or clipped down, it doesn't distort much.

But in the steel plant's rolling mill, the rail is 'cambered' = bent upwards a little in the middle, so that it is flat/ straight after it cools - the bulky head portion holds more heat longer than the thin web and base portions - so it cools slowly, after they have - as a result, the head portion gets shorter and tends to curl the rail inwards in that direction.  Likewise, when thermite-field-welding rail ends together, the joint needs to be 'crowned' a small fraction of an inch so that it too will be flat/ straight when it cools.

Once in a while we had to [EDITED] use a cutting torch to remove one side of the base of a rail to use it as an inner guard rail for some sharp curves - so that portion of the base flange wouldn't interfere with the inside spikes of the running rails - or for a road crossing flangeway guard, etc.  Fortunately, the cooling/ contracting effect there tended to curve the guardrail horizontally in the direction that was needed/ desired, so that the remaining flange was towards the outside of the curve, and the cut-off base flange was towards the inside.

- Paul North. 

[ericsp]

Forget everything else, the real reason is that nobody wants to calculate the stress in an asymmetrical part. It is about as fun as banging your head against a wall. Fortunately, I have not had to do that since college.

[Deggesty]

CShaveRR

Two sides, two uses.  Maybe now Eli can quiet back down.  Perhaps he needs a shot of cotton gin to help that process.

Cotton gin? How do you make it; ferment cotton seed meal and distill the product?