On Tue, 21 Aug 2007 09:47:43 -0400, Wes wrote:
"Oppie" wrote:
"Mike" wrote...
On Mon, 20 Aug 2007 14:05:35 GMT, "Oppie" wrote:

Yes, the rail is thermite welded to a continuous rail except for insulated
signaling joints. They ship it as 1000ft sections though.

No, that webpage clearly shows a welding machine at the end of the
welded rail transport train - they preheat the ends to be welded with
gas torches, then it looks like it's either stick,flux-core or MIG
welded. The pictures didn't get close enough to tell.

Thermite may be fun to use, but I'm betting it's not their preferred
method if you have the right heavy equipment available. Especially
here in So-Cal, where the Smog Police ;-) (SCAQMD) would object to any
pollution source, no matter how small.

1000ft?
How?

It rather amazed the crud out of me to see these long flatbed cars - about
ten altogether, with the rails laid in grooves on the deck. I guess there is
a limit on how many rails they can cary and still make it around a curve...

The bare rail is rather flexible when it isn't bolted or spiked down
to the sleepers. It's not like thinwall tubing where it will buckle
if that much bending force is applied, even at the very large bending
radius of the average existing track.

How does this deal with the expansion and contraction of rails from daily
and seasonal temperature changes? Somewhere there needs to be some sort of
slip joint to deal with it.

Nope. A conventional slip joint on long sections of welded rail
wouldn't work, there's too much weight involved - a train hitting the
brakes or throttle would slam the ends around something fierce. I
suppose they have something that will work for bridge approaches where
you have to allow free movement...

They would need to have an active (powered) hydraulic buffer at each
joint to compensate for the trains braking or acceleration forces, and
the power cylinder would shift it back into position after the train
passes. The joint would require a long section of 'split rail' like a
switch frog (with the wheel flange retaining lip) so the two halves
can move a foot or more without the train falling off the track.
Trains falling off the track is bad. ;-P

A powered joint every mile or two with a computer interface to run
it remotely from the railroad's Control Center (like a remote switch
machine) would be really expensive to implement, and require
continuous power and data line availability along the tracks - and
unless they used backup batteries and 12V pumps to run the joints
(like they use 12V equipment for crossing signals & gates) a power
outage stops the trains cold.

And that expansion joint would magnify the problems they had with
the old bolted rails - the track sags at each joint, and that causes
the roadbed to subside at that spot, and then the train rocks as it
goes over the dip. Rock a high-speed passenger train or a light
freight running flat out, and they tend to fall off the tracks. As
was pointed out earlier, trains falling off the track is bad. ;-)

One trick I've seen (but not everywhere) is not to run the track
arrow-straight for 50 miles, put slight bends into the track so if
they're going to kink they have a safe place to do it... Even a
short-line built with sectional rails like Ventura County Railway does
this on the long straight sections - every 1/4 mile or so the tracks
shift over 20 feet on the ROW and straighten out again.

Otherwise the tracks can grow a foot a mile in a severe temperature
swing, and you end up with a "Sun Kink" (Google the term) where the
tracks move sideways on their own to release the compressive stress.
If the engineer is lucky they are visible and they can get the train
stopped (or slowed) in time, and the tracks stay attached to the
sleepers and stay in gauge so the train doesn't fall off the tracks
when they hit the sudden S-curve...

And trains falling... I won't say it a third time. ;-)

-- Bruce --