Curiosities & Oddities

[landnrailroader]

When I digitized the Anchorage to Fairbanks line, I was very careful to mark as close as possible to the center line of the track as seen on Google Earth, especially on curves. I am seeing things I never saw working on lines in the "lower 48". Perhaps it is because of the proximity to the north pole or ? While some curves compare correctly between what it takes in track pieces to fit to a digitized curve and what the track chart says it is, this is not always the case. And sometimes the difference is large. Take the left hand curve at mile 182 plus a couple hundred feet. The track chart says this is a 4 degree 8 minute curve. Thus I would expect a radius of about 422 meters. However this produces a much less sharp curve than what will fit the carefully digitized line. This would require a 8 degree or 218.25 meter radius curve. I suspect there is a drafting error. Although this curve is on a .53% gradient and I know from experience that a SD unit going downhill in dynamic will tend to make a curve sharper at the lower end - this was a source of some derailments in the late 60's on the Southern and careful measurements showed this anomaly after a heavy train passed and it led to a lot of extra work to anchor the track, this is probably not a factor. Modern units, from around 1975 onward have improved trucks that prevent this issue. And when the curve is digitized, it is found to be a smooth curve. So guys, when some of you eventually download this unfinished route, no, I'm not insane, though evidence might point that way.

Eric, do you know of some strange scale factor between Google Earth & TSRE, or would such strange things be a fact of life near the earth's poles. For example there is another curve a mile or so further on, near mile 183. Track chart says that this is a 5 degree, 59 minute curve. Measuring it on Google Earth by striking radii toward the center suggests that it is more like an 8 degree curve. And then to fit the digitized line in TSRE for the best fit requires using 150 meter radius which is about a 11.5 degree curve. If you want to take a look at this area, milepost 183 is at -150.0068/61.7114. I know that map projections get real touchy near the poles and this is the only thing I can think of but it wasn't a problem south of Anchorage.

Jerry Sullivan, P.E. (retired from SR & CSX)

[landnrailroader]

Eric answered my question and the oddities that I am seeing are related to the difference between the coordinate system TSRE/ORTS uses and Google Earth. It is odd in that some curves are almost perfect, while others are considerably different. In any case, when I release the route I will point this out in documentation. The only practical thlng would be that in doing activities, there will have to be speed restrictions on curves. For example on those above, the track chart indicates 40mph freight or passenger while the degree of curve required to match the digitized points limits to 10mph on jointed rail or 15 on welded rail. Oddly it is not always the sharper curves that are the problem. Some it is, other areas, slight curves have to be made sharper. And another thing is that there are several passenger trains on this line. I believe there was one tunnel, located along one of the rivers at a point where it was visible from the Parks Highway bridge over the river but it has been daylighted. There are grades up to 2% and the elevation ranges from 10 feet or so to over 2000ft. above sea level.

Jerry

[landnrailroader]

Another anomaly that we encounter has to do with gradients. The value shown on track charts, at least those of most railroads, is the "adjusted value". When curves are present, the friction between the wheels and the curved rail has the effect of increasing the gradient. For example at mile 233 on the Alaska RR, going north, there is a short stretch of 1.03% grade on a 1 degree, 47 minutes curve, i.e. 1.78 degrees. At the start of the curve, the elevation is 403.6 feet and generally these values are close to what Google Earth shows. If I use 1.03%, then at the end of the curve, instead of havinig 406.4 feet, I have 413 feet, so obviously this is a adjusted gradient. I reduced it to 0.6% and then lowered 10meter pieces of the curve by .1% and came out the end very close to the desired 406.4 feet. So usually it will be necessary in the presence of curves to reduce the value of the simulated route gradient, else we come out of the curve too high.

And for this curve, I used M-tracks, 1 degree 45 minute curve.

Jerry