I am working on the boundary for a piece of property that is the remaining portion of two tracts in two different subdivisions. The property in question has been leased for many years with an option to buy, so now we are preparing a partition plat.
The trouble lies along the east boundary which is described as XXX ft more or less at a 10 ft offset when measured at right angles to the center of the track. The as-built the centerline of the existing track at approx 30ft intervals along the curve is giving me fits. There is a 350ft +/-tangent in and the arc length is 300ft +/-.
I have no design data for the rail. This is a rail line that served a civilian airbase during WWII and later was turned into an industrial park. There is one other survey (ca 2000) that developed curve data for the railroad, but my as-built radius is longer by approx 26ft (460 vs 486).
I am not content with the "best fit" curve holding the tangent in and the tangent out and my centerline points. Some of my as built centerline points miss the best fit by 1.5 ft!
How can I tell if this is (or should be) a spiral curve based on the as-built data?
I would expect my best fit centerline to fit the geometry +/- 0.20ft. Are my expectations too high?
Any recommendations?
I will post some maps when I am able.
You are probably looking at a spiral curve. I have seen very few curves on a railroad that were not spiral. With that said you can not offset a spiral. Good luck
Matt
Thanks for the reply.
I am going to try fitting a Searles Spiral to my as-built.
See this thread...
[msg=179413] http://beerleg.com/index.php?mode=thread&id=179413#p179503 [/msg]
You can offset simple curves. You can offset straight lines. There are straight lines in the Tangent portions. There is a simple curve, in the middle of every spiral curve. So, this leaves the transition from tangent to Simple, that is a spiral. You can approximate this portion with a number of solutions. A series of tangents. Or with a series of simple curves, with different radius points, and different radai. (or radiuses!)
I am sure others will have a better answer, but That is the direction I'd head.
N
I was going to point this out. Most R/R curves are actually 3 curves ... the spiral curve is a transition from a straight segment to a radial curve, and another spiral to transition from a radial curve back to straight.
So a R/R curve normal goes spiral - radial - spiral.
I was able to best fit two tangent compound curves to my as built with field tied centerline points matching my developed geometry +/- 0.10ft. There was one outlier that was out by 0.3ft, but I can live with that.
Thanks for the input fellas. This was a weird one since there was no record data to work with or against 🙂
There is not a simple curve inside a spiral. A spiral starts at tangent and progresses through an increasing degree of curve until it reaches a degree of curve equivalent to the simple curve it may or may not transition into.
> to the simple curve it may or may not transition into.
OK, those that I have tied, did in fact have a simple curve, in the middle, between the 2 spirals. I guess you are correct, there does not really have to be a simple one there.
Hmmmm where is there a spiral in, and a spiral out, with NO simple one in the middle?
Probably are some like that, but I have not yet seen them.
N
