At the risk of beating the same old drum.....
Subsequent to our discussion about Sharbot Lake and HFR, an unnamed source sent me a very old document that gave very exact detail about the curves on the Havelock Sub east of Havelock. (I would stress that this source is not related to either VIA or CP, and there was no breach of any organization's confidentiality in giving it to me. The document is roughly 100 years old, but happily the track hasn't moved since the data was recorded, so it's still valid today.)
Anyways, based on this "actual" data, I was able to do some amateur calculation of likely track speeds, predicting a) how fast a train could go on the curves as they existed before the track was torn up and b) how fast a train might go on a rebuilt line.
First, here's the summary of what I found in the document :
View attachment 268257
Translation: There are 86 segments where the as-exists curvature exceeds 1.5 degrees. At that curvature, a "typical" train would have to reduce speed to negotiate the curve. My assumed "typical" train is the equivalent of an LRC with tilting deactivated - ie, assumed track superelevation (banking) of 3 inches, and equipment with a "cant deficiency" rating of 3 inches. (I pulled the LRC specs from memory, and may not be accurate, but it's just a straw man anyways) The speeds shown under the "FRA 3 inch" collumn are the speed allowed by the US Federal Railroad Administration for that track profile. (Note that the length of the curve does not affect the restriction - although a longer curve will require holding that train to the restricted speed for a greater distance).
So, under this baseline, in the 94 miles from Havelock to Glen Tay, a "typical" train would have to navigate 3.8 miles at 47 mph, 10.6 miles at 53 mph, and 12.9 miles at 65 mph.
Of the remaining track, trains could run 42.9 miles at 95 mph+. Because of all the slow segments, there are about 23.7 miles of track that is straight, but is adjacent to the slow bits, so trains would have to be accelerating/decelerating through these segments. These segments while straight consequently must be considered "restricted" by all the slowing down and speeding up.
The question is, what can VIA do to improve on this? One solution would be to increase the banking of the curves and/or use a tilting train that can handle curves faster. The collumn "FRA 6 inch" shows the speed restrictions that would translate to "LRC train with banking operating as designed" - ie 3 inch superelevation, 6 inch cant deficiency. As you can see, if VIA either found a tilting train, and/or banked the curves more, speeds through all those tight curves would rise.
The other alternative for VIA is to physically straighten some of the curves. It would take some pretty hefty engineering expertise to model that, but one can see that with the sheer number of restrictions, correcting a small number of curves would not produce any meaningful straightaways, and the time gained by fixing any one restriction is a matter of seconds. One would have to build a fairly long stretch of new track to gain a better end to end timing.
So, what does this mean for end to end time? I took
@reaperexpress's chart, and did some of my own calculations. As a baseline, I assumed that VIA would not upgrade the existing line east of Smiths Falls, so current timetable timings would remain unchanged. I assumed that the section following the CP Belleville Sub through Perth would be good track, but with a speed restriction through town (old timetables tell me that CP imposed a 50 mph restriction on passenger trains through Perth). I did a Google Map arbitrary assessment of curves and speeds west of Havelock, using data from another old CP timetable which restricted RDC equipment to 60 mph on certain curves - meaning these were all likely in the 2 degree range.
View attachment 268258
The bottom line? By my data, the "uncorrected" best time I could model, excluding time lost in meets, station stops, and contingency, was 3 hours 55 minutes. That doesn't sound all that appealing. However - using my "LRC with Tilting" scenario for only the Havelock-Perth segment, the time improved significantly to 3 hours 35 minutes. I have shown my data against
@reaperexpress 's estimate, and also against the best timetable time I could find from CP days. (The latter probably assumed 75 mph top speed, where I assumed HFR would achieve 95-110 mph speeds on tangent, but I was fairly conservative around the accel/decel issues). If one assumes, as
@reaperexpress did, that more speed can be squeezed out of the segments east of Perth and west of Havelock, things look much better.
There are so many assumptions and unknowns that one could poke huge holes in this data, and I'm not claiming any expert knowledge. But with some declaring that "tilting is the magic solution" and others arguing for straightened tracks, I though this at least put some perspective and bounds around what reality might be.....please, Ottawa, tell us the straight facts.
Food for discussion, and by all means rip it to shreds.
- Paul
