The MBTA has a pretty good explanation of this issue here, but I've got some more technical details.
Basic physics: make something cold and it shrinks, make something hot and it expands. The Framingham - Worcester line, as most rail lines nowadays, is constructed with 1/4 mile long pieces of continuous rail. That's correct, they come from the factory as 1/4 MILE long pieces. This is a pretty good video of how the rail is transported and laid out: https://www.youtube.com/watch?v=5xxVD2L_veM And here is a picture of a rail train:
A piece of rail has a "neutral" temperature ("rail neutral temperature" (RNT)) - the temperature at which it is neither in compression or tension. Once it is anchored down at this temperature, any changes in temperature create stress in the rail. Too much compression stress (from expansion in heat) = sun kink. The general rule of thumb is that if the rail temperature gets 40 degrees above RNT, there is a chance of a sun kink. Too much tension stress (from shrinkage in cold) = a broken rail. The good thing about a broken rail is that it usually breaks the track 'circuit' and changes all the signals to red. So there is some mitigation for the cold side of RNT. There is also evidence that rails are better able to stretch than compress - meaning you have more 'room' on the cold side of RNT than the hot side. In other words, a higher RNT is more desirable.
There has been much research into this topic, and the Federal Railway Administration (FRA) now requires all railroads to outline their plan for managing the risk of CWR (continuously welded rail). The target RNT in our area is now 105. Note that this is the RAIL temperature - not the air temperature. Everything we're talking about here is the temperature of the rail - which can be quite a bit higher than air temperature on a sunny day.
Prior to the MBTA purchase of the entire line, CSX dispatched all the trains. They had a policy of limiting speeds according to hot temperatures, with the theory that a slower train does a number of things:
1) does not impart more heat onto the tracks by the faster friction of speeding train;
2) allows more reaction time for an engineer to react to a sun kink; and
3) mitigates the chance of a derailment over a sun kink (slower train has better chance of dealing with a track defect).
Now that the MBTA owns the entire line and dispatches it, they can set and follow their own policy. However, their policy has to follow some logic or reasoning to comply with the FRA rules. For whatever reason, CSX or Conrail did not record or have the RNT of the Framingham - Worcester line. With an unknown RNT, the MBTA is forced to 'play it safe' and slow trains down like CSX did.
The good news is that the MBTA is spending a few million dollars to fix this situation on the Framingham Worcester line. Details are here. The "destressing project" essentially raises the RNT to a documented 105 degrees. The rail is released from the ties and heated to the new RNT. This usually requires cutting a small segment from the rail to allow it room to expand. Once it is at the new desired RNT, it is reaffixed to the ties and welded to the next section.
Note that they are only fixing track 2 from Boston to Framingham and track 1 from Framingham to Worcester. These are the tracks that carry MOST outbound evening commuter trains. So the project will fix the delays for MOST outbound trains in the evening, but it doesn't fix everything.
And yes, this is a real problem. The following accidents have been blamed on sun kinks:
- April 18, 2002 Amtrak Auto-Train derailment, off CSX tracks, near Crescent City, Florida. 4 deaths.
- July 29, 2002 Amtrak Capitol Limited derails, off CSX tracks, near Kensington, Maryland.
- July 8, 2010 CSX train derails off tracks in Waxhaw, North Carolina.
- July 6, 2012 WMATA Metrorail train derails off tracks near Hyattsville, Maryland
Sources / regards: