MBTA Commuter Rail recently tweeted out two photos of the rail destressing project. More details about the project here. This also explains the station closure - you really don't want to be near chopsaws and open flames while waiting for your train. 

Please read the original post: http://framwormbta.weebly.com/blog/whats-with-the-heat-restrictions
and "Take 2:" http://framwormbta.weebly.com/blog/heat-restrictions-take-2

According to scanner chatter on Monday 7/20/15, there are now four places that are subject to outbound rush hour heat restrictions. I'm not sure exactly where they are, but based on my observations, they may be:
1) Allston single track (from CP 3 to around milepost 5)
2) On either side of and through the switches at CP 11 (Weston)
3) From Framingham to Worcester

I'm not sure where the fourth one is, and I'm not even sure if that list is correct. I'll try to find out the real list, but the weather forecast indicates heat restrictions are probably not in the cards for the rest of this week. 

On Tuesday, 7/21, I watched the PTIS (GPS) data for the outbound rush hours trains and created this table:

This table shows the scheduled vs. actual arrival times for the outbound PM rush hour trains on Tuesday 7/21/2015. Heat restrictions were in place from about 3:25 PM until after 8 PM. As noted in the table, there were other problems with P583 and P531, so that data may not be useful in our analysis. I gave up after P531 arrived in Worcester, so I don't know how late P533 was at Worcester. 

The revealing data is that P525, P527, and P529 essentially arrived in Framingham on time - while all arrivals at Worcester were at least 20 minutes late. From my unscientific and only periodic watching of the train data, it appears that all trains operated up to 60 mph from Boston to Framingham, but no train exceeded 30 mph from Framingham to Worcester. 

So perhaps my theories in "Heat Restrictions: Take 2" were correct, and they're making progress with the entire project. 

Note that as I explain in previous posts, the heat restrictions are VERY complicated, which is why the MBTA justs advertises it as affecting the whole line. Issues:

1) Late outbound trains create late inbound trains due to equipment 'turns.'
2) There are NO repairs being made to the tracks used for inbound trains. So while an outbound train may be able to go 60 mph on track 2 from Boston to Framingham, an inbound train on track 1 on that same stretch is probably limited to 30 mph. 
3) Late trains screw up the sequencing and shuffling of all trains at switches, meaning that an outbound train just going to Framingham may STILL be affected by the whole mess even if it doesn't have a reduced speed restriction. 

Good news is PROGRESS! And it's great news for pax between Framingham and Boston, since heat restrictions appear to have a much reduced effect on them. Even Worcester pax benefit from being able to go fast all the way to Framingham. 

NOTE: the current 'turn table' is maintained here

With the summer heat restrictions in effect very early today, I thought it would be good to re-post the 'table of turn times,' even though I've blogged about it before. 

The basic usefulness of this table is that the lateness of an outbound train may affect an inbound train - and this table will tell us which outbound train affects which inbound train, and how long the delay needs to be before the inbound train is definitely affected. 

For example, Keolis is currently advertising P515 as "15-25" minutes delayed. So that could be a problem for P522, but it will be close. 

The current "Table of Turns:"

500, 502, 506, and 508 originate in Worcester after 'laying over' through the night
501 becomes 504 (Worcester turn, 10 min)
581 (non-revenue) becomes 582 (Worcester turn, 12 min ***SEE NOTE BELOW)
503 becomes 512 (Worcester turn, 22 min)
505 becomes 516 (Worcester turn, 13 min)
507 becomes 510 (Framingham turn, 11 min)
509 becomes 514 (Framingham turn, 12 min)
511 becomes 518 (Worcester turn, 16 min)
513 becomes 520 (Worcester turn, 19 min)
515 becomes 522 (Worcester turn, 24 min)
517 becomes 524 (Worcester turn, 24 min)
519 becomes 526 (Worcester turn, 37 min)
521 becomes 528 (Worcester turn, 12 min)
523 becomes 532 (Worcester turn, 25 min)
525 becomes 530 (Framingham turn, 52 min)
527 terminates in Worcester (lays over, usually becomes 508 equipment next AM)
529 becomes 534 (Framingham turn, 32 min)
583 goes back into Boston as 584 (non-revenue)
531 terminates in Worcester (lays over, usually becomes 506 equipment next AM)
533 becomes 536 (Worcester turn, 10 min)
535 becomes 538 (Worcester turn, 13 min)
537 becomes 540 (Worcester turn, 35 min)
539 terminates in Worcester (lays over)
541 becomes 542 (Worcester turn, 10 min)
543 becomes 544 (Worcester turn, 10 min)
545 terminates in Worcester (lays over)

***581 is a non-revenue move that departs Boston at 4:45 AM and arrives in Worcester at 5:53 AM. These times can be helpful for us...theoretically, if it is after 4:45 AM and MBTA / Keolis has NOT cancelled 582, there is a good chance it will operate, since the equipment has probably left Boston en route to Worcester. That theory is only partially true since the following things could happen:
1) MBTA / Keolis decides to delay sending out the cancellation notice for 582. 
2) 581 might break down somewhere between Boston and Worcester or in Worcester.
3) If some other layover equipment in Worcester is dead, MBTA / Keolis may decide to use the 581 / 582 equipment from Boston as a different train. I'm not sure that this has or would ever happen, but it is a possibility. 


_________________________
original post (from 5/17/15, provided here for the more thorough explanation): 

Recently on Twitter a fellow rider and I had a discussion about trying to predict if an inbound train would be late based on the lateness of an outbound train. This conversation prompted me to go through and figure out (and write down) the whole equipment usage for the Framingham Worcester line. Here it is...with some explanation. 

First of all, we need to get some definitions out of the way:
"Equipment" (as used in this post) = one train set, consisting of a locomotive and coach cars. 
"Turn" = the process of changing one set of equipment from an outbound to an inbound (or vice versa) train.
"Non-revenue" = movement of a train set from one place to another, BUT without passengers. A non-revenue move is a regularly 'scheduled' move with specific times, but isn't on the public MBTA schedule (it is on the employee timetable). Non-revenue moves also don't show up on the GPS tracking apps. 
"Layover" = overnight 'parking' for a train set on a dedicated siding where it can be plugged into ground power. At the current time, the Worcester MBTA yard has capacity to layover 4 train sets. No trains currently layover in Framingham or anywhere else on the line. 

As I noted on Twitter, it is impossible to predict turns at South Station. Equipment is used interchangeably (well almost, there are some restrictions or considerations) on all south side lines. In other words, the equipment used on Framingham / Worcester throughout the day does NOT stay dedicated to the Framingham / Worcester line. A set arriving at South Station from Worcester may go back out to Needham or Providence, and a set originating in South Station bound for Worcester may have come from Providence or Needham or elsewhere. I'm sure there is a master plan of entire south side turns, but I'll bet changes have to made almost daily to accommodate late (or broken) trains. Therefore, since the South Station turns are so dynamically managed, it wouldn't help us very much to understand the entire south side master plan. But it is worth noting that delays on other lines CAN result in delays on the Framingham / Worcester line.

However, at Framingham and Worcester, the turns are very predictable and can help us understand delays. So without further rambling, here is the entire schedule of weekday equipment usage on the Framingham / Worcester line. The list above notes the location of the turn and the duration of the turn (the difference between the scheduled arrival time of the outbound train and the scheduled departure time of the inbound train). The duration of the turn can help us understand the impact of a delay. For example, if train 513 is only delayed 5 minutes, it might not delay the schedule of 520. But if 513 is delayed 30 minutes, it will almost certainly delay the schedule of train 520. 

Acknowledgements: I first posted this to the railroad.net forum for confirmation and clarification of some aspects of my list. Thanks to the contributors there.