Acela derailment

Amtrak Unlimited Discussion Forum

Help Support Amtrak Unlimited Discussion Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Status
Not open for further replies.

jis

Permanent Way Inspector
Staff member
Administator
Moderator
AU Supporting Member
Gathering Team Member
Joined
Aug 24, 2003
Messages
39,111
Location
Space Coast, Florida, Area code 3-2-1
Westbound Acela 2171 split a switch and derailed at relatively slow speed about at CP Davis west of Providence RI a little after 4pm. Passengers were transferred to a Regional (175?). 2173 behind that was running about an hour late according to reports from those on board 2173. No injuries.
 
Last edited by a moderator:
LINK AND VIDEO.

On the 11:00 news, it was still there. And I'm very surprised about the slow speed. This is one of the 150 MPH sections!

And it is actually southbound - even though it goes west.

EDIT: It was more then 3 miles south of PVD, as North Kingstown is MUCH further south then that. I am trying to name the place of the incident, but can not figure it out, Maybe by the bridge and houses next to the track, it could be ny Quanset Point, which would be almost 10 miles south of PVD.
 
Last edited by a moderator:
the_traveler said:
1333117752[/url]' post='357421']LINK AND VIDEO.

On the 11:00 news, it was still there. And I'm very surprised about the slow speed. This is one of the 150 MPH sections!

And it is actually southbound - even though it goes west.

EDIT: It was more then 3 miles south of PVD, as North Kingstown is MUCH further south then that. I am trying to name the place of the incident, but can not figure it out, Maybe by the bridge and houses next to the track, it could be ny Quanset Point, which would be almost 10 miles south of PVD.
Davisville, RI. It happened where Davisville Road meets Old Baptist Road and parallels the tracks for a short distance.
 
EDIT: It was more then 3 miles south of PVD, as North Kingstown is MUCH further south then that. I am trying to name the place of the incident, but can not figure it out. Maybe by the bridge and houses next to the track, it could be by Quanset Point, which would be almost 10 miles south of PVD.
Davisville, RI. It happened where Davisville Road meets Old Baptist Road and parallels the tracks for a short distance.
Then I'm correct in my guess!
smile.gif
That is the location of the track switch going to Quanset Point Industrial Park! (The sound on my computer is broken and I just caught the end of the news story, so if it was in that clip I didn't hear it.)

Off topic for trivia buffs: Quanset Point used to be a US Navy Base, but has since been decommissioned. (It is now an Air National Guard Base and an industrial park.) Quanset Huts (those long buildings with a round roof) were first built on this Navy Base during WWII - and some of them are still used!
cool.gif
 
the_traveler said:
1333166593[/url]' post='357569']
ParrotRob said:
1333165057[/url]' post='357562']
the_traveler said:
1333117752[/url]' post='357421']EDIT: It was more then 3 miles south of PVD, as North Kingstown is MUCH further south then that. I am trying to name the place of the incident, but can not figure it out. Maybe by the bridge and houses next to the track, it could be by Quanset Point, which would be almost 10 miles south of PVD.
Davisville, RI. It happened where Davisville Road meets Old Baptist Road and parallels the tracks for a short distance.
Then I'm correct in my guess!
smile.gif
That is the location of the track switch going to Quanset Point Industrial Park! (The sound on my computer is broken and I just caught the end of the news story, so if it was in that clip I didn't hear it.)

Off topic for trivia buffs: Quanset Point used to be a US Navy Base, but has since been decommissioned. (It is now an Air National Guard Base and an industrial park.) Quanset Huts (those long buildings with a round roof) were first built on this Navy Base during WWII - and some of them are still used!
cool.gif
They didn't come out and say it, but the house they showed in the video is plainly visible on a satellite pic. It looks like the train was derailed by the crossover immediately north of the Quonset turnoff. There's a crossover between the center track and the westernmost track
 
This is another proof positive that human ingenuity can always defeat any safety system. According to reports posted on trainorders this is what happened.....

Switch failure reverse. Maintainer cranked the switch over and spiked and wedged it, but forgot to crank over the movable point frog. Engineer got a 241 past the signal and did not look out for switch not properly lined....
I bet there are several red faces and probably one or two disciplinary actions involved.

Location was Davis interlocking I believe. The derailment was contained on a single track and hence effect on service was relatively less severe.
 
This is another proof positive that human ingenuity can always defeat any safety system. According to reports posted on trainorders this is what happened.....

Switch failure reverse. Maintainer cranked the switch over and spiked and wedged it, but forgot to crank over the movable point frog. Engineer got a 241 past the signal and did not look out for switch not properly lined....
I bet there are several red faces and probably one or two disciplinary actions involved.

Location was Davis interlocking I believe. The derailment was contained on a single track and hence effect on service was relatively less severe.
What is shocking to me is that, apparently, there was no formal checklist procedure for a maintenance operation on 150mph track. Coming from an industry where worker and public safety was paramount, nothing took place on our facilities that was not formalized by a reviewed, written procedure and worker checklist. That way, nothing could be "forgotten." "I've done this a thousand times. I don't need a checklist.", are famous last words before an accident.
 
Just for general information, here is the relevant portion of NORAC Rule 241, with a portion that I have bolded - the portion that appears to have been violated by some one or more people involved:

241. Passing a Stop Signal

To pass a Stop Signal, a train must have verbal permission of the Dispatcher (or Operator when authorized by the Dispatcher). Permission must not be given or accepted until the train has stopped at the signal. A member of the crew must contact the Dispatcher or Operator and follow his instructions.

a. Giving Permission to Pass

Before giving permission to pass the Stop Signal, the Dispatcher (or Operator) must determine that:

1. Affected appliances are properly positioned. If the position of a switch cannot be determined, the route must be inspected.

2. No opposing or conflicting movements have been authorized.

3. Blocking devices have been applied to protect against opposing movements whenever the Stop Signal involved governs entrance to a track where Rule 261 is in effect.

The Dispatcher (or Operator) must give permission to pass a Stop Signal in the following manner:

“No. 5316 engine 4129 pass Stop Signal on No. 2 track at Rare and proceed east to No. 1 track.”

The receiving employee must repeat this permission and the Dispatcher or Operator must then confirm it.

b. Movement After Permission Has Been Confirmed

After permission has been confirmed, the train must operate at Restricted Speed until the entire train has cleared all interlocking or spring switches and the leading wheels have:

1. Passed a more favorable fixed signal,

OR

2. Entered non-signaled DCS territory.

OR

3. Entered Rule 562 territory with a Form D authorizing Rule 563.

In CSS territory, trains with operative cab signals must not increase their speed until they have run one train length or 500 feet (whichever distance is greater) past a location where a more favorable cab signal was received.
 
Last edited by a moderator:
This is another proof positive that human ingenuity can always defeat any safety system. According to reports posted on trainorders this is what happened.....

Switch failure reverse. Maintainer cranked the switch over and spiked and wedged it, but forgot to crank over the movable point frog. Engineer got a 241 past the signal and did not look out for switch not properly lined....
I bet there are several red faces and probably one or two disciplinary actions involved.

Location was Davis interlocking I believe. The derailment was contained on a single track and hence effect on service was relatively less severe.
What is shocking to me is that, apparently, there was no formal checklist procedure for a maintenance operation on 150mph track. Coming from an industry where worker and public safety was paramount, nothing took place on our facilities that was not formalized by a reviewed, written procedure and worker checklist. That way, nothing could be "forgotten." "I've done this a thousand times. I don't need a checklist.", are famous last words before an accident.
As I'm fond of saying; "This is why pilots have a check list. You put the wheels down 100 times in a row for landing on the 101st time, you think that you put them down because you're remembering all the other times and scrape!" "o
 
It's pretty much impossible to that accidentally these days, especially on a passenger plane. The computer would have warned the pilot well before the scrapping occurred.

Is it possible to know, electronically, that the switch and the frog aren't in sync? If so, someone could have known automatically.
 
JIS, also in the trainorders thread:

The cause of this mishap turns out to be a rules violation, not paying attention to the route running at restricted speed.
When putting this with you quote, assuming they are both true, of course, you see a living example of the norm that it usually takes multiple rules violations to create a disaster. We must say "assuming these statements are true" at this time, because we cannot know for sure that they are as yet. They are reasonable and even likely probabilites, but not yet proven.

If the signal maintainer had cranked over the frog, failure on the part of the engineer to observe the switch and frog position would have not caused the derailment. If the engineer had observed the switch and frog positions as running under restricted speed rules required, he would have stopped short of the frog, so no derailment. It took two screw ups to cause the derailment. Note that they were not by the same person, either.

I have seen a turnout with movable frog in UP tracks that had signs on both sides of the track at the switch point warning that the frog must be cranked over when the switch is cranked over.

Is it possible to know, electronically, that the switch and the frog aren't in sync? If so, someone could have known automatically.
Yes. The issue did not apply to this situation as the operation was supposed to be restricted speed, which means by line of sight, making sure everything is as it should be. (The rule gives a list.)

The basic concept is fail safe. That is if anything fails, the system will not give you clearance to use the turnout.

A normal AAR switch machine has three rods: A drive rod, a lock rod, and a detector rod. Some have two detector rods, one for each switch rail. If these three (or four) do not agree, then a clear signal cannot be given across the switch. For turnouts with multiple switch machines, and for crossovers, for the turnouts on both ends, all switch machines must be saying that their switch (or frog) is where it ought to be for whatever the move is supposed to be before a clear signal can be given. This is the meaning behind the common term, "Interlocking" In essence an interlocking system says that conflicting moves cannot be permitted by the signal system, and also that the arrangement of all switches, etc is set so that the permitted movement can be safely done.

The rods do exactly what their names imply. The drive rod moves the switch points. The seperately attached lock rod is moved by the switch. If it is not properly positioned, the lock cannot close, so no clear signal. Once locked, it holds the switch point in place, power or no power, and stops any attempt to move the switch. The detector rod is also moved by the switch rails, and its proper position closes a circuit indicating the switch position. The system is very difficult to defeat. I do not know how, and would not ever say how if I did find out.
 
JIS, also in the trainorders thread:

The cause of this mishap turns out to be a rules violation, not paying attention to the route running at restricted speed.
When putting this with you quote, assuming they are both true, of course, you see a living example of the norm that it usually takes multiple rules violations to create a disaster. We must say "assuming these statements are true" at this time, because we cannot know for sure that they are as yet. They are reasonable and even likely probabilites, but not yet proven.
You are quite correct of course!

If the signal maintainer had cranked over the frog, failure on the part of the engineer to observe the switch and frog position would have not caused the derailment. If the engineer had observed the switch and frog positions as running under restricted speed rules required, he would have stopped short of the frog, so no derailment. It took two screw ups to cause the derailment. Note that they were not by the same person, either.
Good point! And furthermore even after the multiple screwups, the final failure was relatively harmless, since at least the restricting speed part of the rule was followed though with lack of attention to the track condition.
 
Is it possible to know, electronically, that the switch and the frog aren't in sync? If so, someone could have known automatically.
Yes. The issue did not apply to this situation as the operation was supposed to be restricted speed, which means by line of sight, making sure everything is as it should be. (The rule gives a list.)
Perhaps I'm misunderstanding, but if the engineer could have been automatically warned, then, the fact that he did not visually note the situation may have been of less consequence.
 
The proceed through red is adequate indication that something might go wrong. Think how cautiously you'd run a red traffic light you thought was broken.
 
Perhaps I'm misunderstanding, but if the engineer could have been automatically warned, then, the fact that he did not visually note the situation may have been of less consequence.
But my impression is that the automatic warning was disabled by the maintainer while he was working on the switch, hence the need for the engineer to visually confirm the switch position.

EDIT: My impression was wrong. George Harris and Jis explained it better.
 
Last edited by a moderator:
The proceed through red is adequate indication that something might go wrong. Think how cautiously you'd run a red traffic light you thought was broken.
Not really. The train had permission to pass the signal. Otherwise it would not have. He was operating under a restricted speed rule, which has a maximum allowed speed, usually 15 or 20 mph plus a requirement to stop within half the distance you cn see looking for misaigned switch, broken rail, etc., etc.
 
When Rule 241 is invoked that implies that the train has just been allowed to pass a signal that shows red, and has no further automatic protection of any sort until it gets past the section protected by that signal. The rule specifies the restrictions under which the train operates and at what point it can revert back to normal operation.

In some of the responses in this thread I see a tendency among people to believe that automation systems can do more than they actually can. All automation systems have their limits and they are designed to fail to a degraded mode with clear notification. It is absolutely critical for an operator using the automation system to have situational awareness of what automation mode they are operating under, and follow the rules of the mode. Absent that disaster is a likely outcome.

The sort of thing that can happen with loss of such situational awareness is illustrated amply by the crash of AF 447 over South Atlantic. Apparently the PIC (Pilot In Control) was acting as if the plane was operating in Normal Mode, whereas it actually was in Alternate I mode, which does not provide the safety envelope protection that the pilot was depending on to recover from a nasty situation. So the net result was that the plane stalled and fell out of the sky.

Fortunately on railroads there is always a safe "stop" state, which is where the Acela was at Davis. And then it was instructed to follow NORAC Rule 241 to get it across the failed portion of the system, and apparently the invocation of the Rule and exercise of the Rule was not quite according to the Rule. In railroad lingo that is called "rules violation" as George mentioned earlier in this thread.

The proceed through red is adequate indication that something might go wrong. Think how cautiously you'd run a red traffic light you thought was broken.
Not really. The train had permission to pass the signal. Otherwise it would not have. He was operating under a restricted speed rule, which has a maximum allowed speed, usually 15 or 20 mph plus a requirement to stop within half the distance you cn see looking for misaigned switch, broken rail, etc., etc.
This is a critical point that needs to be understood by all. All signals are passed under some prevailing rule. A Green (Clear) signal is passed under NORAC Rule 281. The Rulebook specifies what must be done to comply with the rule. Depending on the situation, multiple rules may be applicable with precedence governing which takes higher precedence. Once one is in Rule 241, one must comply with 241 subsections until the specification of 241 says that you are released from 241 to move onto some other rule.

So in this case if there was no problem and the signal at Davis indicated Clear then the train would have passed it under Rule 281. Because of the problem it first faced a Stop (Rule 292) and then was allowed to proceed under Rule 241. Upon meeting the requirements of 241, had the train not derailed, it would most likely have moved on to Rule 281 at the next block signal.

NORAC experts please feel free to correct me. I am just an amateur trying to explain how things are.
 
Last edited by a moderator:
Jis:

Sounds like a really good explanation to me. However, my expertise, such as it is, starts where the wheel meets the rail and goes down (geographically) from there.

I don't know how it could be made clearer. The situation was nothing like running a red light. It is more like running into pothole because you were not looking for it despite having a warning sign for it.
 
I don't feel like people are responding to what I'm saying.

I understand that the engineer (or should I say, motorman) did not verify the switch as per the prevailing rules. I was asking about a system that would have warned about the switch situation, irregardless.
 
I don't feel like people are responding to what I'm saying.

I understand that the engineer (or should I say, motorman) did not verify the switch as per the prevailing rules. I was asking about a system that would have warned about the switch situation, irregardless.
There isn't.
 
I don't feel like people are responding to what I'm saying.

I understand that the engineer (or should I say, motorman) did not verify the switch as per the prevailing rules. I was asking about a system that would have warned about the switch situation, irregardless.
There isn't.
Hence the signals dropping to the fail-safe "stop" indication.
 
I don't feel like people are responding to what I'm saying.

I understand that the engineer (or should I say, motorman) did not verify the switch as per the prevailing rules. I was asking about a system that would have warned about the switch situation, irregardless.
There isn't.
Hence the signals dropping to the fail-safe "stop" indication.
Exactly! I wonder which part of "no protection in Rule 421" was not understood by sechs. But words are cheap, so repeating a critical point is worth it I suppose. :)

In ACSES radio equipped section it is likely that the train computer receives a radio message from the dispatcher's computer asking it to ignore all warnings until CTC signal from the next section is established and 421 is canceled by a followup message or explicitly by the engineer. This is because any warnings may be spurious and would just gum up the system, when the underlying system is known to be broken. I am not quite sure about the details of how 421 is handled. But in case of Form D radio messages are involved.
 
Status
Not open for further replies.
Back
Top