Train Crash in India 1

[dik]

New...

"The death toll in the train crash in India has risen to 120 people according to the Hindustan Times this evening, Friday, June 2.

More than 800 passengers are said to have suffered injuries. A huge rescue operation is expected to continue through the night with many people still feared trapped in the derailed cars."

https://euroweeklynews.com/2023/06/...80-injured-after-horror-train-crash-in-india/

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[dik]

update... more than 230 fatalities.

https://www.bbc.com/news/world-asia-india-65793257

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[dik]

Death toll now 288. This is a bit of an explanation about what happened.

https://www.bbc.com/news/world-asia-india-65796173

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[spsalso]

"Coromandel Express somehow ends up on loop line and hits goods train"

The "somehow" is pretty clear, I think. The switch was thrown for the siding, and the train took it and rear-ended the stopped freight.

It could have been "thrown" due to a mechanical failure, but the odds are low. But existent. And there's always vandalism/sabotage.

The freight most likely had aken the siding to clear the track for one or more following trains. Such as this one. Of interest is whether the track switch is remotely operated or manually operated. Also of interest is the type of signaling and occupancy detection and the level of interlocking.

I do recall reading an account of a similar rear-ending: The brakeman (in the US), as the following train approached, thought he might have forgotten to throw the switch. He ran back and threw it just before the train got there. Unfortunately, he had thrown it correctly the first time. In his panic, he mistook the track alignment (which was correct) as being in error, and threw the switch.


spsalso

 

[Murph 9000]

It could have been "thrown" due to a mechanical failure, but the odds are low. But existent. …

The odds should be low if the track infrastructure is properly maintained, but it does happen:

RAIB Report 20/2008: Derailment at Grayrigg

The above report contains plenty of technical detail for those unfamiliar with the way railway points work (there may well be some design differences between the British points in the report and the Indian points in the incident, but the basics should be quite similar).

 

[dik]

Thanks, Murph...

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[spsalso]

Murph,

Thanks for that paper. Very interesting. I tucked it away for later.

Currently, the Indian government seems to be saying the Coromandel Express's engineer saw a "proceed" signal, and was doing so when passing that signal. And also that the signal changed after he/she passed. This fits well with the track switch being thrown just before the train arrived.

A single loose point would normally never change a signal indication, as noted above as happening. But there MAY have been some other mechanical failure.

One can ask how it is they know what the engineer saw. Is this person lucky enough to still be alive? If they are asserting the signal change without his/her input, I do wonder how they know that happened. If they DO keep careful record of signal indications, that should also reveal how it is the signal changed. Signals are not known for being willfully driven to changing their indications on their own.

The authorities are saying there was a failure in the signal system. That system does not throw track switches; it reacts to such an event. The true failure was having a track switch thrown for a siding that contained a stopped train.


spsalso

 

[JohnRBaker]

Despite my many trips to India, I never had the pleasure of riding on a train. Planes, cars, taxis, even a bus once, but no trains. Now. I've ridden trains in Japan, China, England and all over Europe (and of course, recently here in the US).

John R. Baker, P.E. (ret)
Irvine, CA
Siemens PLM:

The secret of life is not finding someone to live with
It's finding someone you can't live without

 

[FacEngrPE]

Depending on hoe the signaling system is defined, a false command issued to a switch engine might be called a signal system fault.

I was looking for information related to how India handles railway accident and found the INDIAN RAILWAYS ACCIDENT PROTOCOL (attached). This document indicts that the "The Commission of Railway Safety" is charged with accident investigation. Looking around it is not obvious if the accident reports are publicly released.

https://sagalnews.com/2023/06/04/india-train-disaster-signal-fault-the-likely-cause-minister-says/[/URL]]Ms Verma Sinha said there was “no issue with the electronic interlocking system” and said investigations indicated “some kind of a signalling interference” rather than failure.

A confusing statement.
Quoted link has the most complete description of the technical aspects of the accident to date.

 

[Murph 9000]

Currently, the Indian government seems to be saying the Coromandel Express's engineer saw a "proceed" signal, and was doing so when passing that signal. And also that the signal changed after he/she passed. This fits well with the track switch being thrown just before the train arrived.

With a typical automatic signalling system, the block detection would reset the signal to red/danger moments after the leading vehicle passed the signal. That's normal behaviour, if that is what is being described. The block detection / track occupancy sensors should have combined with the interlocking to prevent a green/proceed signal aspect while the points were set for the goods loop, as well as preventing the points from being changed until the passenger train had cleared the section.

If it was a manual signalling system, with no occupancy detection, it's possible for the signaller to do something really bad by throwing levers early. The reference to "electronic interlocking" makes me think it's an automated signalling system, which should prevent this type of incident if everything is working correctly.

 

[zeusfaber]

"Signal went red in my face" is eerily reminiscent of the 1988 Clapham Junction crash. That one was caused (in big handfuls) by defective logic in the track circuit element of the interlocking.

A.

 

[Murph 9000]

Investigation into the Clapham Junction Railway Accident (aka the "Hidden Report") for anyone who wants to read the detailed investigation and analysis.

 

[spsalso]

Also of interest is the Amtrak crash in Niles, Michigan in 2012:

https://www.jonroma.net/media/rail/accident/usa/ntsb/RAB1306.pdf

While there may well have been a failure in the signal system, there remains the question of how and why the track switch was in the diverging position. The signals did not MAKE it be in that position because they cannot. A person put the switch in the diverging position (unless there was a mechanical failure).

Also, we likely do not and can not know the signal indications the engineer saw, unless there was a recording forward-looking camera that survived the crash. It's unlikely the engineer did.



spsalso

 

[Murph 9000]

While there may well have been a failure in the signal system, there remains the question of how and why the track switch was in the diverging position. The signals did not MAKE it be in that position because they cannot. A person put the switch in the diverging position (unless there was a mechanical failure).

With an automated signalling system (implied by the mention of "electronic interlocking"), points/switches on the main line (i.e. outside of yards) are typically controlled by the signalling system. The points should have detection on them which feeds into the interlocking, preventing a green/proceed signal if they are not set and locked correctly.

I guess it's possible the signalling system is both electronic and a bit simpler than we have in Europe (not counting the 21st century stuff that's replacing lineside signals with magic boxes on the trains communicating via digital radio, e.g. ETCS).

 

[waross]

It could have been "thrown" due to a mechanical failure

OR
It could have been "NOT thrown" due to a mechanical failure.
I am thinking a pin or bolt may have dropped out of the linkage.
The order was given to change the switch.
The automatic operator cycled as it should and reported that the switch was changed.
Due to missing pin in the linkage, one or both points failed to move when the operator cycled.
Just thinking about possibilities, and remembering machine elements that were not where the control logic thought they were due to a missing bolt or pin.
Or something else.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[waross]

Another issue that I have seen more than once is sticking limit switches.
The switch points move properly but the stuck limit switch does not report the position correctly.
With the vibration of the approaching train, the limit switch makes its move, just as the engine had passed the signal.
Or something else.
These are suggestions, not solutions that I will strongly defend.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[waross]

This allowed the left-hand switch rail to move into an unsafe position close to its stock rail, without losing signalling detection.

That's what I was trying to say.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[spsalso]

There is no detection for points position. By that, I mean that there is no sensor on each point that has an input to the signaling sysem.

If a point moves to an unsafe position, that fact is not input into the system. As far as I know.

In an integrated signaling system, it is assumed that the track switch position is correctly input to the signaling system. This input starts as a position switch linked to the throw bar.

It is certainly possible for the throw bar to be in a specified position without the points being in THEIR specified position. This would be a failure of the mechanical linkage, which has been commented on here previously.


If there is a true points position switch, I am unaware of such a thing. It does sound like a good idea--sort of the "last mile" of signaling safety.


I look forward to an investigation of this event by the Indian authorities. In particular, I look forward to the release of facts of the matter. I note that the Greek investigation of a similar event on February 28, more than three months ago, has revealed next to nothing. I believe that will be reflected in this case. And for similar reasons.


spsalso

 

[Murph 9000]

British points on the mainline have detection which feeds back into the interlocking. See various discussion of the detector rod in the Grayrigg report (it had become detached from the switch rail). I'm not certain if Indian points are built to the same standard.

 

[Alistair_Heaton]

They will be at the standard that was in force when India became independent.