Metro signaling system continues to malfunction

By BRIAN WESTLEY – 1 day ago

WASHINGTON (AP) — Signaling equipment that is supposed to detect stopped Metro transit trains continues to fail periodically in the area where a deadly crash occurred in Washington, officials testified Tuesday.

National Transportation Safety Board member Debbie Hersman told a House subcommittee that investigators are replacing various pieces of equipment in an attempt to stop the problem from occurring on a portion of the red line near the Maryland border.

So far, she said, nothing has worked.

"It's a mystery as to what's going on here," Metro board chairman Jim Graham told lawmakers.

Nine people were killed and more than 70 injured June 22 when a Metro train slammed into another train stopped on the tracks.

Metro's signaling system is designed to prevent crashes by generating speed commands and not allowing more than one train to occupy a section of track. But problems began occurring after a piece of equipment was replaced five days before the crash.

Investigators are continuing to examine how the system was functioning at the time of the accident. If the system was malfunctioning, the oncoming train could have lacked information that another train was stopped on the tracks.

Investigators have said there is evidence the operator of the train that was moving, Jeanice McMillan, applied the emergency brake before the crash.

McMillan was among those killed in the accident.

Metro General Manager John Catoe said Tuesday that trains continue to run manually systemwide until the cause of the crash is determined. Added precautions are being taken near Fort Totten, where the crash occurred.

"Operators have to go at slower speeds and need permission from the radio control center to proceed," Catoe told reporters. He said that is causing ongoing delays on the red line, and he asked passengers for patience.

Tuesday's hearing was held one day after the NTSB urged Metro to upgrade its signaling system with continuous backup protections. Investigators want Metro to evaluate the system in real time so that the operations control center is immediately alerted about a problem. Trains would then be stopped or slowed to prevent a crash.

"You've got to get an alert when something fails," Hersman said. "We want to make sure that when the system isn't functioning as it was intended, there's some way to find out about it so we can intervene."

Metro officials planned to meet Wednesday with vendors for help developing such a system, Catoe said. For now, Metro says it is only able to review data once a day to determine whether the system is functioning properly.

At the NTSB's request, the Federal Transit Administration has sent a letter to transit systems nationwide asking that the agencies also make sure they have adequate backup protections in case their train control systems fail.

The American Public Transportation Association says technology similar to Metro's is used in other big cities, such as Atlanta, Baltimore, Boston, Miami, Philadelphia and San Francisco.

Copyright © 2009 The Associated Press. All rights reserved

v

From: Harvey Kahler <harvey.kahler@...>
Subject: Re: [CHICAGOTRANSIT] Re: Automatic Train Operation [1 Attachment]
To: CHICAGOTRANSIT@yahoogroups.com
Date: Thursday, July 16, 2009, 12:23 PM

 

I've followed a parallel discussion on Trains discussion site. The audio frequency block signal system employs the basic railroad-type electro-mechanical track circuit relay; but uses electronics instead of relays for "codes." While blame has been assessed to a relay failure, the logic escapes me. If the struck train had braked to a stop in the block approaching the occupied block at the station, the following train with ATO getting no code should have begun braking before the operator put the train into an emergency stop.  A de-energized on-board, not wayside, relay that failed to break contact for mechanical reasons could be the culprit. If the struck train had pulled up to its leader at the station, the approach block would continue getting no code to clear the restictive indication.  The following train would get the same restictive signal to brake to a stop before entering the occupied block. If the struck train failed to shunt the approach block track circuit, the following train would get the same restictive signal to brake before the operator put the train into an emergency stop, and would have done so at less than track speed.  A de-energized on-board, not wayside, relay that failed to break contact for mechanical reasons, "sticking," could be the culprit. I got around to diagrams of fault conditions that are attached.  At least some of you should get this. --- On Wed, 7/15/09, icrelic <icrelic@yahoo. com> wrote: From: icrelic <icrelic@yahoo. com> Subject: [CHICAGOTRANSIT] Re: Automatic Train Operation To: CHICAGOTRANSIT@ yahoogroups. com Date: Wednesday, July 15, 2009, 11:29 AM   --- In CHICAGOTRANSIT@ yahoogroups. com, "Adam H. Kerman" <ahk@...> wrote: > > At 8:44am -0500, 06/25/09, Dennis McClendon <dennismcclendon@ ...> wrote: > >Larry M wrote: > > >>Anytime you have something like that that operates less than 100 > >>percent you are putting human lives into harm's way. Anythine less > >>thatn 100 percent is totally unacceptable. > > >Surely you're not saying that manual train operation would be better! > > Larry's right, Dennis. Control systems in a railroad environment must be > robust and reliable. False information is worse than no information. > > I changed your Subject to Automatic Train Operation, which is what WMATA > calls its signal and control system, although I think the term is being used > in a proprietary manner. > > I discussed this with a friend who had a bit of information. When operated > manually, the speed limit is 15 mph. When I first heard about it, it was > clear that the following train had collided at speed, not at 15 mph. > > The system reported that the block was unoccupied. When the operator saw the > obstacle, she attempted to take the system out of ATO so she could apply the > emergency brakes but couldn't disengage. > > On a steam road, shunted track circuits are used to detect the presence of > train in the block. The wheels create a shunt. Only an energized track > circuit reports the ABSENSE of a train, so when it fails, the block must be > assumed occupied. > > ATO relies upon relays on board to report the train's presense. If this > fails, the block is assumed to be unoccupied. > > Rohr told WMATA that the relays were good for 70 years, even though a > railroad is a harsh environment and electrical and electro-mechanical > components are never known to last that long. Rohr, itself, wasn't in the > signalling business that long, and having gone out of business... > > The relays on the original fleet were problematic from the beginning and > have been known to require replacement for a number of years. WMATA didn't > have the money to replace them on the oldest equipment, and kept operating. > > In addition, WMATA operates longer trains today than the system's original > design assumed. This changes the braking profile. Was the system updated > with newer braking profiles? This may have been another contributing factor. > > CTA doesn't use this signalling system. But BART and MARTA do. I don't know > how similar it is on the other two systems. > > Now, everyone pretend Al replied and told me I everything I just said is > wrong. > Don't worry I happen to know he's on the other side of the world talking about that subject that seems not to be allowed here. You missed him at the APTA workshop on that topic at the Hilton last month!