By Lena H. Sun
Washington Post Staff Writer
Saturday, July 4, 2009
Twelve days after a deadly Metro crash on the Red Line, federal investigators are trying to answer several key questions that will help determine whether the cause was systemic failure in an aging railroad or a "freak occurrence," as Metro officials claim.
Although investigators have determined that a component designed to prevent crashes failed, they have not concluded what caused the June 22 accident that killed nine and injured 80 when a train rammed into an idling one north of the Fort Totten Station in Northeast Washington. The crash was the deadliest in Metro's 33-year history.
"The guiding principle of investigating accidents is that there's never one cause," said Elisa Nichols, a transit system safety consultant. Short of a meteor falling out of the sky and hitting the train, she said, the cause is always a combination of factors: "All the bad things line up at once."
Investigators from the National Transportation Safety Board are examining everything: the condition of the trains, tracks and signals; the actions of the operator of the striking train, Jeanice McMillan, who died in the crash; communication from her downtown supervisors; and the systems that control train movement and are supposed to automatically prevent crashes.
Among the critical questions that remain unanswered:
Is the Wee-Z bond to blame?
On June 17, a Metro crew replaced a device known as a Wee-Z bond, a crucial part of the track circuit system that maintains a safe distance between trains. The device, about 18 inches square and six inches high, is positioned between the running rails. It is a metal container that houses electrical and electronic components and is connected to the running rails by cables. Wee-Z bonds define a block, or distinct section of railroad. Each block contains at least one track circuit. Together with equipment in the rail car, the system is designed to prevent a train from entering a block occupied by another train. The system uses audio frequencies transmitted between the train and the steel rails to note the presence of a train and automatically transmit signals to the next train down the line. If the following train gets too close, the system sends a "zero" speed signal that forces the train to stop.
Shortly after the repair, the track circuit intermittently lost its ability to detect a train. After the crash, that same track circuit failed to detect a test train stopped in the same location as the idling train in the accident. The findings by the NTSB suggest that if the circuit was malfunctioning on the day of the crash, the system would not have detected the idling train and would have sent a "clear" signal to the striking train. Onboard computers would have set the train to 59 mph, the speed limit along that stretch, and into the path of the idling train.
The Wee-Z bonds are manufactured by Union Switch & Signal, Metro officials said. A spokesman for Ansaldo STS, the parent company of Union Switch & Signal, said he could not discuss the Wee-Z bond or the company's contract with Metro because of the ongoing investigation. Metro is replacing the devices throughout the railroad because many are approaching the end of their usefulness, officials said.
Investigators have stopped short of saying the malfunctioning circuit caused the crash.
Transit officials would not say whether they believe the malfunction in the track circuit was a result of faulty equipment or poor installation.
Is the work crew that made the repair to blame?
Metro rail chief Dave Kubicek said that the crew tested the equipment and that "everything tested okay upon installation." Transit agency procedures typically require this type of work to be checked on-site by a supervisor.
What caused the track circuit to malfunction?
Investigators don't know if the Wee-Z bond or some other component of the track circuit caused it to perform poorly. Metro General Manager John B. Catoe Jr. said investigators and transit agency crews know the problem is in the track circuit but have not been able to isolate the cause.
"We could just replace the parts, but we need to understand what caused it," Catoe said in a statement Thursday. "You don't just change the parts. We must find the cause."
Federal investigators are collecting data to analyze each component in the train detection system, including the Wee-Z bonds and trackside cables.
After the repair and before the crash, did the malfunctioning track circuit fail to detect other trains?
Metro officials would not say whether other trains experienced similar trouble along that stretch of track. Kubicek said Metro did not know the track circuit was only intermittently detecting a train -- "fluttering" on and off -- until after the accident, when officials ran analytical reports that showed the pattern.
Federal investigators are reviewing the performance of the track circuit before and after the equipment was replaced June 17.
When did McMillan see the idling train before she activated the emergency brakes?
Federal investigators say that remains unclear. Her train was operating in automatic mode, its movements controlled by onboard computers and a system in the downtown operations center. Investigators intend to perform tests over the July 18 weekend to establish whether the curve in the tracks, or anything else, obstructed her view.
Why was the operator of the idling train in manual mode?
Trains are supposed to operate in automatic during rush hour, when the crash occurred. Investigators say the operator of the hit train was operating manually during his entire shift. He has been a train operator for more than seven years. Train operators say they switch from automatic to manual if trains are experiencing brake problems or not operating smoothly; manual provides more control, they say.
Operators are required to get permission from downtown controllers to change to manual. If the operator had changed to manual only when he entered the crash area, it might suggest that something was amiss, Metro sources said.