By Lena H. Sun and Lyndsey Layton
Washington Post Staff Writers
Thursday, July 2, 2009
Five days before last week's deadly Red Line accident, a Metro crew replaced a key piece of equipment designed to prevent crashes, but the circuitry malfunctioned and no one at Metro detected the problem, investigators and transit officials said yesterday.
The findings raise new questions about whether Metro officials should have discovered the hazard before one train rammed into another June 22, killing nine and injuring 80. It also puts a spotlight on Metro's maintenance crews and the design of a highly automated subway system that is supposed to be "fail-safe."
Transit officials would not say yesterday whether they believe the malfunction was a result of faulty equipment or poor installation, citing the investigation by the National Transportation Safety Board.
In the aftermath of the crash on the Red Line between the Takoma and Fort Totten stations, Metro officials analyzed track circuit data and found that one circuit in the crash area intermittently lost its ability to detect a train. The circuit would report the presence of a train one moment, then a few seconds later the train would "disappear," only to return again.
The problem started shortly after June 17, when a Metro crew replaced a device known as a Wee-Z bond, a crucial part of the system that maintains a safe distance between trains, said Dave Kubicek, Metro's rail chief.
Instead of completely failing, the track circuit "fluttered" on and off so quickly that, Kubicek said, the failure would not have been obvious in Metro's downtown operations center, where controllers monitor real-time movement of trains by watching an illuminated graphic depiction of the 106-mile railroad.
"It was happening so fast, you would just blink and miss it," he said. "Realistically, you had to be looking at the exact area at the exact place" at the exact time.
A controller would have to be staring at something the size of "a button on a BlackBerry" to detect the malfunction, he said, adding that Metro did not realize that there was problem until officials began examining data after the accident.
Under normal conditions, if a track circuit goes "dark" or stops working, downtown controllers will see an indicator change colors on the illuminated screen before them. In addition, if the circuit stops working, the adjacent track circuits will automatically force approaching trains to stop before they reach the "dark" stretch. As a backup, train controllers can also intervene to redirect train movement.
But the "fluttering" was so fast and subtle that none of those auxiliary safety measures were activated, Kubicek said. He stressed that the malfunction was an "anomaly" and that the rail system is safe. He said Metro is testing daily to determine whether this "fluttering" is occurring elsewhere. Metro workers have been inspecting all of the transit system's nearly 3,000 track circuits. They have checked more than 65 percent of them and found no problems, Kubicek said.
"From what we have discovered so far, it appears to be a freak occurrence," said John B. Catoe Jr., Metro's general manager.
The Frederick News Post first reported on the Wee-Z bond issue Friday.
Metro officials would not say whether in the days before the crash, other trains experienced trouble along the stretch affected by the malfunctioning circuitry.
NTSB officials said they are reviewing the performance of the track circuit before and after the equipment was replaced June 17.
Metro's rail system is divided into blocks and is designed to keep at least two blocks of distance between trains to prevent a crash. Each block contains at least one track circuit that detects the presence of a train using audio frequencies transmitted between the train and the steel rails.
Within each track circuit are two devices, the Wee-Z bonds, which are about 18 inches square and six inches high and are often mounted on the wooden cross ties that secure the rails. The Wee-Z bonds 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 Wee-Z bond sends a "zero" speed signal that forces that train to stop.
After the installation, Metro crews tested the equipment. "Everything tested okay upon installation," Kubicek said.
Metro is replacing Wee-Z bonds across the railroad because many are approaching the end of their usefulness, said David Couch, who is in charge of Metro's infrastructure projects.
In last week's crash, the train heading toward Fort Totten slammed into the train in front of it, which had been idling on the Red Line outside the station, waiting for a third train to depart. NTSB investigators said the track circuitry did not detect the presence of the idling train. That would mean the striking train received a "clear" signal and its onboard computers would have been automatically set to travel at 59 mph, the speed limit along that stretch.
Federal investigators say the emergency brakes on the striking train were engaged, but it is unclear at what point the novice operator, Jeanice McMillan, saw the idled train before she deployed them. She was among the nine who died in the crash, which was the deadliest in Metro's 33-year history. NTSB investigators intend to perform tests over the July 18 weekend to establish when McMillan would have been able to see the train ahead of her and measure that against the point where the brakes were deployed, about 425 feet before impact.
Many portions of Metro's automated train control system rely on original equipment based on 100-year-old relay technology that transit officials would like to replace. As part of the agency's major capital improvement starting in 2011, officials want to upgrade track circuits.
According to an internal Metro report assessing the reliability of various track systems, problematic track circuitry stood out in the fiscal year ending in June 2008. Of 668 incidents that caused delays last year, track circuits accounted for 337, or more than half. The station with the highest number of track circuit delay incidents was the Takoma station.
Kubicek declined to comment, saying he had not seen the report. "We have lots of circuitry, lots of equipment," he said.
The problem circuit in the crash area has been disabled but not replaced because officials are trying to pinpoint what caused the malfunction. As investigators continue to examine the signal system at the scene of the crash, Metro officials have ordered an "absolute block" through the area, which means that Red Line trains must proceed one at a time through the affected stretch. That has considerably slowed speeds. All trains have been operating manually on all lines since the crash, and Red Line's top speed limit is set at 35 mph, which has also created significant delays.
In a statement yesterday, Catoe apologized to riders for the inconvenience but said, "This is critical to gaining a full understanding of why this happened."
Meanwhile, Carolyn B. Jenkins, mother of crash victim Veronica DuBose, filed a wrongful death lawsuit in D.C. Superior Court against Metro and Alstom Signaling, her attorney, Stephen D. Annand, said yesterday. The complaint alleges negligence on the part of both parties. It is at least the third claim filed.