ORLANDO — According to the Orlanda Sentinel, ballast — the chunks of crushed stone holding the company’s track in place — kept slipping away no matter how much work the railroad did on the short section of track just south of Crescent City.
And maintenance crews worked a lot on that track. They were out in October, November, February and twice in March. At one point, they dumped 90 tons of new ballast and even built a makeshift retaining wall to hold the material in place.
But a month after the March work was complete — and a few hours after the curve was inspected — an Amtrak Auto Train running just below the 60 mph speed limit derailed as it made its way through the banked, 1,400-foot curve. Four people died, and more than 100 were injured. Twenty-one rail cars, including 14 of the train’s 73-ton Superliner passenger cars, left the tracks, many tumbling onto their sides.
The wreck, which National Transportation Safety Board investigators say occurred about 4 feet from the makeshift retaining wall, has resurrected questions about a company whose maintenance has been criticized for years.
“They were really having trouble maintaining the ballast,” said Russ Quimby, the lead NTSB investigator on the derailment, “particularly on the inside of the curve.”
But in trying so hard to shore up the narrow, sloping rail bed, CSX Transportation may have set the stage for the April 18 wreck. NTSB says the railroad never performed a critical step designed to prevent tracks from buckling in extreme heat. CSX says it did — at least twice.
Key to investigation
The procedure — called maintaining “neutral rail temperature” — essentially pre-stresses the steel so that it will not expand too much and buckle when exposed to high temperatures. The engineer of the Amtrak train reported seeing a buckle, or “sun kink,” just before his train hurtled off the track in the piney woods of Putnam County.
The fuzzy, tough-to-explain notion of neutral rail temperature has emerged as a key element in the government’s probe of the Amtrak derailment. Investigators want to know how CSX maintenance might have affected the neutral rail temperature and whether the company ever adjusted it. Investigators have consistently said CSX didn’t perform the procedure, but a company spokesman disagreed.
“I don’t want to get into a dispute with the NTSB,” CSX spokesman Gary Sease I said, “but our folks are confident we made those adjustments.”
Railroads typically recalibrate neutral rail temperature a couple of times a year, but researchers say it should be done more frequently when rails and rail beds are being disturbed. In the months before the Amtrak wreck, CSX crews lifted the track at least four times.
“There’d been no adjustment of the rail temperature,” Quimby said this week. “And if that’s not maintained, that’s when you get track buckles.”
Buckles are a well-known risk associated with the type of track used south of Crescent City. Called continuous welded rail, it has no joints.
Instead, strips of rail are welded together, forming ribbons of steel that can stretch for thousands of feet. Welded rail is cheaper to maintain, reduces wear and tear on rail cars and provides a smoother ride.
But as welded rail expands in the heat, there are no joints to provide breathing room. Instead, the expansion puts pressure on the steel — 2,500 pounds for every 1-degree increase in temperature. And after a day spent soaking up the sun, rail temperatures can reach 160 degrees. (The day after the April 18 derailment, investigators measured the track temperature at 120 degrees, though the air temperature was only 88.)
Eventually, the pressure becomes too much, and the rail buckles, sometimes bowing out by more than a foot. When a train runs over such a buckle, especially at high speed, derailment is almost certain.
Researchers say the keys to preventing sun kinks are proper installation and careful maintenance. Rail must be laid at the so-called “neutral temperature,” one that accommodates expansion and contraction.
Crews installing rail heat the steel and, after it reaches the right temperature, anchor it to the ground. Periodically, they go back and, if necessary, set a new neutral temperature. That’s particularly important in areas of extensive work. Anything that disturbs the rails and rail bed, researchers say, can change neutral rail temperature.
Sun kinks are not inevitable.
“Good track does not buckle!” wrote Andrew Kish, an authority on continuous welded rail who does research for the U.S. Department of Transportation, in 1995. “Sound maintenance, installation and inspection practices will prevent catastrophic derailments.”
Maintenance, though, has been a CSX sore spot for several years. Since the late 1990s, federal investigators have criticized CSX for shoddy work and poor-quality track.
An audit done two years ago concluded that the “number of serious” track defects found by Federal Railroad Administration investigators indicated CSX had “systemic” problems with its inspection program. FRA said many of the track defects were “easily preventable” with better inspections. Just as troubling, FRA found that even when problems were identified, they didn’t always get fixed.
In March of 2000, for example, inspectors in Florida’s Panhandle found 23 spots where the track’s gauge — the distance between its rails — was too wide. All of the sites, the audit said, had been reported at least three weeks earlier but CSX had failed to repair them. Some spots on the CSX system, the report said, appeared as if they “had been neglected for many years.”
Report blames ‘short cuts’
FRA blamed much of the track trouble on CSX’s efforts to save money. Inspection and repair crews were cut by as much as 35 percent in some places, FRA said, creating “an environment where track maintenance forces regularly take short cuts.”
The 2000 audit was stinging but not surprising. FRA had been monitoring CSX since 1997, requiring the company to craft a strategy to combat track problems.
That proved only marginally effective: In 1998, the number of track-caused accidents increased. The next year saw little change.
From the mid-1990s through 2001, track-caused derailments on CSX climbed every year but one. In 2001, derailments blamed on bad CSX track were up 274 percent from the level in 1995, according to FRA data. No other major railroad in the United States did as poorly.
In the past two years, though, FRA has become convinced that CSX has significantly improved its track and track-safety procedures and agreed to end two years of special supervision it had put CSX under in 2000. The decision came two days before the Amtrak crash.
In addition to the track issue, NTSB said recently that it was studying the train’s brakes. The train was equipped with a special supplemental braking system that had been disconnected months before the wreck. The brakes would probably not have prevented the accident, NTSB has said, but they might have reduced the severity of the crash.
NTSB is expected to wrap up its fact-finding for the Crescent City crash this summer. A final report may be released in September.