Train derailment points to need for research into railroad safety

Even in the age of computer-engineered sophisticatry, train derailment is a problem that still besets the railroad industry. But on the horizon are innovations that may help keep more trains on track in the future.

The problem was once again highlighted Saturday when a mudslide tore off a 500-foot section of railroad track near Williston, Vt., causing the crash several hours later of an Montreal-bound Amtrak Metroliner. Five people were killed.

''It was a once-in-a-lifetime event,'' says Thomas Simpson of the Federal Railroad Administration (FRA). Nevertheless, Saturday's incident has sparked calls for mandatory weather monitoring by rail carriers.

Officials of Central Vermont Railway, which owns and operates the track, apparently ''weren't aware of the weather conditions,'' says Mr. Sheehan.

The FRA, which establishes and enforces safety regulations in the rail industry, does not currently have a regulation requiring train operators to monitor weather conditions.

Two similar incidents in 1982 prompted recommendations from the National Transportation Safety Board (NTSB) that Amtrak require companies from which it leases track to monitor a weather service. Yet both Amtrak and Central Vermont have disavowed any responsibility for monitoring local weather reports.

One FRA official said he thought mandatory weather monitoring regulation would soon be formulated, thus preventing such confusion in the future. ''It certainly could have prevented this accident,'' added Ira Fulman of the NTSB.

But most derailments have nothing to do weather. Last month, for example, another Amtrak train skipped off the tracks near Philadelphia. In May, 11 cars of an 83-car Union Pacific freight train spilled into Columbia's Oregon River. Of 3,906 train accidents reported to the FRA last year, 3,004 were due to derailments.

Usually, it's a bit of broken track or some unexpected warp in the rails that sends a train bouncing off the rails. So emphasis is being placed on inventing more effective ways to identify flaws in railroad track.

For example, railroads measure the stress on rails caused by heat and pressure with so-called strain gauges permanently clamped at strategic locations on a rail. Modern welded-steel rails, while stronger than the old kind that were bolted at the seams, do not allow much room for natural flexing. If too much heat is applied to a rail segment, it can warp.

The problem with the strain gauge system is that it is not mobile - an important consideration with the US's 278,000 miles of track. So transportation engineers at Texas A&M University have developed a portable device that uses ultrahigh-frequency sound waves to spot buckles and cracks in railroad tracks.

And researchers at the University of Houston are working on a device that will check to ensure that the ballast tucked underneath the track ties to secure them is of consistent density. If the ballast is too loose, then the tracks could shift in their beds. If it is too densely packed, the rails might not have room to naturally flex with pressure and temperature changes.

The current method of checking is to eyeball the rail bed - just as it has been done for over 150 years. The new device would ''listen'' to rail ties' resonance when tapped with a plunger to assess their density.

But these projects are not being carried out in conjunction with the industry - a circumstance which the researchers would like to alter. ''No doubt, the industry can be called research-poor,'' says Edward Reinschnidt, research director at the Association of American Railroad's (AARR) Track Testing Site in Chicago.

Deregulation, he explains, has put the railroads at cross-purposes as they try to hold down current costs while acknowledging the importance of basic research.

''But I think it's slowly changing,'' he adds, pointing to a research association the AARR formed for the first time last year with four universities.