Good morning Chairman Franks and Members of the Committee. It is a pleasure to have the opportunity to represent the National Transportation Safety Board before your Committee regarding the important matter of railroad safety.

Today, I will limit my testimony to hardware and mechanical issues that affect railroad safety. We understand that there will be an opportunity to discuss human factors and other issues at subsequent hearings.

This discussion of hardware issues is broader than it may at first seem. The Safety Board normally identifies human factors as the most prevalent issue in all transportation accidents. This especially holds true for railroads. However, mechanical (or hardware) issues are also extremely important because they have a dual role. In some cases, a mechanical failure can be the direct cause of an accident. In others, appropriate hardware can be used to prevent accidents. I have examples of both situations to share with the committee today.

The Safety Board’s Most Wanted list of transportation safety improvements contains two items that are directly relevant to today’s proceedings. They are positive train separation and the safety of passengers in railroad passenger cars.

POSITIVE TRAIN SEPARATION

Positive Train Separation (PTS) has been on the Safety Board’s Most Wanted list since its inception in September of 1990. Eight years later, this remains vitally important because tragic railroad accidents that could have been prevented by a PTS control system continue to occur. Roughly 50% of the significant railroad accidents that the Board investigates each year could have been prevented had a PTS system been in place.

Positive Train Separation is a system that backs up the action of the human operator to prevent collisions or overspeed derailments. The need for PTS has been highlighted over and over in our railroad investigations since 1969. The Safety Board has long advocated advanced control systems that will provide PTS and act as a safety net for human performance failures in the operation of trains. These advanced electronic systems can reduce both the number and severity of train accidents caused by the failure of a crew member to take action to prevent a train collision.

The Safety Board for many years was discouraged with the pace at which the railroad industry and the Federal Railroad Administration were developing a train control system that could provide PTS. However, in recent years, there has been activity that may eventually result in the development and utilization of a viable PTS system on our nation's railroads. We are pleased that there have been some important developments including:

• UP/BNSF Pacific Northwest PTS Project: The building of the hardware and software for PTS began in 1995. Deliveries, installations and testing began in 1996 and is proceeding in four planned phases. Eventually, 845 miles of track in Oregon and Washington will be equipped with an overlaid PTS system.
• Amtrak’s Incremental Train Control System (ITCS): Installation of the ITCS wayside equipment in the test bed was completed during 1996, and initial testing of the system began that same year. Production of hardware and software for equipping 71 miles between Kalamazoo and New Buffalo (Detroit-Chicago Corridor) is proceeding. Testing is scheduled to resume later this year.
• UP/Amtrak Advanced Train Control System (ATCS): About 111 miles of track on the Chicago-St. Louis Corridor are part of a four-year project to develop and demonstrate full Positive Train Control in revenue service and interoperability with other train control systems. A contract for ATCS hardware and software is expected with testing to begin in late 1998.
• Amtrak’s Advanced Civil Speed Enforcement System: Testing on the Northeast Corridor is expected to begin this year. A Federal Register notice was published and a public hearing was held regarding this system.
• NS/CSX/CR Joint PTC Project: A vendor was selected in 1997 to provide the first phase design specifications for an on-board platform that is not specific to any PTS protocol. The plan is to develop an economical, interoperable on-board system that features common PTC interfaces.
• Other Activities: Along with the test beds mentioned, other railroads (Alaska, New Jersey Transit, New York City Transit, CSX) have budgeted PTC projects as part of their capital and safety improvement projects.

The Safety Board eagerly awaits implementation of these PTS systems. Until this implementation is complete on all railroads, accidents like those that occurred in Secaucus, New Jersey; Silver Spring, Maryland; Devine, Texas; and Delia, Kansas, will continue to cost human lives.

SAFETY OF PASSENGERS IN RAILROAD PASSENGER CARS

The Safety of Passengers in Railroad Passenger Cars was added after the tragic February, 1996, Silver Spring, Maryland, accident. This collision of MARC commuter and AMTRAK trains resulted in 11 fatalities. As a result of its investigation of this accident, the Safety Board recommended that safety standards be established for railroad passenger cars. The Federal Railroad Administration proposed safety standards for passenger cars last fall. The Safety Board made comments on the proposed rule and is currently awaiting the final rule to be issued.

Simultaneously, the American Public Transit Association is producing safety standards for the commuter rail industry. The APTA PRESS (Passenger Rail Equipment Safety Standards) Committee is updating obsolete standards for heavy rail passenger cars from the Association of American Railroads and adapted them for current commuter rail cars. The standards had been dormant since the early 1970s when most of the AAR’s members stopped running passenger trains.

Progress is being made on passenger car safety standards. It is important to note, however, that developing such standards is not a one-time activity. The passenger car safety standards and the commuter rail safety standards adopted by PRESS need to be continually reviewed and revised, especially in light of lessons that are learned through accident investigation and research. If the FRA and the industry are not willing to periodically modify these standards as needed, then we will end up with yet another set of obsolete regulations and requirements.

LOCOMOTIVE CRASHWORTHINESS

A companion issue to passenger car safety standards is locomotive crashworthiness standards. The NTSB has maintained an interest in locomotive crashworthiness since 1970, when it issued several recommendations to the FRA and the industry to improve locomotive crash survivability. These regulations resulted from an investigation of a caboose over-ride and a crushed locomotive operating cab that occurred during a collision between an Illinois Central freight train and an Indiana Harbor Belt freight train in Riverdale, Illinois. Since that time, there have been many additional recommendations that have addressed such issues as locomotive collision posts and crashworthy cab structures, readily accessible crash refuge (compartments), application of shelf couplers, and improved locomotive fuel tank design.

One of the more recent Safety Board cases to address locomotive crashworthiness, and the issue of fuel tank safety specifically, was the Silver Spring, Maryland, accident where an Amtrak intercity train collided with a MARC Rail commuter train. In that case, the fuel tank of the leading Amtrak locomotive was ruptured by the raking action of the underframe of the leading car of the MARC train during the collision. A large amount of diesel fuel sprayed out and onto the MARC car. The fuel ignited almost instantly and engulfed the car in which 11 persons perished. Among the many recommendations issued by the NTSB as a result of this case, the Safety Board requested that the FRA and industry conduct research to determine if locomotive fuel tanks could be improved to withstand the forces encountered in the more severe locomotive derailments.

The FRA initiated action on this through its Railroad Safety Advisory Committee (RSAC) which organized a Locomotive Crashworthiness Working Group to examine the current status of locomotive crashworthiness. This was based upon a detailed review of crash case histories, review of existing industry and Federal / state regulatory technical standards. It also sought to identify potential non-regulatory alternatives for the improvement of locomotive crashworthiness. The Safety Board’s participation has been limited to discussions of NTSB recommendations, providing historical rail accident case reviews and safety study information, and to conveying insights into NTSB investigative activities.

POWER BRAKE LAW

A discussion on railroad hardware issues would not be complete without a discussion on power brakes and the power brake law. This is an area where the Safety Board believes that the railroad industry could do a much better job. This is also an area where human factors and hardware issues are sometimes directly related.

There are three areas of the power brake law that I would like to discuss today:

• Locomotive Dynamic Brakes,
• Brake Tests and Inspections, and
• Brake Repairs

Dynamic brakes apply braking to the locomotives only, not the cars. They use the kinetic energy of the train to generate electricity through the traction motors, which is then run through resistor grids and dissipated as heat. The kinetic energy used to turn the traction motors retards the movement of the train, causing it to slow or brake. As long as the dynamic braking system works, total dependence on the air brakes and their associated higher fuel use and maintenance can be avoided.

The railroad industry has maintained that dynamic braking is a non-critical feature not required for safety or train control. It contends that the main purposes of dynamic brake use are fuel economy and maintenance reduction. Because regulations require that trains be safely handled with the air brake system alone, railroads do not acknowledge that dynamic brakes have become an important safety and train-handling feature. Actual railroad rules and train-handling routines, however, indicate that, in practice, dynamic brakes have become essential to train handling. The Safety Board has found that railroads are operating trains in situations in which loss of dynamic braking may very well result in loss of train control.

As a result of the investigation of an accident that took place at San Bernardino, California, in May 1989, the Safety Board recommended that the FRA:

The FRA placed the recommended action with its RSAC for handling. The RSAC was unable to develop a satisfactory solution to the problem of providing for functioning dynamic brakes.

Because no progress on Safety Recommendation R-90-024 had been achieved in approximately seven years, the Safety Board concluded that the FRA should separate the dynamic braking function component from the power brake rulemaking process and promulgate regulations to require, on locomotives equipped with dynamic brakes, that the dynamic brakes are functioning properly before trains are dispatched.

The Safety Board also believes that the FRA and AAR should develop procedures to ensure that, before a train equipped with dynamic brakes is dispatched, all the dynamic brake systems on the locomotives are functioning properly. Also as a result of the San Bernardino accident, the Safety Board made recommendations to both the FRA and the AAR that they should:

Study the feasibility of developing a positive method to indicate to the operating engineer in the cab of the controlling locomotive unit the condition of the dynamic brakes on all units in the train. (To the FRA, R-90-023. To the AAR, R-90-026).

The FRA also considered this issue part of the proposed revisions of the Power Brake Law, which, as noted above, have not been successfully advanced.

Despite these recommendations, reliable information on the status of a train’s dynamic braking is still not available to the engineer. Although engineers are instructed to assume that their brakes are operational, they have no means of checking whether they are working properly aside from conducting "seat-of-the pants" tests while running the train. There is no requirement for equipment in the lead locomotive to provide the engineer with information on the train’s dynamic braking status.

In recent years, an effective and reliable device to display the real-time dynamic braking performance of trailing locomotive units has been developed. Such a display permits an engineer to modify his train-handling strategy based on the information it provides, before being surprised by the failure of a dynamic braking system.

The Safety Board concluded that installing a device in the cab of each controlling locomotive to indicate the real-time condition of the dynamic brakes on each locomotive unit in the consist would give valuable information to the engineer on train dynamic braking capability at any given moment. Since no progress had been made by the FRA on Safety Recommendation R-90-023 that addresses developing a positive method to indicate to the engineer the condition of the locomotives’ dynamic brakes, the Safety Board classified R-90-023 "Closed – Superseded." The Board believes the FRA should now require railroads to ensure that all locomotives with dynamic braking be equipped with a device in the cab of the controlling locomotive unit to indicate to the operating engineer the real-time condition of the dynamic brakes on each trailing unit. Further, the Safety Board believes that all railroads should equip all lead or controlling dynamically-braked locomotive units with real-time displays capable of indicating to the engineer the dynamic brake condition on each trailing locomotive unit in the consist.

The Safety Board is also very concerned about brake tests that are performed at a train’s initial terminal and enroute. The Safety Board has found evidence in a number of accidents that the required tests of the brake system are not always performed or are not performed adequately.

The Code of Federal Regulations [(CFR) Title 49, Part 232 - Power Brakes and Drawbars] sets out requirements for the proper inspection of freight trains (freight cars and motive power/locomotives). Two specific sections provide for inspections and testing of air brake equipment by qualified personnel at initial terminal and intermediate terminal (or at a maximum of one thousand miles from initial terminal).

Because of a series of 15 accidents that occurred on Union Pacific Railroad (UP) lines in a one-year period, the Safety Board recently held a public hearing in Springfield, Virginia, to gather facts for its investigations of the accidents.

During the hearing, the NTSB questioned Union Pacific officials concerning causes for the increase in accidents and other related safety problems. Testimony was provided by Union Pacific witnesses regarding the need to conduct thorough inspections of freight cars and locomotives before leaving rail yards and terminals.

During one discussion regarding a head on collision accident in the Houston, Texas area, UP witnesses were asked to comment regarding evidence that indicated that a departing train had received initial terminal inspection lasting less than 10 minutes (event recorder information indicated approximately three minutes). The UP witnesses admitted that such an abbreviated attempt to inspect trains was inadequate, not UP policy, and that a proper inspection could not be accomplished within such a limited time frame. Preliminary NTSB findings indicate that an air brake line blockage within the locomotive consist prevented normal and safe operation of the train by significantly diminishing braking capability. A proper inspection of the locomotives during the air brake and end-of-train-device testing and inspections might have prevented the collision.

As a result of testimony during the hearing, and from evaluations of other investigations (including other railroads) the Safety Board strongly believes that all railroads need to make significant improvement in the inspection of air brake systems. Such improvements must include enhanced training programs for all railroad employees that have responsibility for the inspection of air brake systems components - and all other safety related inspections of freight trains. The NTSB believes it is vitally important during the current power brake law rulemaking process that the emphasis and efforts for strengthening and enforcing rules governing the inspection of train air brake equipment receive significant priority attention.

The Safety Board has recently investigated two accidents where crimped hoses within the train consist prevented the train crewmembers from braking the train properly. In both cases, the hose length that was installed was different in length than was specified by the manufacturer; and most likely the result of an improper repair.

On February 14, 1996, a Burlington Northern Santa Fe Railroad (BNSF) freight train entered the Canadian Pacific Railway System (CPRS) Pig’s Eye Yard, in St. Paul, Minnesota, in a run-away condition. The train collided with three standing CPRS locomotive units and a standing CPRS train. A third standing CPRS train was also struck by derailed locomotive units, A pedestrian bridge and a yard office building were The cost associated with the accident were $3.6 million.

Post-accident investigation revealed that the BNSF train only had effective brakes on the locomotives and first seven cars in the train at the time of the accident. The 7^th car, which was equipped with 15-inch travel end-of-car cushioning devices, had been repaired at the BNSF Northtown Yard before the train departed, and a 60-inch flexible air hose had been applied to the car. The manufacturer of the cushioning device specified that a 39-inch air hose be used with this type cushioning device.

The second accident occurred on March 23, 1998, in Herrington, Kansas. A Union Pacific (UP) train which consisted of two locomotive units and 59 freight cars, was proceeding westward from Topeka to Herrington where it entered the yard in a run-away condition and collided into a standing train. The standing train was parked on the main track and was in the process of being given a 1,000 mile inspection by two carmen. One carman and one train crewmember sustained minor injuries. Damages are estimated at $100,000.

This accident is currently under investigation. However, preliminary investigation revealed a crimped intermediate hose on the lead end of the seventh lead car. The hose length was about 36 inches, and the other end of the car was equipped with a 41-inch hose. Manufacturer hose length specification and repair history from the car owner are forthcoming. A post accident reenactment revealed that as the train traveled down the descending grade with the train in compression, the crimped air hose restricted the air flow to the remaining 53 cars in the train.

Only the lead locomotive unit was equipped with dynamic brakes. When the engineer made an emergency brake application, the dynamic brake was "cut out", and only the lead six cars had emergency braking effort. The rear of the train was equipped with a two-way end of train device, which allowed the train crewmembers to activate the emergency brakes from the rear of the train. Post accident testing revealed that it functioned properly. However, both train crewmembers stated that they did not even think of activating the brakes from the rear of the train.

The Safety Board is on record from the Cajon, California, and Silver Spring, Maryland, accident investigations that address issues identified in these accidents. From Cajon, the Board recommended two-way end-of-train device use and utilization training, and air brake hose configuration. From Silver Spring, the Board recommended that all locomotives be equipped with functional dynamic brakes.

TRACK SAFETY STANDARDS AND INSPECTIONS:

Track Safety Standards and Inspections is another area in which the Safety Board has concern. The current climate of railroading in the United States, with increased car loading and rail traffic, has placed greater demands on rail performance. Railroads today are trying to get a billion gross tons of traffic out of mainline rail. Rail cars weighing up to 315,000 lbs. - over 150 tons - are now common on our railroads. The Safety Board has investigated many accidents where undetected rail defects caused the accident, and has other accidents under investigation where undetected rail defects appeared to have caused the accident. Also, investigators have recently noted in some cases that while some of the defects are relatively small, the railhead is very worn. Thus, rail defect detection is more important than ever before. Constant improvement in testing and analysis of rail condition is critical to prevent the reoccurrence of these accidents.

Some of the more notable accidents the Safety Board has investigated that were caused by an undetected rail defect included:

1. Superior, Wisconsin, June 30. 1992. More than 40,000 people were evacuated when a tank car of hazardous material (mostly benzene) entered the Nemadji River and caused about $1 million in damages.

2. Weyauwega, Wisconsin, March 4, 1996. About 3,155 people were evacuated, seven tank cars of LPG derailed and caught fire, and caused about $19.6 million in property damage.

3. Marshall, Missouri, September 10, 1996. The train was traveling at a recorded speed of 43 MPH in 45 MPH territory when the derailment occurred. Thirty six of the 37 derailed cars were destroyed. There were no injuries, but the costs associated with the derailment exceeded $1.7 million.

4. Lomira, Wisconsin, November 22, 1997. This accident is still under investigation, but preliminary investigation revealed a broken rail with an internal defect found in the wreckage. During the derailment sequence a freight car struck an nearby factory, fatally injuring one employee and injuring four others.

Rail defects can occur as a consequence of the rail manufacturing process, but usually result from fatigue accumulation under repeated loading. The inspection of the rail for the purpose of detecting these defects before they progress to complete failure is addressed in the Track Safety Standards Section 213.113. However, present day regulations and inspection procedures do not facilitate the identification of rail defects under some field conditions. The most insidious and prevalent field condition is shelling, because detection signals generally won’t pass through a shelled railhead.

The Board has made two recommendations to the Federal Railroad Administration (FRA) that specifically address shelling.

R-94-001

Research and develop, with the assistance of the Association of American Railroads, inspection methods that will identify internal defects in rail that has significant shelling and other surface conditions.

R-94-002

Perform the necessary research and develop standards that (1) provide defined limits of allowable rail surface conditions (such as shelling) that can hinder the identification of internal defects, and (2) require remedial action for rail with surface conditions that exceed defined limits.

Advances in rail defect inspection technology need to address the locating of defects that are currently being missed or improperly interpreted by the operators. When most of the available detection equipment tests rail with shelling, the most likely scenario is that shelling is masking a rail defect from detection. However, the rail was inspected as required by the current regulations, but allowed to continue in service for normal operations. Until this inspection equipment technology limitation is addressed, railroads need to limit the catastrophic consequences of accidents caused by undetected rail defects that progress to sudden failure.

GRADE CROSSING SAFETY

The Safety Board has had a long-standing concern about grade crossing safety, which continues today. Since its creation in 1967, the Board has investigated more than 400 accidents that occurred at railroad grade crossings. The Board also has conducted several safety studies on topics related to grade crossing safety. As a result of our accident investigations and safety study results, the Safety Board has issued more than 100 grade crossing safety recommendations.

On March 9, 1997, just over a year ago, a van occupied by a husband, wife, and 3 teen-aged children was struck by a CSX freight train at a passive crossing in Kirkwood, Ohio. The family of 5 was killed. This tragic accident is just one of the more than 4,000 accidents involving motor vehicles at grade crossings that occurred last year. Seventy-five percent of the nation’s grade crossing are passive. More than half of the approximately 500 fatalities that occur at grade crossings each year are at passive crossings.

The Board initiated a passive grade crossing Safety Study in February of 1996. In May of last year, the Safety Board convened a 2-day public forum in Jacksonville, Florida, to gather information about issues affecting safety at passive grade crossings. Witnesses included experts from the railroad industry, law enforcement, research groups, Operation Lifesaver, and Federal, State, and local government agencies. Highway victims of grade crossing accidents and train crews involved in grade crossing accidents testified about their personal experiences. In addition, representatives from Canada and Italy discussed passive grade crossing issues and experiences in their countries.

Our investigative staff has completed the investigations of 60 accidents for this study, and we are currently preparing the report that is scheduled to be presented to the Board this summer.

Some of the issues being developed are: regulatory responsibilities regarding grade crossings; physical characteristics of grade crossings; signage; audibility of train horns; train conspicuity; communication between railroad and highway communities; crossing closures; and, law enforcement at grade crossings.

How do we solve this important problem? The passive grade crossings can benefit from hardware solutions. Elimination would be our first choice. Secondly, we feel that where possible, crossing gates and lights could be added. There is also technology available on the vehicle side. Proximity warning devices may provide the warning necessary to prevent some of these tragic accidents, especially those that occur on hazardous materials carriers and school busses.

Who can forget the tragic photos of the October 1995 Fox River Grove, Illinois, school bus accident? The body of the bus was torn from the chassis and 7 young lives lost. Our investigation found problems with the crossing signal system and the interface with the traffic lights. Even more surprisingly, however, was the fact that it was extremely difficult to hear the train horn inside the vehicle. The Safety Board feels that a good job was done handling the signal issues on the Fox River Grove accident. However, train and school bus collisions continue to occur. In the last few months the Safety Board has investigated two additional school bus train collisions.

A school bus carrying 15 members of the Beesville, Texas high school girl’s track team and three adults was struck by a train in downtown Sinton, Texas at an active crossing. According to the Sheriff’s Office, the bus entered the crossing while the warning lights were flashing and the warning bells were ringing. The train clipped the right rear of the bus. Twelve students and the driver were taken to local hospitals. Eleven of the students and the bus driver were treated and released; one student was held overnight for observation.

Passengers on a school bus in Buffalo, Montana were not as lucky.

On March 10, 1998 at approximately 7:35 a.m., a 48-passenger school bus from the Hobson Public School District, carrying 5 children, was eastbound on Buffalo Canyon Road, near Buffalo, Montana. The bus stopped at the highway/railroad grade crossing with the Burlington Northern Santa Fe railroad tracks. But the busdriver failed to open the passenger door, as required by State law. A 15-year-old male student requested the busdriver to play a cassette tape in the bus’ stereo system. While the bus remained stopped at the grade crossing the student brought the tape forward and the bus driver inserted the tape into the tape player. The music was too loud so he reduced the volume. The student returned to his seat in the back of the bus. The bus’ heater fan was operating. The bus driver proceeded east, across the tracks, and into the path of an on-coming BNSF freight train. The bus was struck on the left side between the rear axle and the rear bumper by the plow on the front of the train. The bus was spun around and struck the train a second time along the left side of the lead locomotive. The cab of the bus separated from the chassis and came to rest on its’ left side. Two children in the back of the bus received fatal injuries. The driver received moderate injuries. Three other children received minor injuries.

Mr. Chairman, that completes my testimony, and I will be glad to respond to any questions you may have.