NEW YORK STATE PUBLIC TRANSPORTATION SAFETY BOARD RAIL SAFETY SECTION PTSB CASES #7367 AND #7373

INVESTIGATION OF TWO DERAILMENTS INVOLVING THE MTA - RAIL ROAD ON #73 SWITCH IN JAY INTERLOCKING, JAMAICA, NEW YORK ON JANUARY 23, 2003 AND JANUARY 25, 2003

SYNOPSIS On Thursday, January 23, 2003 eastbound equipment train #3800 derailed in the Jay Interlocking which is located in Jamaica, Queens, New York. The rear wheel set on the third car and the lead set on the fourth car derailed as they passed over #73 switch. Two days later, on Saturday, January 25, 2003, eastbound equipment train #8738 derailed four wheels on the lead truck of the first coach car (the second car in the consist) as it too passed over #73 switch. Neither train was carrying passengers at the time it derailed. In both cases, at least one car passed over #73 switch before the derailments occurred. Both derailments caused moderate damage to the track and car equipment. The cause of the derailments was not determined until after the second derailment on January 25, 2003 when the investigation discovered that a newly installed switch point on #73 switch's right hand switch point was not manufactured to the correct specifications . The defective switch point and its companion stock rail were installed one day before the first derailment on January 22, 2003 . INVESTIGATION Accident Scene : The Jay Interlocking is located west of , in Queens, New York. The station and interlocking is one of the railroads busiest hubs with all trains routed through this interlocking going to the various branch lines . The speed limit throughout the interlocking is 15 mph. The #73 switch is a left hand, straight split switch with a Sampson undercut point. Number 73 switch rests on a 1 .8 percent descending grade sloping towards the station and is on a 4 degree curve to the left. A super- elevation of one inch exists between the rails of the diverging routes. On January 22, 2003, a work order was completed by LIRR Track Department personal which included replacing the right side sections of the #73 switch including the thirty-nine foot-long right hand straight stock rail and a 16'- 6" right hand curved switch point. Accident Descriptions : First Derailment - January 23, 2003 At approximately 6 :51 a.m., train #3800 made up of eight M-3 type coach cars and carrying no passengers, left from track #17 at Penn Station heading to Long Beach Station. The trip prior to Jay Interlocking was normal and uneventful . Train #3800, traveling east at approximately 13 mph, entered Jay Interlocking on main track #4 and was routed by the operator in Jay Tower via #73 switch to Jamaica Station track #6 . As the train approached the station platform, the engineer

received a derailment hand signal from a track worker in the interlocking and a simultaneous radio message from the Jay Tower operator to immediately stop the train. The engineer applied the train's emergency brakes in response to the communications received. As the B-end truck of coach car #9252 negotiated the left turnout of #73 switch, the wheel flange of the first south wheel climbed the right hand switch point resulting in the wheel riding up on top of the rail . The wheel traveled across the top of the rail for a distance of approximately two feet, then dropped to the south side of the turnout rail, adjacent to the insulated joint at the south heel block, approximately 12 feet past the switch point. The following set of wheels took the same route as the first wheels to derail. Simultaneously, the left hand side wheels of the B-end truck dropped inside the gauge. The B-end truck of the following coach car #9251 also climbed the right switch point causing the wheel to ride up and across the rail, dropping to the south of the turnout at about the same spot as the first truck to derail. The train traveled approximately 150 feet before stopping causing significant damage to the track, railroad ties, third rail, third rail protection boards, switch mechanisms, and signal wires. Car equipment damage included damage to the two trucks, wheels, and undercar equipment of coach cars #9252 & #9251 . The track and third rail damage caused by the derailment was repaired, and the switch and surrounding track was returned to service at 4:03 p .m. that same day. The Railroad estimated that there were over thirty train movements made safely across the #73 switch from the time the new switch point was installed until the first derailment. The coach cars (#9252 & #9251) involved were taken out of service and sent to a Maintenance Facility for further inspection and repairs . The weather at the time of this incident was partly cloudy and cold, with a temperature around 10 degrees . Based on a review of all track measurements, equipment measurements and operating moves, no conclusive decision as to why the train derailed was determined and the track was returned to service after all repairs were made. Second Derailment - January 25, 2003 At approximately 7:14 p .m., eastbound train #873 8, carrying no passengers, left the Morris Park Yard bound for Speonk Station. The train consist was made up of a duel power locomotive (#401) pulling four C-3 type coach cars. The train was routed through Jay Interlocking over switch #73 in a facing move for Jamaica Station track #5 . The engineer was operating from the locomotive and was traveling at approximately 13 mph as the train approached the station. The engineer stated that after passing over the #73 switch, he felt the train shake violently and he immediately applied the emergency brakes. As the wheels of coach car #4005 passed over the switch, the wheel flange of the B-end truck climbed up onto the right hand switch point, similar to that of the derailment two days earlier. Investigation showed that the wheels dropped off the track adjacent to the #2 switch rod, approximately two feet from the end of the switch point. All four wheels of the B-end truck on coach car #4005 dropped to the ground on the south side of the turnout rails. The derailment caused minimal damage to the B-end truck, wheels, drawbar assembly and undercar equipment of coach car #4005. Track damage was also termed as minor. The equipment was sent to a Maintenance Facility for repair. The #73 switch was immediately taken out of service and an in depth investigation ensued. The weather at the time of both derailments was clear and cold, with the temperature around 30 degrees.

Derailment Investigation: The initial investigation after the first derailment on January 23, 2003 did not find a definitive cause for the derailment . Post accident measurements taken by the LIRR Track Department showed that the track gauge through the turnout was within LIRR standards for Class I track. The track profile and alignment leading up to and beyond #73 switch also met with LIRR standards. The Railroad's Track Geometry Car tested the tracks in the interlocking one week prior to the first derailment with no exceptions taken. Since all measurements taken met with LIRR track specifications, and there was nothing which indicated a problem with #73 switch, the investigation shifted from a track related problem to a potential car equipment problem involving the wheels or trucks.

The suspected equipment problem was short lived however and, after the second derailment on January 25, 2003, the investigation shifted back to the track and particularly #73 switch. Wheel markings left in grease that was applied to the right hand switch point after the first derailment indicated that the point of derailment was at the switch point. At this time, a decision was made to replace the switch point and stock rail that was only installed just three days earlier on January 22, 2003 . On January 26, 2003, the recently installed 3 9' section of the right hand straight stock rail and the 16' 6" right hand curved switch point of the #73 switch was removed and replaced. A visual inspection of the right hand curved switch point after it was removed showed that the switch point taper appeared to be different from the other switch point being installed. The point taper appeared too thick, instead of the knife edge taper that was normal for this type of switch installation. The excessive thickness of the switch point caused a narrowing of the gauge at the critical point where the wheels deflect past the switch point on to the diverging route. The narrowing ofthe gauge, along with a combination of track geometry, local gradient, and track super-elevation all contributed to allowing wheel climb on the right hand curved switch point at slow speeds. Split Switch Design: A switch point is used to deflect the wheels of a train from the tracks upon which they are riding to the tracks of a diverging route . Switches are designated either right-handed or left handed, as determined by which way the equipment diverges. The switch points are the movable rails of a turnout which actually cause the train's wheels to diverge. To protect the switch point from wheel batter, the companion stock rail is milled away and the switch point is beveled to fit snugly against the milled area of the stock rail. This style of switch point is known as a Sampson undercut. A bend in the stock rail is made on the turnout side at a predetermined distance in front the switch point. This measurement is referred to as the vertex distance . The vertex distance, as well as the degree of bend in the stock rail, determines the thickness of the switch point. The larger the degree of bend, the thicker and blunter the shape of the switch point must be to fit snugly against the stock rail. Switch Point Procurement: The switch point involved in both derailments was manufactured by Rail Products and Fabrications of Seattle, Washington. The switch point was 16'-6" in length, right hand curved cut from 119 pound rail and milled to match a 39 foot long, undercut, right hand straight stock rail of similar weight. The 3 point was part of a larger order placed in April, 1999 by the Railroad's Procurement Department to the LB Foster Company, who is a major manufacture of switch points and rails in the United States. The order, purchase order #19255, was for 83 various switch point and stock rail packages. LB Foster sub-contracted the order to Rail Products and Fabrications Inc., who was, at that time, a newcomer to the manufacturing of track equipment. Switch points and stock rails are manufactured as matched sets and delivered as fully assembled stock rail/switch point sets, with fixed heel block assemblies. The Railroad supplies the switch manufacture with a manual of drawings, material specifications, references and requirements from which the manufacturer makes the engineering drawings necessary to fabricate the pieces. The drawings provided to the manufacturer for the fabrication of the 16' 6" straight split switch was titled LIRR drawing No. 110-102 and based on American Railway Engineering and Maintenance of Way Association (AREMA) standard plans. The switch point detail on the drawing used to manufacture the switch point in question referred to AREMA design #5100 which called for the thickness at the top of the tapered point to be zero or what is known as "knife edge". The drawing also showed an incorrect vertex distance of 8 and'/4 inches to be used on the bent stock rail. By using this vertex distance, the bend in the stock rail diverged 3/8's of an inch by the time the point matched up with it 8 and'/4 inches from the bend. This design is not compatible for use with the #5100 switch point design. According to AREMA drawings, for point planning, the vertex distance should be zero for the bent stock rail when using switch point detail 5 100 (the rail bend should begin at the switch point). When the vertex distance of 8 and 1/4 inches is used, - AREMA design switch point detail #4000 should have been used which calls for a point thickness of 1/4" at the top of the taper. Rail Products and Fabrications incorrectly manufactured the switch points using the wrong vertex distance with the AREMA design for switch point detail #4000 . The and LB Foster Company were aware of the incorrect vertex distance measurement noted on LIRR drawing #110-102 but both failed to relay this information to the sub-contractor. The correspondence between LB Foster and Rail Products and Fabrications showed that the sub-contractor asked both the LIRR and Foster for clarification on the discrepancy . Although final resolution on this confusion could not be determined from the documentation, it appears that the sub-contractor proceeded on the work order and fabricated the switch point thicker than specified by design to compensate for the bend in the right hand stock rail at the wrong place. In a fax from Rail Products and Fabrications sent to the LIRR in August of 1999, a spokesman for Rail Products and Fabrications states, "this is not a normal AREMA point but exactly per your drawing 110-102" Quality Control: In July of 1999, Steel Services Inc., a firm hired by the LIRR to do their Quality Control, rejected fifty-two, 16'- 6" switch point and stock rail assemblies due to non-conformance with LIRR drawing number 110-102 specifications. The inspection report stated, "the switch pointdesign fabrication for the front partition of the switch point has a clearance round off cut that is different from LIRR plan #110-102" . The point appears to be "fat" because the stock side of the point was incorrectly "widened" to meet the diverging stock rail. This was done so that the switch would maintain the proper gauge at the point. Steel Services Inc . requested that the LIRR make the final decision as to whether or not they should accept this material. 4

On September 21, 1999, the LIRR Procurement Department issued another purchase order (#29553) for the rework and machining of all of the rejected switch point assemblies . The purchase order stated that "this requisition is issued for a change to Order PO#19255. The change calls for a change implemented by the LIRR to correct manufacturing errors in the interpretation of the specifications. Both the Long Island Rail Road and L B Foster have contributed to the need to issue this change order." These modifications were to be completed in accordance with a revised LIRR drawing No.110-102, Rev. 2, dated September 15, 1999. This purchase order was issued with no inspection requirements called for other than a first article sample to be laid out for LIRR inspection. The rejected assemblies were reworked and shipped back to the Railroad with no further quality control inspections made by Steel Services Inc . Additional Facts: In a letter dated September 27, 1999 from Rail Products and Fabrications to L B Foster, Rail Products and Fabrications stated that after receiving the revised LIRR drawing 110-102 Rev 2, dated 9/15/99 for the 16-6' straight switch points, they revised their in-house drawing (#LI 24905 16' 6" RH Switch Point) . At this time, there were two sets of Rail Products and Fabrications production drawings being used for the same RH Switch Point - LIR24905 Rev A, and LIR24905 Rev B. The Rev A drawings reflected the configuration that was produced after they reworked the 52 rejected "fat" switch points. Rev A called for a modification to the vertex distance from 8 and 1/4 inches to 29/32 of an inch. Rev A did not address the extra thickness that was "added" to the switch point to compensate for the extra wide angle created by the 8 and'/4 inch vertex distance or the 1/8t'` inch deeper undercut on the RH stock rail. Drawing LIR24905 Rev B correctly returned the 1/8 inch stock from the rail undercut and incorporated the correct knife edge switch point taper and zero vertex distance. Rail Products and Fabrications stated that Rev B drawings would be used in all future 16'-6" straight switch points manufactured by Rail Science and Fabrications since their drawing, LIR24905 Rev B matches the details now incorporated in LIRR drawing 110-102 Rev 2 dated 9/15/99.

The switch point taper thickness for drawing LI 24905 Rev A is noted as t/4 inch. The switch point thickness for drawing LI 24905 Rev B is 3/32 of an inch (also called a knife edge) . Between January 10 and 24 of 2000, the LIRR accepted delivery on all of the reworked switch point assemblies, even thought the LI 24905 Rev A switch point tapers did not meet AREMA specifications . None of the 52 switch point assemblies that were delivered in January of 2000 were installed on LIRR property until the last quarter of 2002. After the discovery was made that the LIR24905 Rev A switch point taper was manufactured too thick for the 16' 6" switch point installation, the Railroad's Engineering Department ordered its track personnel to locate all non-conforming installations of the switch point sets manufactured under LB Foster purchase order #19255, marked Job No. 24905, dated 6-99. The search included the entire system and the findings were reported immediately to LIRR supervision. The search revealed that six of the incorrectly manufactured Rev A switch point assemblies were put into service with the longest being in use without incident for around three months. Four of these switch points were installed in the Jay Interlocking, including one on #73 switch. All four of the 16' 6'switches installed in Jay Interlocking were replaced by January 31, 2003 . The remaining two switch assemblies were installed 5 on the Main Line; one at the Hooker Chemical siding in Hicksville, NY and the other in KO Yard in Ronkonkoma, NY. The Hooker Chemical siding switch was spiked out of service on January 29, 2003 . The Railroad determined that since the KO Yard installation also had a switch point protector installed that there was no reason to remove that switch from service at this time noting that the point protector was sufficient to prevent any potential derailments. The rest of the Rev A switch point assemblies were located in storage at Belmont Yard and were returned to LB Foster who has agreed to rework the switch points to the proper dimensions and tolerances now set by LIRR drawing 110- 102, Rev 2 . CONCLUSION The Public Transportation Safety Board stafffinds the most probable cause of both derailments at #73 switch was an improperly manufactured switch point. The switch point taper was too thick, which caused a narrowing of the gauge at the critical point where the wheels deflect past the switch point onto the diverging route. The narrowing of the gauge, along with a combination of track geometry, grade and super-elevation contributed to allowing wheel climb on the right hand curved switch point at slow speeds . Contributing to the cause of these derailments was the failure of the Railroad to provide correct engineering drawings to the sub-contractor from which critical measurements and information could be extracted for the manufacturing of the switch components. In addition, also contributing to the cause of these derailments was the failure of the Railroad to properly perform a quality control inspection on materials procured; allowing them delivered to railroad property, put into inventory, and eventually used in service even though they did not meet AREMA specifications RECOMMENDATIONS Based on the findings developed during this investigation, the Public Transportation Safety Board staff recommends that the Long Island Railroad : 7367-1 -Ensure that all engineering documents contain the proper American Railway Engineering and Maintenance of Way Association and Federal Railroad Administration specifications .

7367-2 -Review and update procedures and standards used by personnel in the Procurement & Logistics Department regarding the acceptance of materials.

7367-3 -Replace the defective switch point assemblies installed in KO Yard 2/4 and Hooker Chemical siding. Prepared by: Robert Maraldo

PROPERTY RESPONSE TO RECOMMENDATIONS

The Long Island Railroad responded to the draft recommendations made by the PTSB staff on November 13, 2003 . In their response, the Railroad stated that their Engineering Department would be updating and consolidating all reference drawings and specifications over the course ofthe next 12 to 18 month period as it revises its CE-1 standards manual. By doing so, the Railroad is confident

that all of their engineering documents will contain the proper American Railway Engineering and Maintenance of Way Association and Federal Railroad Administration specifications as requested in Recommendation 7367-1 . As for Recommendation 7367-2, the PTSB staff recommended that the Railroad update and review their procedures and standards used by their personnel when accepting materials according to AREMA standards . The newly formed Materials Group within the Engineering Maintenance of Way section has been assigned this task. In addition, the Railroad also contracts out the inspection of track related materials and inspection reports are regularly forwarded to the Railroad with the proper documentation as required by the Railroad.

Finally, in Recommendation 7367-3, the PT SB staff recommended that the two remaining defective switch points still in service be replaced. The Railroad reviewed each of the two remaining defective switch point locations and decided that since one switch point location is protected by an auxiliary point protector and the other location is spiked out of service, replacement of the points, at this time, was not warranted. Since extra precautions are in place and it is unlikely that the out of service switch will be restored to service, PTSB staff accepts the Railroad's position and will continue to monitor the two locations.

Based on the response supplied by the Railroad to the draft recommendations, the Public Transportation Safety Board staff makes no further recommendations .

SUBMITTAL

This report is hereby submitted by Jerry Shook, Director, Rail Safety Bureau, to the Public Transportation Safety Board for further action.

DATED: October 15, 2003 S ofk, Director Safety Bureau