Angels Flight Accident


Figure 1: A view down the Angels Flight incline to the location at which Sinai (left) and Olivet (right) came to rest. Note the slack wire rope cable leading to Sinai, whereas the wire rope cable to Olivet is taut.	Figure 2: A view up the Angels Flight incline to the location at which the cars came to rest.

Figure 3: The impacted cars on the Angels Flight incline (Olivet is at left).

Shortly after noon on February 1, 2001, an accident occurred during the operation of the Angels Flight funicular railway when a passenger car (named Sinai) broke loose, rolled down the inclined tracks and impacted the other car (named Olivet) that had been brought to a halt by the inclined railway’s braking system, Figures 1 through 3. Unfortunately, as a result of this accident, one passenger was killed and seven others suffered injuries.

Immediately following the accident the National Transportation Safety Board and several other entities began to investigate its cause. There were two major questions regarding the accident scenario that had to be answered: 1) what occurred that allowed Sinai to separate from the Angels Flight control system; and 2) why didn’t the emergency braking system prevent Sinai from rolling down the incline.

The Consultants’ Bureau was asked to provide expertise to assist one of the entities involved with Angels Flight. Our team was comprised of Dr. Bill Jones (mechanical engineer), Mr. Norm Schutzberger (electro-mechanical-hydraulic systems), Dr. Larry Kashar (metallurgy and materials science) and Mr. Chip Pedersen (graphics & animation).

Angels Flight is a historic funicular railway in downtown Los Angeles that was originally built in 1901 to move passengers between the commercial district at the bottom of Bunker Hill and the residential area then at the top of the hill. It was dismantled in 1969, but the original cars, named Olivet and Sinai, and other equipment were placed in storage. The funicular was rebuilt close to its original site and was re-opened in 1996 using the original two cars still with wire-rope haul cable, the same track layout, but a different haul drive system.

The investigation at the scene showed that the wire-rope cables controlling the movement of both Sinai and Olivet had not broken nor detached at either end, but the cable to Sinai had almost completely unreeled from its cable drum, indicating that the drum had been free to turn, see Figure 4 and the drive system schematic in Figure 5.

Figure 4: The small amount of wire rope cable remaining on the drum for the Sinai car after the accident. The arrow shows the location of the emergency brakes on the Sinai cable drum.

Figure 5: A schematic of the Angels Flight drive system.

Disassembly of Sinai’s cable drum from the gear box showed that its female spline which transmitted the rotational motion from the gear box to the cable drum had suffered a massive failure of its splines, Figures 6 through 9. The loss of these splines explained why the motion of Sinai was not controlled by the gear drive system, but did not explain why the spline failure had occurred or why the emergency brake system failed to activate and prevent Sinai from accelerating down the incline into Olivet.


Figure 6: The inside of the female spline gear (split in half) from the Sinai drive train system.	Figure 7: Detail of the failed Sinai female spline gear showing a major amount of missing splines and a small area of intact splines that had not been in contact with the mating male splined shaft.

Figure 8: The cross-section of a portion of the intact female spline from the Sinai drive system, showing the shape of the as-machined spline teeth.	Figure 9: The cross-section of a portion of the failed female spline from the Sinai drive system, showing the remaining “nubs” of the spline teeth.

There were two independent braking systems used on Angels Flight: a) the service brake, mounted next to the electric motor, used to stop and hold the cars in the stations at the end of each trip, and b) the emergency brakes, mounted on each car’s cable drum. In the days immediately following the accident, it was found that the service brake system worked as designed, but that the emergency brake system did not operate at all and, based on the oil film on the emergency brake rotors, had not operated for some time. Follow-up investigation found that the solenoid valve used on the emergency brake system had a burnt-out valve activation solenoid coil mounted on a normally-closed valve, rather than the required normally-open valve (see the brake system schematic, Figure 10, and photos of the system equipment, Figures 11 through 13). This resulted in the valve staying in the closed position, unable to relieve the hydraulic pressure in the emergency brake system, causing the brake calipers to remain open.


Figure 10. Schematic diagrams of the hydraulic systems for the Service Brake (top) and Emergency Brake (bottom).

Figure 11. The equipment cabinet housing the hydraulic controls for both braking systems on Angels Flight: at top, the Service Brake system; at bottom, the Emergency Brake system. This photo was taken several days after the accident when the Emergency Brakes had been activated for testing. Note that when the one of the brake systems is activated, there should be no hydraulic pressure in that system.	Figure 12. A more detailed photo of the Service Brake hydraulic system, taken several days after the accident. The arrow denotes the accumulator in this system. The solenoid valve is obscured by the accumulator.

Figure 13. A more detailed photo of the Emergency Brake hydraulic system,taken several days after the accident. The arrow denotes the solenoid valve for this system.

In addition, our examination showed that the solenoid and valve did not fit properly and had actually been forced together. We contacted the manufacturer regarding this observation and learned that in 1997 the solenoid valves had been redesigned such that the component parts produced after the redesign were not interchangeable with those produced prior to the design change. Our research of the identification numbers on the solenoid coil and on the valve showed that the solenoid coil had been manufactured prior to the design change in 1997, whereas the valve had been made after the design change. Because of the dimensional changes involved in the redesign, the new valve could only have been assembled into the old solenoid by forcing it in, and the nut used in the redesign to hold the new valve in the solenoid can was barely larger than the hole in the old solenoid, which had been designed for the larger threaded shaft of the old design. Wrench marks were found on the exterior of the solenoid, indicating that the valve body had been forced into the solenoid coil, although with properly matched parts, the solenoid was supposed to be assembled to the valve without the use of any force, Figures 14 and 15.


Figure 14. The longitudinal marks on the “old” design solenoid casing appear to be the result of gripping with a tool used to force a fit with the “new” design valve body.	Figure 15. The nut on the threaded end of the “new” design valve is barely larger than the hole in the casing of the “old” design solenoid.

Examination of Angels Flight maintenance records showed that the emergency brake solenoid valve had been worked on or changed in December, 1998, and again in September, 1999. The solenoid valve assembly in that system on the day of the accident was found to be the forced assembly of a burned-out (electrically open) solenoid coil to a normally closed valve instead of to a normally open valve. With these inoperative and incorrect components, neither Sinai nor Olivet was protected by the emergency brake system on the day of the accident. Based on our analysis of these records, the emergency brake system on Angels Flight had been non-functional from either December 1998 or September 1999 (the maintenance records lacking sufficient detail) until the date of the accident on February 1, 2001. The incompetent maintenance of Angels Flight had resulted in its operating for 17 to 26 months without the emergency brake protection that it was designed to have, and in this period the maintenance personnel never noticed that the emergency brake system was inoperative. Had the maintenance personnel ever checked the brake pads visually or verified the hydraulic pressure in the two systems, by looking at the gauges shown in Figures 11 through 13, when they conducted the daily brake test, they would have seen that the emergency brake system was not operating.

These findings explained why, once disconnected from the drive gear system by the failure of the female spline teeth, Sinai had not been stopped by the emergency brake system and had rolled down the incline into Olivet. The cause of the failure of the female spline teeth, however, still remained to be determined.

Metallurgical examination of the Sinai’s failed female spline showed that it had been made from AISI 1018 steel, a soft, un-heat-treated, low-carbon steel. Two engineering drawings were found for this spline among the engineering drawings for the system: one showed the material to be used was un-heat-treated AISI 1018 steel, but the other drawing indicated that it was to be made from surface-hardened AISI 8822 steel, an alloy steel. Because the identical component female spline used on the Olivet drive system did not show any evidence of spline failure (see Figure 16) one of our theories was that the Olivet spline, which showed no signs of deterioration, may have been made from the surface-hardened AISI 8822 alloy steel. However, when the Olivet spline was examined and analyzed, it was found to also have been made from the soft AISI 1018 steel. The cause of the difference in performance of the two splines remained an unanswered question. It should be remembered that, in the operation of Angels Flight, when one car descends, the other ascends, so that the number of trips, and therefore, the load history that the cars have experienced should be exactly the same.

A careful study of the maintenance evidence, including records and depositions, did reveal one difference in the utilization of the two cars, and that occurred every morning when the entire braking system for Angels Flight was tested operationally by the maintenance personnel. To perform this daily test, the maintenance person would position Sinai at the bottom of the incline and, with Olivet at the top, start the cars in motion and then hit the emergency button, which cuts off electrical power from the entire system, stopping the drive motor and activating both the emergency and service brakes. Our analysis of the motion of the two cars during this test showed that Sinai’s momentum continues to move it uphill after the power is cut, allowing its cable to slacken until it comes to a stop. But with the cable slack and the emergency brake disabled by the faulty maintenance, there is nothing that prevents Sinai from sliding back down the hill until the cable snaps taut, causing a sudden impact force to be applied to the entire Sinai drive system. This would not occur on the downhill-bound Olivet system because its cable is never allowed to go slack during this procedure and its momentum is retarded by a slight stretch of its always taut control cable.

This sudden impact force resulting from the daily brake test on the Sinai system would have no effect on the gears and splines as long as the emergency brakes are operational and clamp onto the cable drum. The emergency brake clamping onto the cable drum prevents this impact loading from being transmitted to the spline and gear train. However, after the emergency brake solenoid valve was disabled and Angels Flight no longer had an operating emergency brake, this sudden impact force was then transmitted to the entire Sinai gear drive and spline system during every morning’s brake “test”. The female spline gear, apparently the weakest link in that system, failed after the 17 to 26 months of these daily tests, as the result of the daily impact loading that occurred without an operating emergency brake system. In these 17 to 26 months, not only was the emergency system inoperative, but the maintenance personnel never properly inspected nor individually tested it, and therefore apparently never noticed that it was inoperative. If the maintenance personnel had ever just checked the hydraulic pressure gauges, shown in Figures 11 – 13, after conducting the daily brake “test”, they would have seen that there was no pressure in the service brake system, indicating that the service brake had activated, but that there was still operational pressure in the emergency brake system, showing that the emergency brakes had not been applied. Although the fracture evidence was obliterated in the failure process, it is believed that the repeated impact loads caused the splines to suffer fatigue cracking and the separated pieces of the spline were pulverized during continued use into the fine debris particles observed during the disassembly of the system, Figure 17.

As a final part of our investigation, we wanted to determine whether the female spline would have lasted longer than the subject failed spline if it had been fabricated using the surface-hardened AISI 8822 alloy steel, which most metallurgists would have preferred to use instead of the soft, un-heat-treated AISI 1018 steel. Interestingly, metallurgical testing of surface hardened AISI 8822 alloy steel coupons showed that, if this material had been used for the female spline, the thin, hardened case would have been less resistant to the impact loads resulting from the improper daily “brake tests” and would have resulted in the cracking and failure of the female spline sooner than it did with the subject, softer (but tougher), AISI 1018 steel spline.

Our examinations and analyses showed that inexcusably poor maintenance caused the emergency brake system to become inoperative and that the maintenance personnel apparently never noticed this critical problem. This failure of the emergency brake system resulted both in causing the failure of the female spline, disconnecting the Sinai cable drum from the gear drive system, and in failing to prevent Sinai from rolling down the incline and impacting with Olivet. As an added point of interest, the general design of the braking system was so robust that, with only the service brake system operative during the accident scenario, Olivet was not only brought to a halt on the incline, but after being impacted by the careening Sinai, the service brake was able to hold both cars in their positions on the incline thus preventing further harm. This final impact did cause indentations on Olivet’s male and female splines, Figures 17 and 18, indicative of the force involved in the collision.


Figure 18. Indentations on Olivet’s hardened alloy steel male splined shaft, resulting from the collision of Sinai into Olivet.	Figure 19. Indentations (shown in mirror) on Olivet’s soft carbon steel female spline, resulting from the collision of Sinai into Olivet.

There was one other maintenance procedure mandated by the Operations & Maintenance Manual for the Angels Flight system that had not been adhered to. This maintenance procedure required that oil samples be withdrawn at regular intervals (about every six months) and sent to a laboratory for the analysis of wear particulate matter. This type of analysis is a normal procedure for large equipment that may suffer wear problems. By regularly conducting this type of analysis, the maintenance personnel would have been able to get non-invasive advanced notice if some portion of the system began to wear. This was done during the first years of Angels Flight operation, but for unknown reasons was discontinued long before the subject accident. The last oil analysis report was performed in May 1998 and, because the amount of particulate material was found to be increasing compared with prior analyses, the contract laboratory recommended that the time intervals between analyses be reduced. Instead, the oil analyses were discontinued. Had these analyses been continued and had the maintenance personnel realized the import of these analyses, a complete inspection of the Angels Flight system could have been undertaken to determine the cause of the particulate debris and could have corrected the problem, thus completely averting this tragic accident.