Swissair 111: A Needle Found in a Haystack
The Canadian government’s probe into the fatal inflight fire was that nation’s most exhaustive ever; the accident was a "wake-up call" regarding flammable materials on airplanes.
By Jan W. Steenblik, Technical Editor
Air Line Pilot, August 2003, p.12
in the pilots’ training would have alerted them to the possibility of a
fire in the cockpit or cabin attic areas—they would not have been aware
that large quantities of flammable material resided in those hidden
—Vic Gerden, Investigator-in-Charge, TSB
Accident investigation is seldom, if ever, easy. The Transportation Safety Board of Canada (TSB) investigation into the loss of Swissair Flight 111 on the night of Sept. 2, 1998, was particularly difficult: The MD-11, carrying 215 passengers and 14 crew members, plunged into the cold waters of the North Atlantic about five miles southwest of Peggy’s Cove, a picturesque fishing village on Nova Scotia’s south coast.
On March 27, 2003, the TSB held a press conference in Halifax, N.S., to announce completion of the arduous investigation of the accident. Camille Thériault, TSB chairperson, declared, "Our investigation took more than four years to complete. Consider its complexity:
• "millions of pieces of shattered aircraft needed to be recovered from the sea;
• "of the more than 250 km [~155 sm] of electrical wire, much was recovered, and it all needed to be identified and examined;
• "extensive series of flight and laboratory tests had to be conducted; and
• "exhaustive analysis of events—even when little information existed —needed to be carried out."
The TSB retrieved about 98 percent of the airplane, by weight, despite the fact that the wreckage rested 55 meters (180 feet) down on the ocean floor. The recovery and investigation cost approximately C$57 million, paid for by Canadian taxpayers. All told, it was the largest, most complex safety investigation that the TSB has ever undertaken; the agency released a 338-page report that elaborated on its exhaustive efforts.
Vic Gerden, the investigator-in-charge (IIC) of the TSB investigation, provided a superb summary of his team’s work.
Swissair 111 was on a scheduled flight from New York to Geneva, Switzerland, when, approximately 53 minutes after takeoff, while cruising at FL330, the pilots smelled an abnormal odor and noticed a small amount of smoke in the cockpit.
The pilots determined that the odor and smoke came from the air conditioning system. They briefly discussed returning to New York, but decided to divert to the closest and most suitable airport, and were cleared to Halifax International Airport.
Less than 9 minutes after beginning their descent, however, the situation in the cockpit rapidly deteriorated. The pilots declared an emergency and told the air traffic controller that they needed to land immediately. Radio communications with the airplane ended about 20 seconds later. Soon afterward, while the airplane was still at 10,000 feet, Swissair 111’s transponder signal suddenly vanished from the controller’s radar scope, and the airplane’s flight data recorder and cockpit voice recorder abruptly stopped. During the last 5½ minutes of the flight, the airplane made a 360-degree descending turn and crashed into the ocean at an estimated 300 knots. The airplane was destroyed, and all 228 persons aboard died.
"Early in the investigation," Gerden explained, "we knew, from seeing heat damage on some pieces of wreckage, that a fire occurred in the forward overhead area of the aircraft….Our analysis indicates that the fire started in a hidden area above the cockpit ceiling, on the right-hand side, near the rear wall of the cockpit."
The TSB eventually determined that the fire started from electric wires arcing, which ignited flammable cover material on thermal acoustic insulation blankets. The material was metalized polyethylene terephthalate (MPET).
"Our tests showed that these insulation blankets were readily ignitable from an electrical arc," Gerden continued, "and that they invariably sustained and propagated a fire. Without the presence of this and other flammable material, this accident would not have happened."
The fire, the TSB’s painstaking analysis found, spread rearward across the surface of the MPET blankets, out of the cockpit attic and into the attic area above the forward passenger cabin. There, other flammable materials helped the fire spread and intensify. As a result, the cockpit environment deteriorated, some of the airplane’s systems failed, and the entire airplane and its occupants were lost.
issued 23 aviation safety recommendations as a result of the Board’s
investigation of the crash of Swissair 111:
• two regarding thermal acoustical insulation
materials and flammability test criteria for materials used on new
Regulatory authorities, airlines, and aircraft manufacturers have
already adopted several of these recommendations, which has improved the
safety of aviation worldwide.
For example, the inflight entertainment network system was removed voluntarily from Swissair aircraft. Insulation covered with metalized polyethylene terephthalate (MPET) and used on thermal acoustical insulation blankets has been ordered removed from all aircraft.
Flammability standards for materials used in aircraft are being upgraded, and inflight firefighting procedures have been subjected to intense review and changes have been made. Other recommendations are also being implemented.
Initial arcing event
Of the 155 miles of wire installed on the MD-11, the TSB recovered much, using six different methods of wreckage recovery.
"From all this wire," Gerden said, "we ultimately found 20 pieces that displayed regions of melted copper wire, indicative of arcing damage. Wire arcing creates very high temperatures that melt a localized region of the wire.
"One particular piece of wire had two arc sites about 50 cm (20 inches) apart. We determined that one of these arc sites was located just above the cockpit ceiling, just in front of the rear cockpit wall, on the right side of the cockpit. We believe that this is the area where the fire started, and that this arc site was associated with the lead arcing event."
This arc site was found on one of the wires that supplied power to the inflight entertainment network (IFEN), a supplemental system installed in the business and first-class sections of the cabin. Gerden stressed, however, "It’s unlikely that this entertainment system power supply wire was the only wire involved in the lead arcing event. We strongly suspect that at least one other wire was involved—either an aircraft wire or another entertainment system wire." Unfortunately, the TSB was not able to find the other wire(s) involved.
When the initial arcing occurred and the MPET started to burn, some of the odor spread forward into the cockpit, the TSB believes. Some intermittent smoke moved into a small area of the cockpit, probably through openings in a panel at the cockpit ceiling, near an air conditioning outlet, according to the TSB analysis.
"We suspect this is one reason why the pilots initially believed the odor and smoke were from the air conditioning system," Gerden explained, "and the reason why they did not view this anomaly as life-threatening, given that anomalies associated with air conditioning systems are routinely considered to be benign.
"Also," he continued, "no warnings or alerts on the pilots’ panels indicated any electrical system anomaly. Nothing in the pilots’ training would have alerted them to the possibility of a fire in the cockpit or cabin attic areas—they would not have been aware that large quantities of flammable material resided in those hidden areas."
In addition, no smoke or fire detectors that could have provided critical information to the pilots were required to be above the cockpit or cabin ceilings. None were installed.
"We believe that the initial, small creeping flame front was prevented from moving further forward into the cockpit attic by the geometry of the aircraft and by physical barriers," Gerden said. "However, the fire propagated rearward, into the attic area above the forward passenger cabin, through a cutout at the top of the cockpit rear wall. This cutout was normally filled with a foam material; in our testing, this foam material was found to be flammable."
|The TSB concluded that, even if the pilots had foreseen the disastrous consequences of the fire, they would not have been able to land safely in Halifax because the fire progressed so rapidly.|
Detailed airflow testing and fire modeling showed that most of the odor and smoke from the creeping fire would be (1) drawn away from the cockpit into the avionics compartment below and dumped overboard, or (2) drawn rearward to the recirculation fans in the cabin attic area. This airflow pattern, Gerden asserted, "would have helped to conceal the fire—and deprived the pilots of additional cues, thereby supporting their initial belief that the odor in the cockpit was related to air conditioning and that they had time to prepare for the precautionary landing."
The TSB believes that, after the fire passed behind the cockpit rear wall into the area above the forward cabin ceiling, it began to grow quickly because of the amount of additional flammable material that was available.
"The probable early failure of a silicone elastomeric endcap, used to cap off the unused vent duct of the deactivated Number Two galley, created a significant vacuum just aft of the initial fire area," Gerden continued, "that further led to the evacuation of the smoke and odor from this area. This also would have contributed to the delay" in the pilots’ detecting the fire.
Conditions in the cockpit deteriorated as the pilots were following the emergency checklist for smoke. The checklist required the pilots to turn off the cabin electrical busses, which are connected to (among other circuits) the air recirculation fans above the cabin ceiling.
"When the recirculation fans stopped," Gerden explained, "the airflow in the cabin attic area reversed, and the fire was drawn forward, into the attic area above the cockpit ceiling. The flight data recorder, for the first time during the flight, started to record aircraft systems failures that we believe were the result of the rapidly increasing damage caused by the fire."
The TSB also believes that parts of the cockpit ceiling liner material would have sagged and started to give way shortly after the FDR began to record systems failures. This would have been around the time that both pilots declared an emergency.
Moments later, the air traffic controller lost communications with Swissair 111. The pilots lost most of their flight instrumentation and ability to navigate. Within minutes, the pilots were no longer able to maintain controlled flight. The MD-11 struck the ocean in a steep right turn, with the nose down about 20 degrees, at about 300 knots.
"Throughout this investigation," Gerden acknowledged, "people have asked, ‘Could [the Swissair 111 pilots] have landed safely if they had been aware of the fire earlier?’"
In a word, no. Detailed, theoretical "minimum-time-to-landing" descent calculations showed that the earliest Swissair 111 could have landed was about 4 minutes before the MD-11 crashed into the ocean. However, Gerden emphasized, these calculations were based on ideal conditions—a perfectly functioning airplane and no disruption to the pilots’ ability to fly and navigate. Moreover, he said, "we know that by 10:25 [p.m. local time, 2 minutes before the ideal earliest possible landing time], conditions in the cockpit were far from ideal.
"A number of aircraft systems—the autopilot, the primary flight instruments, and communications and navigation systems—had failed. Other systems normally used for landing likely also continued to deteriorate. Also, the fire was starting to breach the cockpit ceiling, which would allow a significant amount of smoke and heat to enter the cockpit, substantially affecting visibility—and possibly preventing the pilots from seeing adequately outside the aircraft to navigate and land safely."
The TSB concluded that, even if the pilots had foreseen the disastrous consequences of the fire, they would not have been able to land safely in Halifax because the fire progressed so rapidly.
As a result of this investigation, the TSB issued 23 safety recommendations—14 before releasing its final report, and another 9 concurrent with the report (see sidebar). As with many major aviation accident investigations, some of the recommendations were directly associated with the causes of the accident, and some were associated with other safety risks uncovered during the investigation.
"The use of flammable materials in aircraft has been an important focus of this investigation," Gerden pointed out. The TSB asked regulators and the airline and aircraft manufacturing industries to
• reduce the risk associated with the use of MPET-covered insulation blankets in aircraft;
• develop more rigorous flammability testing for all materials; and
• prevent use of any material that sustains or propagates fire.
In June 2000, regulators—including the FAA and Transport Canada—ordered removal of MPET-covered insulation blankets from more than 1,900 aircraft. The agencies also have taken steps to improve flammability testing of other types of insulating blanket materials.
"While these are positive steps," Gerden acknowledged, "the TSB remains concerned about the flammability of certain cover materials on insulation blankets [currently] in use."
Potential ignition sources also were central to the investigation. Early in the investigation, in December 1998, the TSB issued a safety advisory that led to comprehensive additional inspections of MD-11 wiring and electrical components. Boeing worked with the FAA to issue more than 50 airworthiness directives that ordered examination of wiring and components to find and eliminate potential ignition sources.
The TSB also called upon regulators and aircraft manufacturers to conduct tests designed to evaluate the characteristics of wire failures and to mitigate the risk of igniting nearby materials. This work is under way, although more needs to be done.
The Board also examined inflight firefighting measures. As a result, it made a number of recommendations to regulators and the airline and aircraft manufacturing industries, including a recommendation to install built-in equipment in hidden areas to detect and suppress fire and smoke.
The TSB has urged regulators and manufacturers to review the design of standby flight instruments—especially their positioning and power source—and to ensure that flight crews are properly trained to be able to continue to fly if the primary flight instruments fail.
"We’ve also uncovered a number of safety deficiencies that are not directly related to causes or contributing factors to the accident," Gerden noted.
"For example, while examining how the entertainment system was installed, we discovered a design flaw in the way the entertainment system had been wired into the airplane’s electrical system." Namely, the design used an electrical bus that would not shed power to the entertainment system when the cabin electrical busses were turned off!
"This flaw did not play a role in the initiation or propagation of the fire on Swissair 111," Gerden stressed, "because the fire was well under way when the cabin busses were turned off. However, it did point to deficiencies in how supplemental type certificates are approved for add-on systems.
"While regulators have made some improvements in policies and procedures in this area," he added, "the Board is calling for sufficient scrutiny on every supplemental system being certified, to ensure that it is properly integrated into the aircraft design."
Another area of concern to the TSB during this investigation was procedures regarding resetting circuit breakers. Although the Board turned up no evidence that the pilots of Swissair 111 reset any circuit breakers, the investigation revealed variations in procedures for resetting circuit breakers in the airline industry.
Gerden said "much has been learned about the risks of resetting circuit breakers, and work is taking place to improve circuit breaker design so that they better protect electrical wiring." The TSB also called for clearer guidelines for flight crews and maintenance personnel regarding procedures for resetting circuit breakers.
During the Board’s search for potential ignition sources, investigators found safety deficiencies with the MD-11 flight crew reading lights (map lights). Even though the map lights did not play a role in starting the fire on Swissair 111, regulators and manufacturers acted quickly to address these problems; the map light has been substantially redesigned, and the lights are being replaced in MD-11s.
"As these examples indicate," Gerden declared, "the investigation of this accident has led to many important changes in how aircraft will be designed, constructed, equipped, and operated. Changes have also taken place in how flight crews will be trained to deal with potential fire-related situations….
"Twenty pages of our report are devoted to summarizing the safety actions that have already been taken. The good news is that each change marks a positive impact on the safety of air travel—in Canada and around the world."
|ALPA Lends a Hand|
|Vic Gerden, a
seasoned aviation accident investigator for the TSB, was asleep in
Winnipeg on the evening of Sept. 2, 1998, when the telephone rang:
Swissair Flight 111 had gone into the ocean near Halifax—would he
accept the daunting task of serving as the investigator in charge
(IIC) of the investigation? Gerden was up all night, getting his
team of about 40 TSB staffers together, and on a 6 a.m. flight to
Capt. Ken Adams (Delta, now retired) was also at home—in north Georgia, watching television news—the night Swissair 111 went down. Swissair was one of Delta’s code-sharing partners. When the breaking story appeared on the screen, Capt. Adams recalls, "I realized I was the only member of the [ALPA] Accident Investigation Board qualified on the MD-11. It was going to be a long night." He didn’t come home until Christmas. After a week off for the holidays, Capt. Adams returned to the field investigation until April 1999.
Other ALPA members, notably Air Nova (now Air Canada Jazz) pilots who lived in or near Halifax— Capts. Gerry Davis, Ray Gelinas, and Nick Seemel, plus Capt. Gary Demone (Canadian, now Air Canada)—were already on the scene the morning of Sept. 3, 1998 to help in any way they could. They were soon joined by trained pilot safety representatives and accident investigators from ALPA and Swiss ALPA, all under the banner of the International Federation of Air Line Pilots Associations (IFALPA).
Larry Vance, the TSB Deputy IIC for the Swissair 111 accident investigation, came from Ottawa. The ALPA representatives, he remembers, showed up "pretty much immediately" and were at the first operations center that the TSB set up in Peggy’s Cove (they soon moved inland to Dartmouth and later to a nearby military base at Shearwater).
"We explained to them the differences between how the NTSB operates and how we operate," Vance explains. "Up here they could have observer status, but not full participant status as on an NTSB investigation."
Adds Gerden, "The process we used [to incorporate the IFALPA investigators into the investigation] was a combination of ICAO Annex 13 [the international agreement that governs aviation accident investigations] and the Canadian legislation [that authorized the creation of the TSB]."
Vance continues, "We talked to the pilots who came up, the early arrivals, quite a bit about our rules and procedures. Everyone on [the pilot side] and our side wanted the process to go smoothly. We knew it would take a while to build up a level of comfort with the process—this was the first time IFALPA had participated in a major aviation accident investigation in Canada.
"One of the things we were most concerned about was the possibility of information being leaked during the investigation. We did a lot of things to make sure none of the IFALPA folks would be blamed for any leaks."
TSB’s Gerden points out that "the [IFALPA representatives] worked extremely hard. Often they were among the last to leave the hangar at night. Because they were observers, we supervised everything they did. We developed a good relationship."
One of the most difficult parts of the Swissair 111 investigation, says Vance, "was just staying with it for so bloody long, keeping people focused on the task at hand."
Gerden explains, "This investigation involved very complex recovery. That took about 13 months, plus another 2 months of initial sorting. That was something we innovated as we went along."
The pilots’ contributions to the investigation ranged from grunt work—moving and washing wreckage at the dock—to developing sophisticated software. First Officer Jim Shaw (Delta), using TSB computers at the agency’s operations center, developed a system for cataloging wreckage and for tracking wire fragments so that they could be reconstructed as wire bundles.
By July 2003, ALPA had spent about $320,000 of members’ dues on the Swissair 111 accident investigation.
In August 2000, Capt. Ted Murphy (Aer Lingus), IFALPA’s president, presented IFALPA Presidential Citations to 13 pilots from Switzerland, the United States, and Canada for their selfless volunteer service in the Swissair 111 investigation that was still ongoing at that time.
Receiving citations were six ALPA members—Capts. Adams, Davis, Demone, Gelinas, and Seemel, and F/O Shaw—plus ALPA staff engineer Mike Huhn, from the Engineering and Accident Investigation Section of ALPA’s Engineering and Air Safety Department. Also receiving citations were seven Swiss ALPA members—Capts. Andrew Baldwin, Pete Eggler, and Mike Fraser, plus First Officers Alain Escher, Tom Leeman, Guillermo Mulet, and Patrick Sutter.
Capt. Murphy said the citations were being awarded to the Swissair 111 investigation team "for enhancing the professional status of all flight crew by their outstanding contribution[s], for assisting in the recovery and reconstruction of the MD-11 and all of its systems, and for aiding in the development of sophisticated computer technology as a primary investigative tool."
Capt. Eggler told the ALPA representatives, "Our Association represents about 1,100 Swissair pilots. Our pilot safety group is twelve pilots. Your organization was there for us—you comforted us, guided us, and established a working relationship with the Canadian authorities. We feel very lucky to have enjoyed the [relationship with your volunteers], the resources you have made available to us, [and] the support of your membership."
Capt. Adams acknowledged that the protracted field phase of the investigation caused the pilot volunteers "a lot of pain and hardship, being separated from our families." But, he added, "within the first month of the accident investigation, we identified six major issues that affect all our airplanes, not just the MD-11."—JWS