Shown in green are the fuel tanks, including the center wing tank (CWT) in the part of the wing under the cabin floor. The CWT has a capacity of 12,890 gallons, as much as a swimming pool 8 feet deep, 10 feet wide, and 21½ feet long. As with fuel tanks in cars, boats, lawn mowers, etc., aircraft fuel tanks are vented. Without vents, air pressure, which is more than 2,100 pounds/square foot at sea level, would collapse the tanks as fuel is used up. Each 747 fuel tank has a separate vent line to one of the two surge tanks near the wingtips. The surge tanks are vented to the atmosphere through openings, shown here in red, on the bottom of the wing near the wingtips. Because fuel is volatile and evaporates, the openings near the wingtips allow fuel vapors to be vented away from the hot engines. As fuel is used up, air enters the vents and fills the ullage space (space above the fuel) where it mixes with fuel vapor. For technical and financial reasons, large airliners rarely have their fuel tanks completely filled. They are only filled with sufficient fuel for the next flight. Also for technical reasons, fuel is only added to the center wing when the required fuel is more than the main tanks can accommodate. Finally, if fuel is added to the CWT, it is used first. Since the other tanks of a Boeing 747-100 can carry enough fuel for 8 or 9 hours of flight, the CWTs of 747s are often empty, or almost so, having only a small amount of residual fuel that the pumps missed. Not shown are the 747's three air-conditioning and pressurization units (packs) that are mounted under the CWT.
Also in the wing are many miles of electrical wiring. There are so many wires that they are bundled together into hundreds of bundles, some of which contain hundreds of wires. Some bundles are routed together and some touch other bundles where they cross. Some of the wires are for the Fuel Quantity Indicating System (FQIS) that uses a set of probes and compensators in each fuel tank. There are 7 FQIS probes in the CWT. When N93119 was built, most of the aircraft's wiring was insulated with a material named Poly-X. The wiring inside the tanks for the FQIS components used Teflon insulation. Although the FQIS operated at low voltage, some of the FQIS wires were bundled with wires that carried as much as 350 volts. All that prevented a potentially-dangerous electrical short was the integrity of the wire insulation.
On Dec. 8, 1963, Pan American (Pan Am) flight 214, a Boeing 707, landed in Baltimore after a flight from Puerto Rico. There was a small amount of residual fuel in the CWT. The aircraft was then fueled for a short flight to Philadelphia but no fuel was added to the CWT. Thirty-four minutes after takeoff for Philadelphia, the 707 exploded and crashed, killing all 81 people on board.
In those years, civil aviation accidents were investigated by the Civil Aeronautics Board (CAB). Like the National Transportation Safety Board (NTSB) that succeeded it in 1967 for accident investigations, the CAB had no regulatory or enforcement authority; it could only make recommendations to the FAA which was free to follow or ignore the recommendations as it saw fit. The CAB investigators determined that the aircraft had been struck by lightning that ignited fuel vapor in the vent system, triggering explosions in the fuel tanks, including the CWT. 2
On December 17, 1963, nine days after the crash of flight 214, Leon Tanguay, director of the Civil Aeronautics Board (CAB) Bureau of Safety, sent a letter to the FAA recommending several safety modifications for future aircraft. One recommendation was to reduce the fuel vapors in the tank by air circulation or by filling the empty space in the tanks with nitrogen. Since combustion cannot take place in nitrogen, the potential for fuel-tank explosions would be eliminated. At that time, nitrogen-inerting systems were under development for two aircraft which would get very hot at their design cruising speed of three times the speed of sound; the XB-70, prototype of the B-70 bomber; and the SR-71 reconnaissance aircraft. Both aircraft first flew in 1964. The XB-70 system was developed by the Parker Hannifin company. After the crash of Pan Am 214, Parker Hannifin tried to sell its system to the airlines. They met with Juan Trippe, the CEO of Pan American. He was enthusiastic about the system and said that it would have prevented the accident. However, the FAA did not require the installation of such systems and, without an FAA requirement, Parker Hannifin did not continue development of the system for airliners at that time. Only two XB-70 aircraft were built. On June 8, 1966, one of the XB-70s was involved in a mid-air collision with a fighter. The fighter exploded and crashed. The XB-70 was so badly damaged that it went out of control and crashed. There was no fuel explosion or fire. 8
The other early aircraft with a nitrogen-inerting system, the SR-71, flew in U.S.A.F. service from 1966 until 1998. 10
By the mid 1960's, Boeing was developing the 747. Although Parker Hannifin and some safety experts tried to get inerting systems installed on the 747, neither Boeing nor the airlines thought the risk of explosion or fire justified the expense of installing such systems. Neither did the FAA and without an FAA requirement, such systems were not installed.
Although the FAA failed to recognize the life-saving potential of fuel-tank inerting systems, the United States Air Force (USAF) did recognize the potential and installed them on several types of aircraft. By 1977, such systems had been installed in the Air Force's entire fleet of C-5A's; large transport aircraft comparable in size to the Boeing 747.
In 1970, Poly-X wire insulation was adopted for use in some Navy fighters. Within a few years, the U.S. Navy began experiencing Poly-X insulation malfunctions, some of which caused fatal crashes. In 1977, the U.S. Naval Avionics Facility (NAFI) issued a report that stated that the time to field failures for Poly-X in U.S. Navy was between 3 and 5 years in service. This information was not given to the FAA.
During the 1970's, Boeing modified four new 747's for the USAF. These aircraft, designated E-4's, were designed to serve as survivable, mobile, airborne command posts in time of war or threat of war. One contingency was to fly the president of the United States to safety. They were loaded with communications equipment that generated a lot of heat. One aircraft was manned and kept on alert, ready to take off on short notice, 24 hours a day. Air-conditioning packs were kept running to counter the heat of the electronic equipment that was turned on, ready for immediate use. The Air Force discovered that such prolonged ground operations caused the fuel in the CWT to get hot. The Air Force complained to Boeing. Boeing conducted a study from 1979 to 1980 and proposed operational procedures to reduce CWT temperatures. They also concluded that thermal insulation should be installed between the air conditioning packs and the CWT. Boeing did not inform the FAA about any of this.
On March 1, 1976, the Imperial Iranian Air Force took delivery of a Boeing 747-131 that it had purchased from TWA after it had been converted to a freighter. Two months later, on May 9, 1976, it crashed as it was approaching Madrid, Spain. According to the NTSB accident report,
"Witnesses observed lightning strike the aircraft followed by a fire, explosion, and separation of the left wing." 3
There was an explosion in the No. 1 fuel tank and fires in the No. 2 fuel tank and the left surge tank.
In addition to 747's, Boeing builds other types of airliners including the 737, their smallest. The 737 is the most-widely used airliner
in the world. Two of them also suffered fuel-tank explosions.
On May 11, 1990, a Philippines Air Lines Boeing 737-3Y0 was on the ground prior to departure at Manila International Airport in the Philippines. The temperature was 95° F and the air-conditioning packs under the CWT were running. During pushback from the gate, there was an explosion in the CWT. Eight passengers were killed. The accident report concluded that "the vapors ignited probably due to
damaged wiring." 4
On March 3, 2001, a Thai Airways Boeing 737-4D7 was parked at a gate at Bangkok International Airport, Thailand, being prepared for a flight. It was destroyed by an explosion and fire. One person was killed and six were seriously injured. The Thai accident investigation found that the probable cause was the ignition of the flammable fuel/air mixture in the CWT, possibly due to contamination of the CWT fuel pump by metal shavings. 5
In 1940, Congress created the Civil Aeronautics Board (CAB). Its responsibilities included establishment of airline safety regulations and to monitor and enforce compliance. Another responsibility was economic regulation of the airlines. The CAB understood that airlines need to be profitable in order to buy new airplanes from time to time, pay salaries that will attract competent air crew and maintenance personnel, and to perform proper maintenance and training. In 1978, Congress passed the Airline Deregulation Act of 1978. It gradually phased out economic regulation of the airlines over a period of years. This unleashed a tremendous increase in the number of airline flights and fare wars to fill them. Prior to airline deregulation, many aircraft components were replaced or overhauled at set intervals. With the reduction in profits, the airlines were not able to replace older aircraft and were forced to cut costs everywhere they could, including maintenance. After deregulation, maintenance procedures were examined to see if they exceeded the FAA requirements and could be cut back to save money. The airlines lobbied the FAA to allow condition-monitored items, meaning that they were removed and replaced only when inoperative; not periodically at designated intervals.
On the morning of July 17, 1996, N93119 was in Athens, Greece being prepared for a non-stop flight to JFK as TWA flight 881. Due to the distance of approximately 5,000 miles and prevailing westerly head winds, the fuel tanks were filled to their capacity of 47,210 gallons weighing approximately 316,300 lbs; 115,000 pounds in each wing and the rest in the CWT. During refueling, an observer near one of the wingtips would have seen the wing flex down significantly as 115,000 pounds of fuel was added to the tanks in that wing. That was completely normal. The wing was designed to be flexible in order to reduce the stress of turbulence on the structure and to provide a more comfortable ride through turbulence for the passengers. As the wing flexed downward, fuel lines, hydraulic lines that carry hydraulic fluid at a pressure of 3,000 pounds/square inch, pneumatic ducts that carry very hot air, and bundles of electrical wire all flexed down with it. All of this, and what happened during the flight to New York, was completely normal. N93119 was 25 years old in July 1996. Except for a 1-year period when it was parked, it had made 16,868 flights and flown almost 93,300 hours; an average of almost 2 flights/day, every day, for 24 years. This extraordinarily "high mileage" was the result of TWA's inability to buy new airplanes due to the financial difficulties inflicted by airline deregulation.
After N93119 was fueled and loaded, the engines were started and the aircraft taxied to the runway. With the load of 349 passenger, 17 crew members, and full fuel, the total weight would have been in excess of 700,000 pounds. The air temperature is not known but likely would have been around 80° F at the takeoff time, shortly after noon local. When the aircraft was on the runway and cleared for takeoff, the pilot flying pushed the throttles forward to the takeoff position. The four engines revved up to produce almost 48,000 pounds of thrust each, straining against the engine mounts to accelerate the aircraft forward. For an observer near the runway, the sound would have been deafening. High-frequency vibrations coursed through the wing structure and components in it. Some components got very hot. At the designated airspeed, perhaps 180 miles/hour, the pilot pulled back on the control wheel, causing the nose of the aircraft to rise and the wings to start producing lift. In approximately four seconds, the lift force was more than the weight of the airplane; more than 350,000 pounds of force on each wing. These enormous forces flexed the wings upward several feet, flexing the fuel lines, bundles of electric wires, etc. with them. As the aircraft climbed, it might have encountered light turbulence. A passenger seated at a window near the wing would have seen the engines nacelles and wings twisting and flexing as the aircraft flew through the turbulent air. About thirty minutes after takeoff, the aircraft would have reached its initial cruise altitude of 29,000 feet or higher. At that altitude, the outside air temperature would typically be -45° F or lower, a drop of more than 120° in thirty minutes. Any atmospheric moisture that had been inside the wings would have frozen and formed frost on the cold structure and components. As the flight progressed and fuel was consumed, the crew would have periodically requested higher altitudes to provide maximum efficiency. The outside air temperature at the higher altitudes would typically be -60° to -70° or colder.
More than 10 hours after takeoff, the flight was cleared to descend for landing at JFK. The pilot reduced the thrust to idle and N93119 descended rapidly from the cold of the stratosphere into the warm humid air near the ground that July afternoon. As it passed through turbulence typical in such conditions, the wings and engines twisted and flexed. The aircraft was much lighter than when it took off, having burned approximately 292,000 pounds of fuel and the wings were flexed upward less than they had been after takeoff. Immediately after touchdown, the pilot lowered the nose and the lift forces that were bending the wings upward vanished. The wings returned to the neutral position, just as they had so many thousand times before. When the aircraft was parked, maintenance personnel went to inspect it. They checked the remaining fuel and noted that 50 gallons remained in the CWT out of the 12,890 gallons with which the flight started; like a puddle in an otherwise-empty swimming pool. Parts of the structure were still relatively cold and condensation dripped off the bottom of the aircraft in the warm humid air. The frost that had formed inside the wings and fuselage during the climb thawed and soaked parts of the structure and various components. The aircraft was scheduled to depart for Paris as TWA flight 800 less than four hours later and remained at the gate. Two of the three air-conditioning packs were kept running for approximately 2½ hours before departure.
A couple of hours later, the crew assigned to TWA flight 800 arrived at the airport. The relatively short flight to Paris, aided by prevailing westerly tail winds, was provided for by 176,600 pounds of fuel; none in the CWT. Neither the crew, nor anyone else at TWA, nor the passengers, knew that the air-conditioning packs were vaporizing the small puddle of residual fuel in the CWT and filling that swimming-pool-size space with an explosive mixture of fuel vapor and air. They did not know about the tendency of the Poly-X wire insulation to crack, thereby increasing the possibility of an electrical short. They did not appreciate the cumulative effect on N93119 of the stress, vibration, flexing, chafing, condensation, pressure changes, and large temperature changes inflicted by 24 years of flight. They did not know about the fuel overheating in the CWT's of E-4's or the history of fuel-tank explosions that had primarily affected Boeing aircraft.
Less than four hours after arriving from Athens, N93119, weighing a calculated total of 590,441 pounds, roared down runway 22R at JFK. It lifted off the ground once more, climbing from the warm and moist air at sea level that evening toward the freezing temperature of the stratosphere. Twelve minutes later, the hot mixture of fuel vapor and air in the CWT exploded.
After the primary focus of the search and recovery efforts to locate and recover the victims, a thorough search was made to recovery the wreckage, which was spread out in an area about 4 miles long by 3 1/2 miles wide in the Atlantic Ocean. This effort took more than 10 months and recovered more than 95% of the wreckage, including both flight recorders and the Cockpit Voice Recorder (CVR). During the thorough, painstaking investigation that began immediately after the disaster, all pieces of wreckage were examined for evidence of an explosive device or missile. No such evidence was found. The investigators carefully examined the data from the flight and voice recorder, maintenance records, aircraft design features and standards, the history of fuel tank explosions, and many other sources of information that might have some relevance to the disaster. The accident report is 341 pages long. It states, "Physical evidence indicated that an overpressure event occurred in the airplane's CWT."
During the accident investigation, Boeing never notified the NTSB about the study that they had done regarding overheating of the fuel in the CWTs of the Air Force's E-4 (B-747) aircraft. It was not until 1999 that the NTSB learned of the study from the USAF Safety Center, Directorate of Engineering and Technical Services.
The data recovered from the flight recorder and CVR indicated that everything was normal until ten minutes after takeoff. At that time, the Captain said, "Look at that crazy fuel flow indicator there on number four ... see that?" Instrument malfunctions are not uncommon and are usually not a cause for concern. About 1½ minutes later, there was a series of three noises lasting for a total of about 30 seconds. Fifteen seconds later, the recorder stopped, the same time when electrical power was lost.
When the evidence indicated that the CWT exploded, the investigators turned their focus to the components in it. Fragments of all seven FQIS probes from the CWT were recovered, and some of the wiring. Except for four probes in the No. 1 main tank, all of the FQIS probes and compensators installed in N93119's fuel tanks were the original units installed when the airplane was manufactured in 1971. Those components were condition-monitored items, meaning that they were removed and replaced only when inoperative. Among the electrical discrepancies found by the investigators were:
"unshielded FQIS wires routed in wire bundles with high-voltage wires."
"wiring documents for the accident airplane did not show all of the wires that were recovered."
"wire configurations in the accident airplane ... did not match those shown in Boeing's technical diagrams."
"... wires were apparently not bundled as shown in the PI (Boeing product illustration) but were randomly laid in trays"
"Several wire repairs were also found that did not comply with standards for repairs or installations used by Boeing or TWA ..."
"Numerous wire splices were covered by a plastic insulating sleeve over metal barrels, such that the ends of the wire splices were open, with no other wrapping. These wire splices were noted at locations throughout the airplane (including areas exposed to fluid contamination, such as above the potable water tanks on the front spar) and occasionally showed evidence of corrosion."
"The connector pins in the recovered fuel totalizer gauge contained excessive solder, which appeared to have inadvertently joined connecting pins/wires from the right wing main fuel tank and CWT FQIS. The excessive solder had cracked between the connecting pins."
"Sulfide deposits were found on some exposed conductors" and on some FQIS wires"
"... FQIS wires and wires routed adjacent to the FQIS wires revealed possible evidence of arcing in several locations."
In July 1997, the NTSB leased a Boeing 747-121. They installed special instrumentation and flew a series of flight tests to gather
data in the CWT under conditions similar to those that existed on TWA flight 800.
During one test flight, TWA flight 800's ground and flight operations were duplicated as closely as possible.
The flight was conducted two days before the anniversary of the TWA 800 flight and within one minute of the takeoff time.
The air-conditioning packs were operated in the same manner and the temperature on the ground was about the same.
The fuel load was the same, including 50 gallons in the CWT that had been obtained from the same source in Athens. The previous
year, on October 1, 1996, the NTSB obtained a sample of fuel from the CWT of TWA flight 881 that had arrived from Athens that day.
The NTSB Safety Board considered this fuel sample to be representative of the fuel in N93119 when the explosion occurred. They had
the fuel tested and the flash point was measured at 114° F. The flash point is the lowest temperature at which it can vaporize
to form an ignitable mixture in air.
The data collected during the flight indicated that the highest temperature in the CWT ullage space (space above the fuel)
was 145° F just before the airplane began to taxi for takeoff. At the time and altitude of the explosion, it was 127°F. Both
temperatures were well above the flash point.
From the beginning of the investigation until 1999, NTSB investigators continued to study the sulfide residue found on some FQIS
wiring and on a fuel probe connector from the CWT. Similar sulfide deposits had been found on the same components recovered from the
Iranian Air Force 747 that blew up and crashed in 1976. The investigators learned that some fuel tank wiring with similar residue had been
removed from an Air Force airplane in 1991 and sent to maintenance. When the maintenance worker applied an electrical tester to the wiring,
some residual fuel vapors ignited. The residue was caused by a chemical reaction between the metal and sulfur in the fuel. The deposit was
very fragile and would not carry an electrical charge if disturbed. In 1999, the NTSB received part of an FQIS part that had been taken off
an airliner because of a short circuit in the CWT. The part had undisturbed sulfide residue like that on the FQIS components from TWA 800.
When the NTSB researchers applied a voltage to the wiring, the residue flashed and made a popping sound. 9
Among the conclusions listed by the NTSB in the comprehensive 425-page report were:
"The fuel/air vapor in the ullage of TWA flight 800's center wing fuel tank was flammable at the time of the accident."
"A short circuit producing excess voltage that was transferred to the center wing tank (CWT) fuel quantity indication system wiring
is the most likely source of ignition energy for the TWA flight 800 CWT explosion."
"Silver-sulfide deposits on fuel quantity indication system components inside fuel tanks pose a risk for ignition of flammable fuel/air vapor."
"A combustible fuel/air mixture existed in all fuel tank air spaces."
During the years between the Pan Am 214 and TWA 800 disasters, some steps were taken to reduce the probability of aircraft fuel-tank explosions,
primarily by design changes to reduce the possibility of ignition in fuel tanks. Thirteen airliners experienced fuel tank explosions during ground or flight operations during that period. 6 Given the history of maintenance errors and component failures, betting the lives of passengers and crew members on the absolute reliability of an electrical system operating for decades under the stressful conditions described seems irresponsible indeed. Furthermore, neither the Thai not Iranian accidents were thought to have been caused by electrical malfunctions of the sort that occurred on TWA 800. In any event, fuel-tank inerting systems would have prevented all of the explosions mentioned in this essay.
One June 29, 2006, the FAA published a fact sheet on fuel tank safety. It contained this paragraph:
On November 23, 2005, the FAA proposed a rule that would require more than 3,200 existing and certain new large passenger jets to reduce flammability levels of fuel tank vapors. The Notice of Proposed Rulemaking (NPRM) would require aircraft operators to reduce the flammability levels of fuel tank vapors to remove the likelihood of a potential explosion from an ignition source. Fuel tank inerting is the best solution to meeting the news [sic] standards outlined in the agency's proposal. 7
On July 23, 2006, the FAA published an Airworthiness Directive on Boeing 747 Fuel Tanks. It contained the following paragraphs: 8
During the investigation, a safety analysis of the fuel quantity indicating system (FQIS) and examinations of 747 airplanes suggested several scenarios in which an ignition source might occur inside of a fuel tank. These involved a combination of a failure or aging condition inside the fuel tank and a subsequent failure or electromagnetic coupling outside of the tank, allowing a high voltage signal on the FQIS wiring.
The 747 center tank scavenge pump assembly, which removes leftover fuel, is designed to be explosion-proof. But a review of the pump design found that unforeseen mechanical failures could cause a spark or flame to travel through the pump inlet line and ignite the fuel-air mixture inside the center fuel tank. While the scavenge pump from TWA flight 800 was never recovered, it is believed that the pump
would not have been operating at the time of the accident.
In 2008, 45 years after Leon Tanguay, director of the Civil Aeronautics Board Bureau of Safety, recommended that the FAA require fuel-tank inerting systems in all new airliners, Boeing finally began installing such systems in new airliners. However, to this day, the FAA still does not require such systems in all fuel tanks in all airliners.
Witnesses
During my late-night research, I might have copied text from another author and inadvertently forgotten to give credit. Please let me know if you find instances of such negligence and I will correct it promptly. I welcome all corrections of facts, spellings, grammar and broken links.
Schematic diagram from NTSB Accident Report. 