When the magnificent Concorde thundered down runway 26R at Paris Charles de Gaulle Airport in July 2000, it became one of the most scrutinized airplane accidents on record. After a spectacular display of flaming fuel, the aircraft climbed to about 200 feet agl, pitched up, rolled inverted and crashed. All 100 passengers, six flight attendants and three cockpit crew members were killed, along with five people on the ground.
Aircraft accident investigators often refer to the chain of errors, preventing any one of which would stop the accident from happening. The Concorde mishap is perhaps an all-time classic in this regard, since there was a clear indication that there were problems with the landing gear and tires.
Where the two airlines and two governments were concerned, modification costs and national prestige appear to have been major considerations, complicated as they were by the aircrafts marginal profitability. After a quarter century of scheduled airline service, the design deficiency that led to the accident had made itself known on several occasions.
But it was not the only error. Working back from the day of the accident, lets examine the various links in the chain.
Dispatch Error
The airplane was programmed for a takeoff brake release weight of 407,851 pounds (185 metric tons). However during taxi the airplane used less fuel than planned. Upon lineup at the end of the runway the Concorde had 2,650 pounds of excess fuel in tank number 11, the rearmost tank.
This in itself is a significant dispatch/management error. The wind was calm, yet Air France dispatch either failed to tailor the fuel required for taxi or miscalculated how much it would take. Thus the airplane was overweight at the time of takeoff.
Shortly before engine start and unknown to the flight crew, 19 undocumented bags were loaded into the rear hole at the last minute. This added an additional 1,100 pounds, which, in our view, could be construed as another sign of lax company discipline.
Those two errors meant the airplane was 3,750 pounds overweight. While that may not seem like much for a 185-ton airplane, it was a contributing link in the accident chain because of where the weight was located.
The estimated departure C.G. was 54.5 percent mean aerodynamic chord, dangerously aft of the 54 percent aft C.G. limit. The original test pilots found that near stall speed an aircraft with a C.G. aft of 54 percent would likely rear up and become uncontrollable.
Interestingly, the French Bureau Enquetes Accidents (BEA) dismissed the overload as negligible, but the remaining events in the accident chain proved the overload was critical.
The accident Concordes front fuel tank was ruptured by rubber chunks thrown up from a blown tire on the left front wheel pair. The resulting massive fuel leak caused the C.G. to move even farther aft. Combined with the loss of thrust on the left wing, these factors caused the airplane to fall grossly behind the power curve, which explains the pitch-up reported by witnesses. The airspeed fell to less than Vmc with two engines out on the same side, and the aircraft rolled toward inverted. The roll wasnt completed because the captain reduced power on the two operating engines in an effort to regain control.
Maintenance Error
As in any transport aircraft accident, investigators carefully examined the airplanes maintenance history. As is so often the case, the Concordes maintenance records contained entries that helped explain the accident.
According to David Rose, writing for the London Observer, Concorde F-BTSC went into maintenance a week before the mishap for a scheduled replacement of the left landing gear beam – the horizontal tube that holds the wheel axles. In the middle is a low friction pivot connected to the vertical leg extending down from the wing.
Areas of the pivot that support the load are reinforced by two steel shear bushes. Normally the bushes are held in position by a gray, anodized aluminum spacer, about 12 inches long and five inches in diameter. Despite the airlines required maintenance quality control checks, the aircraft left the hangar four days before the accident with this spacer missing in the left front wheel pair. In fact, after the accident the spacer was found in the workshop still attached to the old beam.
Even without the spacer, the shear bushes remained intact as the aircraft flew two round-trips to New York. On the ground the shear bushes are opposite each other. However with the landing gear retracted the right hand bush is above the left. After several up and down cycles, the right shear bush began to slip down into the gap left by the missing spacer.
By the day of the crash it had moved about seven inches, with the two almost touching. This allowed the front wheels to swivel approximately three degrees in any direction. Lacking a snug-fitting pivot, there was nothing to keep the front wheels aligned with the back wheels.
An Air and Space magazine article shortly after the accident quoted a study by retired Air France Concorde Captain Jean-Marie Chauve and former Concorde flight engineer Michel Suaud as determining the Concordes initial acceleration was abnormally slow. There was something retarding the aircraft, holding it back. They theorized it was friction from the mis-aligned front left undercarriage. Without the required spacer the left front wheels were slightly skewed on takeoff roll. From the start this caused a steady pull to the left.
Driven by the massive engine thrust, one of the mis-aligned tires failed. Chauve concluded the tire burst at about 174 knots, nine knots higher than V1, the decision speed. Only after the tire failed did it roll over the metal strip blamed by some for the tire failure.
The Chauve/Suaud report shows that the aircraft should have become airborne in 5,506 feet, well before the 5,700-foot point where the metal strip was found.
Once the tire failed, the load on the three remaining tires was unbalanced. Although the left-hand pull was manageable before the tire burst, the two front wheels were now castered hard left. The result was the overpowering left pull experienced by the crew.
The French BEA contends the left pull was due solely to asymmetric thrust from the two failed engines, despite their own photos showing skid marks from the four tires, heading off the runway toward the grass. And BEA published data show that the number two engine failed only one second before liftoff. Meanwhile the number one engine produced normal thrust until ingesting parts of a runway edge light, hit during takeoff rotation.
Retired Concorde Captain John Hutchinson told the London Observer that a double engine failure on one side is no big deal because the Concordes engines are mounted closer to the fuselage than most four-engine airplanes. Hutchinson characterized the yaw as totally containable with two engines out.
With normal acceleration and more than 152,000 pounds of thrust available, the flight could have gotten away with the overweight and out-of-balance condition, but the blown tire made a normal takeoff impossible.
Rejected Takeoff
A rejected takeoff was out of the question because the aircraft had passed decision speed. Unable to steer the aircraft and facing a potential collision with a taxiing Air France Boeing 747 looming ahead, the captain made the only logical choice and attempted to fly the aircraft out of the problem.
The Captain rotated at 187 knots – 11 knots below the programmed Vr of 198 knots. One second before liftoff the number two engine failed due to debris ingested from tire and wheel fragments.
During rotation the left landing gear struck a yellow runway edge light and the number one engine ingested steel fragments from the light fixture: Why this international airport – home base of the supersonic Concorde – had frangible runway edge lights has not yet been addressed.
Crew Factor
When the Concorde taxied out the winds were calm. However during taxi the tower advised of an eight-knot tailwind on runway 26R. With the tailwind, the airplanes brakes and balanced field length put the airplane 11,000 pounds overweight for runway 26R even before considering the unaccounted baggage and extra fuel.
A runway change for an easterly departure would have consumed the excess taxi fuel and put the balanced field length at 12,410 feet – making the 13,550-foot runway adequate.
The captains decision to continue with a downwind departure is perhaps understandable in the context of management demands for on-time departures and arrivals. A 180-degree runway change that disrupts the IFR traffic flow would have taken more than an hour, and many transport pilots will admit when pressed that they have exceeded balanced field lengths at busy airports like JFK, La Guardia, Washington National, OHare, London or Paris because of a tailwind.
Although the dispatcher had an obligation to warn the crew of the wind change and of their over-grossed condition for runway 26R and excessive balanced field length, the captain still had the ultimate responsibility.
Management Responsibility
Although the specifics of the accident flight make it clear the odds were stacked against the flight crew, another factor setting the stage for this catastrophe inevitable appears to have been known.
With the very first wheel/tire failure in 1979, a major design deficiency became clearly obvious. If that one incident was not enough, the Concorde experienced 57 tire/wheel failures during its 25 years in service. Air France experienced 30 and British Airways 27.
In 32 of the blowouts, damage to the structure, hydraulic systems and engines resulted. Fuel tanks were penetrated six times. Yet it wasnt until after the accident that a high British CAA official said, What is uniquely different in this case is that tire debris alone is thought to have led to this catastrophe.
Yet except for the fire, the July 2000 accident was almost identical to the June 1979 mishap and 30 other incidents involving serious damage over the years.
The June 1979 incident occurred during departure from Dulles International Airport. Tires 5 and 6 blew out on the left main landing gear, and debris and shrapnel punctured three fuel tanks, severed electrical wiring and several hydraulic lines, and damaged the number two engine.
A month later an almost identical mishap occurred, again on departure from Dulles. This was followed by two more U.S. incidents, one in October 1979 and another in February 1981.
In a November 1985 incident, a left-hand tire failed and both left engines were damaged. Another incident occurred during landing with another during taxi. Although the taxi mishap was thought to be caused by a locked brake, the number one fuel tank was punctured by a piece of the wheels water deflector. Thus the trends were clearly there in the accident record.
Government/Airlines
Following the two U.S. incidents in the 1979, the FAA urgently telegraphed airworthiness directives to both the British and French airlines, providing procedures for improved tire, wheel and brake checks. The French director general of civil aviation issued an airworthiness directive and Air France issued a Technical Information Update.
Both of these called for a pre-takeoff wheel and tire inspection, which included both tire pressure and temperature. In addition, if wheel or tire problems were suspected, crews were to leave the landing gear extended. Later the Concorde was equipped with roll-on wheel rims, strengthened tires and cockpit-mounted tire-failure warning lights.
In October 1979 a Concorde departing JFK suffered a tire failure and another incident occurred in February 1981 to a Concorde departing Dulles. In both cases the Air France crews ignored the tech instructions and retracted the landing gear.
Retracting the landing gear with a blown tire is dangerous in any airplane because the tire carcass and wheel are probably very hot, possibly even burning. At the very least the distorted shell can cause damage or prevent landing gear extension.
The NTSB found after the 1981 incident that the passengers had not been briefed for an emergency landing and the cockpit voice recorder had been inoperative for several weeks.
Design Changes
Lacking export orders for the Concorde, the manufacturer terminated further design refinement shortly after the airplanes were built. However in the late 1970s engine air intakes were optimized to enhance airflow, and the elevons, vertical stabilizer and rudder were modified, with fuel tank capacity increased slightly.
These changes probably enhanced the economic aspects of the airplanes operation by reducing fuel consumption and improving handling qualities, but nothing was proposed to fix the continuing problems with the wheels and tires.
After the July 2000 accident, Michelin quickly developed new tires that essentially eliminate the risk of blowouts. With these lighter radial tires, combined with other major structural improvements, the Concorde will no doubt fly for another quarter century.
It is noteworthy, however, that these changes were done only after the accidnet in order to reinstate the aircrafts airworthiness certificate.
Conclusions
The French BEA has not yet released its official report, but already French investigators have focused on the metal strip dropped by a Continental Airlines airplane shortly before the Concorde departed. Such a finding would not only protect the national honor and camouflage the very serious and obvious airline and government oversight deficiencies, but also throw the onus of financial liability on a deep-pocketed American airline.
Although the suspect metal strip on the runway may have contributed, errors by Concordes crew, dispatchers, maintenance and management helped spell the end of the Concordes 25-year safety streak. But an important factor in this tragedy was the fact that the French and British airline management opted not to modify a design deficiency that appeared to have been well-established.
When a manufacturer considers a new airplane, designers look at every possible aspect in an effort to find design flaws that could lead to a catastrophic failure. The Concorde underwent very careful study and a wonderful, futuristic airplane resulted.
However the ultimate proof of concept lies in operational experience. And in this case that experience made little difference.
-by John Lowery
John Lowery is a former Air Force pilot, accident investigator and corporate pilot.
