As English is not my first language, but French is supposed to be ... I had a pretty good idea of what I had to translate, but probably not how to put it right.
Hope it won't sound too Chinese to too many ... but welcome any comment to improve the final result.
There is in the following about 85% of the
BEA report
LAST EDIT 2009 FEB 04 :
Thanks to your inputs I did some corrections.
I also did add the paragraph translated by
RatherBeFlying, I hope he won't mind.
SYNOPSIS
On 15 November 2007, the Airbus A-346 F-WWCJ was undergoing static engine ground runs on the Toulouse-Blagnac airfield. The purpose was to test systems with technicians of the airline that had ordered the aircraft. No wheel chocks were used. On completion of these tests, after having stopped and inspected the engines, technicians have started them again for another run at high-power to find the origin of oil seepages.
Approximately three minutes after power up, the aircraft began to move forward. The technician on the left seat perceived the motion and informed the Airbus technician on the right seat. The latter acted on the brake pedals and then released the parking brake. DFDR then indicates a partial release of the brake pedals. The aircraft continuing to move forward, he tried to modify its trajectory by using the nose wheel steering. The nose wheel gear quickly skidded sideways as the aircraft accelerated.
The aircraft struck the slope of the anti-blast wall. Forward fuselage broke and flop on the other side.
There were thirteen seconds between the aircraft early move and the wall collision.
1- GENERAL INFORMATION
Information on staff
The ground tests during customer reception phase are performed under the responsibility of only one ground test technician, an Airbus employee. He was usually accompanied by one or more persons representing the client, and sometimes by other Airbus employees. Airbus had no special qualification requirement toward the client representatives attending testing. The representatives of the customer sitting in the cockpit had normally observer roles, but it could happen that the ground test technician involves a representative of the client, for example by allowing him to taxi.
During this test, the technician in charge of ground testing was on the right seat, an aeronautical technician representing the client was on the left seat and a flight test engineer was on the jump seat.
The client representative and the flight test engineer had no specific function in the aircraft handling. The role of the client representative was to observe the parameters during testing to ensure compliance with expectations.
Persons on the flight deck
Ground test technician on the right seat
Male, 41 years old, Airbus employee, responsible for the test
- Track technician since 1992
- Ground test technician since 1998
- RR Trent 500 familiarization course in May 2000
- Attached to the Flight Test Department / Reception since 2004
- Flight test engineer since 2004
- Recurrent training for A-330/340 engine test in October 2006
Aeronautical technician on the left seat
Male, 36 years old, employee of a maintenance company (GAMCO), which receives and maintains the Etihad Airline fleet.
- Technician for the GAMCO company since 1997
- Courses at Lufthansa Technik and Airbus in 2002
- A-346 static engine ground run training in 2006
Flight test engineer on the jump seat
Male, 42 years old, Airbus employee
- Flight test engineer in 2000
- Attached to the Flight Test Department / Reception since 2000
- Authorized to perform engine tests on Airbus family
- Airplane commercial pilot in 1998
- A-320 type rating in 2004
- ATR-42 type rating in 2006
Engine parameter reference
The thrust of the A-346 engines is measured with the EPR (Engine Pressure Ratio) which represents the ratio of pressure between the total output of the turbine and compressor inlet. This ratio varies approximately between 1 (ground idle) and 1.41 (full thrust, or around 28000 daN).
Weight and balance
The aircraft was at 223 tons including 40 tons of fuel, and the CG was at 25.8%. Ground tests are usually performed with 80 tons of fuel. The maximum certified take-off weight is 380 tons.
Braking system
The A-346 has two Main LG, one on the right side and one on the left, one Central LG and one Nose LG.
Each MLG and the CLG have 4 wheels each. CLG is slightly behind both MLG. Each MLG wheel and CLG wheel is equipped with a braking system, and each brake is powered by two independent hydraulic systems. The NORMAL braking is controlled through the green system. The blue system powers the ALTERNATE braking.
When the parking brake is set, the blue system pressurizes both MLG at 2500 psi. The CLG brakes are not pressurized by the parking brake.
When the brake pedals are pressed, the green circuit is pressurized on both MLG and on CLG with a record depending of the pedals position. The green circuit pressurization is inhibited as long as the parking brake is activated.
If you release the parking brake and simultaneously you press the brake pedals, the system allows both circuits to be pressurized together, while the ALTERNATE circuit depressurizes. This applies only to both MLG and the total amount of pressure from both circuits is limited to 2770 psi.
In addition, the braking of the CLG wheels is automatically reduced when the nose wheels are steered. From a turning order of 20 degrees, the CLG braking is completely inhibited.
Regulation for certification indicates that the parking brake must be designed to prevent the aircraft from moving on a dry runway with one engine at maximum thrust, the others being at ground idle. In these circumstances, the A-346 parking brake must develop a minimal braking force of 28000 daN or 3500 daN by braking wheel. The system was designed to develop a braking force of 8500 daN by braking wheel with a brake pressure of 2500 psi.
Weather conditions
Wind 330/16
Temp 5C
Dew point –5C
QNH 1019
Aerodrome information
The accident occurred on the BIKINI ramp. The area dedicated to testing, is part of the manufacturer's facilities.
No grip data for the surface of the test area was available before the accident. To enable a quantitative analysis of braking performance, it was necessary to undertake measures of slipperiness. These measures were carried out in conditions close from the day of the accident. The friction coefficients were measured between 0.65 and 0.68. These values correspond to the coefficient of a dry track in good condition.
Data recorders
The CVR and FDR have been synchronized using the UTC time registered in the FDR and the “Master Caution” “Single Chime” identified on the CVR.
The aircraft arrived at the BIKINI area around 14:19
It was at a magnetic heading of 312 degrees. The parking brake is set and active.
During the tests between 14:19 and 14:58 the maximum EPR values are between 1.04 and 1.22
The last static engine ground run is started at 15:58
The aircraft is still at full stop.
Between 15:58:10 and 15:59:03 the thrust is increased from idle to a steady value of 1.25 EPR. This engine thrust setting corresponds to a position of thrust levers between MCT (Max Continuous Thrust) and MTO (Max Take Off Thrust).
The ALTERNATE pressure values are close to 2600 psi for the wheels 1,2,5,6 (left gear) and 3,4,7,8 (right gear). They are at 64 psi for the wheels 9,10,11,12 (central gear).
At 16:02:06 the person on the right seat starts talking but is interrupted at 16:02:08 by the person on the left seat who announces :
“euh ... cabin is ... aircraft is moving forward”
The first significant LONGITUDINAL ACCELERATION values showing a forward acceleration of the aircraft are observed around 16:02:07
The recorded ground speed starts to increase at 16:02:09
Between 16:02:08 and 16:02:13 the ground speed increases from 0 to 4 kt.
At 16:02:11 the person on the left seat repeats :
“Aircraft is moving forward”
An action on the brake pedals is recorded from around 16:02:11
The parking brake is deactivated around 16:02:13
The person on the right seat announces :
“parking brake off”
From the moment the park brake is released:
- brake pedals are briefly released on two occasions
- the ALTERNATE braking pressures go below 192 psi
- the NORMAL braking pressures from MLG are consistent with the brake pedals position on both right and left sides, and increase from 300 to 2500 psi in one second
- the NORMAL braking pressure from CLG reach a maximum of 192 psi at 16:02:14 and then decreased to 64 psi and stabilizes at that value
- the wheels speed values which were still recorded to zero (sensors do not work until a wheel speed of 3 to 5 kt) become positive and are consistent with recorded ground speed and aircraft movement
- recorded ground speed increases rapidly from 4 to 31 kt in seven seconds
Between 16:02:13 and 16:02:15 the order given to the right NWS (Nose Wheel Steering) goes from 0 to 75 degrees (full right order). The evolution of the nose wheel angle until impact is consistent with that order. From 16:02:15 the magnetic heading of the aircraft begins to increase; it goes from 312 to 349 degrees in seven seconds.
The angle of the nose gear reaches 77 degrees right at 16:02:19 and remains to that value until the end of recording.
From 16:02:18 we can hear on the CVR severe vibration noises followed by impact noises.
The thrust levers do not move until 16:02:20 when they are retarded in the IDLE detent. The EPR values of the 4 engines start to decrease immediately afterward.
The longitudinal acceleration becomes significantly positive, indicating an aircraft deceleration, around 16:02:20.5
FDR recording ends between 16:02:21 and 16:02:22
CVR recording ends at 16:02:23
Information on the wreckage site
The aircraft was involved in a collision with the anti-blast wall north of the BIKINI area. It came to rest leaning on the wall, heading north. The tail cone and the tip of the right wing are in contact with the ground. Only the right MLG touches the ground.
The aircraft struck the anti-blast wall at an angle of about 30 degrees. The underside of the forward cabin was torn on about fifteen meters and was folded to the ground when passing the anti-blast wall.
The cockpit was crushed against the ground north of the wall. The avionic bay including most of the flight calculators, below the cockpit, has been completely destroyed.
ENG 1 and 2 hit the wall and have many damages. The pylon 2 is twisted. ENG 3 and 4 kept running after impact and did not stop immediately. It was not possible to stop them by pushing the FIRE pushbuttons or by switching OFF the ENG MASTER switches.
Water and foam spray on ENG 4 managed to extinguish it at 18:48
Due to the proximity of the wall a similar scenario was not possible with ENG 3 which extinguished by itself only on November 16 at 01:25 when he had consumed all its fuel supply.
The NWG is broken and separated from the fuselage. The wheels are oriented to the right and have a steering angle close to the maximum value. The wheel tires are tapped and show signs of perpendicular friction to the tread.
Ground tire marking
For the following descriptions, the distance reference is taken from the point of impact on the wall, and up the aircraft trajectory.
A first trace of tire corresponding to one of the internal wheels of the Right MLG is visible at 120 meters over a length of approximately 10 meters. The trace of the external tires is present but less marked. Those marks are directed along an axis with the magnetic course of 330 degrees. No trace of the Left MLG tires was observed.
At 83 meters, we can see the first NWG marks. They curve toward a northerly course, are initially parallel then at 50 meters, converge to leave only one single trace. The NLG is no longer directional.
Symmetrical braking traces from both MLG are present at around 60 meters and down to the wall.
Video Camera
The recording of a video camera permanently filming the BIKINI area has been exploited. You can see the plane during the last test. At first there is a plane slow translation, then a sudden accelerating movement.
While the path begins to slowly turn right, the NLG starts skidding sideways. The plane continues on its path to the wall.
The forward section rises, falls back on the wall and the fuselage breaks. There are flames at ENG 1 and 2 level and on the aft section of the aircraft.
By looking at recorded shots several days before the accident,
we can see that some tests are carried out with wheel shocks and some others without.
Analysis of braking force and surface grip
Braking force
For each of the braking wheels, the maximum braking force created by the brake pressure is determined based on the specification of brakes, depending on the recorded pressure. The overall braking force is obtained by the total amount of braking forces from the 12 wheels. When the parking brake alone is used, the brake pressure on the CLG wheels is void and only the MLG wheels contribute to braking.
Slip resistance force
For each of the wheels, the strength of slip resistance is equivalent to the supported weight by the wheel multiplied by the friction coefficient μ tire/bitumen. The simulation is used to compute the limit friction coefficient value below which the wheels would slip, under certain mass distribution assumptions. In the same way, the forces of slip resistance for each of the wheels are added to obtain the overall slip resistance force.
Engines thrust
The engines thrust was calculated from the recorded EPR parameters and from manufacturer data, based on the day conditions (320 ft, nil speed, ISA -9C, no bleed air from engines).
It has stabilized at around 83500 daN.
Results
The model is used to calculate the theoretical changes in thrust, the maximum braking force developed by the braking system and compare the slip limit force from which the wheels start to slip.
For the aircraft to remain motionless, it is necessary that the thrust is less than both the maximum braking force developed by the system and slip limit force.
Throughout the last test, the engines thrust and the maximum braking force on the parking brake are very close. To obtain under the same conditions a slip limit force equivalent to the thrust force, a friction coefficient μ of 0.687 is necessary. Given the measured friction coefficient values, it is reasonable to believe that the aircraft was quickly on the edge of the slip.
The fact that a balance, even fragile, has existed for about three minutes confirms that the brakes were functioning in accordance with their specifications.
Therefore, modeling allowed to establish, with a reasonable confidence level, that during the last test the thrust and braking forces were neutralized, but that the balance of those forces was particularly precarious.
The aircraft remained motionless with 8 wheels through the parking brake, then started moving. Several factors may have contributed to the aircraft movement, notably :
- the vibrations created by the engines
- the reduction of weight due to fuel consumption (about 1270 kg)
- a slight local brake pressure reduction on one of the wheels
When the parking brake has been released, the application of the brake pedals has never permitted to reach the same level of braking despite the fact that the 12 brake wheels became solicited. This is due to two factors: firstly, the actions on the brake pedals have not been sustained at the maximum level, and, secondly, the action on the NWS very quickly lead to inhibit the CLG braking. The resulting braking during the motion varied between 65 and 95% of the braking level obtained before the aircraft movement.
Information on organizations and management
In the manual delivery, the seat occupied by the representative of the customer during ground test is not specified. In practice, the representative of the customer is generally on the left seat. He takes note with the Airbus technicians of the parameters to be monitored.
The Aircraft Maintenance Manual (AMM) and the CAM (Customer Acceptance Manual)state that the engines tests must proceed with the use of wheel shocks for the MLG.
The CAM uses tasks listed in the AMM. Thus, there is in the AMM a procedure to search for oil leaks called “The Fuel and Oil Leak Test”. This procedure calls for testing with two engines in operation. These engines must be the symmetrical engines of each wing. For the engine on which the oil leak is sought, it must be applied the maximum thrust value of the day (1.25 EPR that day, corresponding to the maximum value for this type of test). For the opposite engine the ERP value to be applied is 1.145
These actions are carried out by memory.
Formation of the ground technicians
After a theoretical phase of familiarisation with the aircraft systems, trouble-shooting and functional tests, ground test technicians take training courses for cabin crew, courses in radiotelephony and taxiing. The next phase is practical while the trainee works with an instructor.
The following points are notably addressed:
- conduct of ground test
- ground taxi
- participation in accelerate-stops
- production flights
- conduct of tests in Customer Acceptance Manual and associated tests
- use of technical documentation and software
- training in pressurization tests
Practical parts of this training are accomplished in the airplane and in the simulator.
A refreshment session is performed in a simulator every two years for ground test technicians.
Additional information
Testimonies
Ground test technician on the right seat
After conducting the ground tests, the client company technicians had noticed an oil seepage on the pipe of one engine. The ground test technician had decided to proceed for another test before leaving.
He increased the thrust in order to heat up the engine oil. After approximately three minutes, as he was looking inside the cockpit, he heard the person sitting on the left seat announcing that the plane was moving. He then noticed himself the movement. He removed the parking brake to use the normal braking. The aircraft continuing to move forward he thought having a brake problem. He then tried to change the trajectory of the aircraft using the NWS. He specifies to have often carried out that kind of test, but at a higher weight.
Aeronautical technician on the left seat
During a last test at high thrust, he perceived the motion of the aircraft by looking outside. He felt hiccups caused by the brakes, he believes. He noted the ineffectiveness on the NWS action. It did not touch any control.
He specifies the high thrust test was performed to loosen the oil in order to detect any seepage.
Flight test engineer on the jump seat
He attended as an observer for his first A-346 delivery. During the last high thrust test, he heard the person on the left seat announcing that the plane was moving. Given the vibrated environment generated by the high thrust setting, he was unable to perceive accelerations caused by the aircraft movement. He only remembers to have grabbed and retarded the thrust levers when seeing the wall very close.
The ground personnel
They watched the end of the tests. Before leaving the testing area, the technician on the right seat told them to move further as he was going to proceed for a last high thrust test. They positioned their car in front of the aircraft then moved further on the side when thrust was set. They then saw the plane moving, initially slowly then faster. They saw it hit the wall. They gave the alert.
The use of wheel chocks is not systematic because their use is binding. Sometimes, it happens they stay trapped under the tires after the tests. This requires to push back the plane to remove them.
Note: It is clear from discussions with other Airbus technicians that this kind of test, including on 4 engines at high thrust, are frequent. All confirm that the use of wheel chocks is not systematic. Finally many underline the pressure from the customers to go and check some details. This leads sometimes to conduct tests outside the scope of the CAM.
2- ANALYSIS
Test procedure
Although the reference documents require the establishment of wheel chocks during engine tests, the investigation showed they were not systematically used. Similarly, during the test for oil leaks detection, it often seems that the procedure to apply thrust on 2 engines only is not respected.
The industrial and commercial issues that are associated to the delivery activities may lead to induce timely pressure on testing technicians during this phase.
The presence of representatives of the customer on board during the delivery phases can create pressures inducing testing technicians to come out of their frame of reference.
Reactions in the cockpit
The ground test technician resources were mobilized for about ten seconds on the braking system. He did not think to retard the thrust levers. This can be explained by focusing on the braking problem, by the dynamics of the situation and by the lack of training in this kind of situation.
The persons on the left seat and jump seat were present only as observers. The aeronautical technician on the left seat did not intervene on the controls until impact.
The flight test engineer intervened, but late, to retard the thrust levers. This can be explained by its statute, the fear of interfering with the actions of the technician and also by the dynamic of the situation.
3- CONCLUSIONS
Findings of investigation
- The aircraft, including its braking system operated in accordance to specifications
- The accident occurred in the delivery phase of a not programmed test
- The procedure was not conformed to the task “The Fuel and Oil Leak Test” listed in the AMM. It was carried out particularly at high thrust on all engines without the use of wheel shocks
- Testimonies and video recordings indicate that engine tests without wheel shocks are regularly practiced
- The thrust used on the engines was at the same level as the nominal braking capacity of the parking brake
- While the aircraft began to move, the ground testing technician pushed on the brake pedals and released the parking brake
- The ground testing technician turned the NWS to the right. This action, by inhibiting the CLG braking, has limited the braking effectiveness
- The actions on the brake pedals were not sustained to the maximum level
- The Flight testing experimenter has retarded the thrust levers when the plane hit the anti-blast wall
Causes of the accident
The accident is due to the realization of a run up on all 4 engines at the same time, without wheel chocks, and during which the thrust was close to the parking brake capacity.
The lack of a detection process and deviation correction in the ground test procedure, in a context of permanent industrial and commercial pressure, did promote the realization of a test outside of the established procedures.
The surprise led the ground testing technician to focus on the braking system; therefore he did not think to reduce the thrust of the engines.
Measures taken following the accident
Customer Acceptance Manual has been revised (May 2008) to strengthen
the instructions to follow when conducting a run up. This includes :
- install wheel chocks to all MLG wheels (and all CLG wheels if applicable)
- mandatory presence of two qualified persons to the controls during run up and taxing.