GLF4, Bedford MA USA, 2014
From SKYbrary Wiki
|On 31 May 2014, a Gulfstream IV attempted to take off with the flight control gust locks engaged and, when unable to rotate, delayed initiating the inevitable rejected take off to a point where an overrun at high speed was inevitable. The aircraft was destroyed by a combination of impact forces and fire and all seven occupants died. The Investigation attributed the accident to the way the crew were found to have habitually operated but noted that type certification had been granted despite the aircraft not having met requirements which would have generated an earlier gust lock status warning.|
|Actual or Potential
|Fire Smoke and Fumes, Runway Excursion|
|Aircraft||GULFSTREAM AEROSPACE Gulfstream 4|
|Type of Flight||Public Transport (Passenger)|
|Origin||Bedford/Laurence G. Hanscom Field|
|Intended Destination||Atlantic City International Airport|
|Take off Commenced||Yes|
|Flight Phase||Take Off|
|Location - Airport|
|Airport||Bedford/Laurence G. Hanscom Field|
|Tag(s)||Ineffective Regulatory Oversight|
|Tag(s)||Post Crash Fire|
|Tag(s)||Procedural non compliance|
|Tag(s)||Overrun on Take Off,|
RTO decision after V1,
Unable to rotate at VR
|Contributor(s)||OEM Design fault|
|Damage or injury||Yes|
|Aircraft damage||Hull loss|
|Fatalities||Most or all occupants (7)|
|Causal Factor Group(s)|
On 31 May 2014, a Gulfstream IV (N121JM) registered to SK Travel and being operated under lease by Arizin Ventures on a private business flight from Bedford MA to Atlantic City NJ initiated a high speed rejected take off in normal ground visibility at night and subsequently overran the end of the 2140 metre-long runway 11 by approximately 570 metres. During the overrun, the aircraft collided with ground obstructions before ending up in a gulley where a post crash fire consumed most of the aircraft wreckage. The three crew and four passengers died and the aircraft was destroyed.
An Investigation was carried out by the NTSB. The FDR, CVR and QAR were recovered from the wreckage and data from all three was successfully downloaded. The QAR data covered 303 hours of data on 176 flights. A Preliminary Report on the initial progress of the Investigation was published on 13 June 2014.
The continuing investigation established that the 45 year-old aircraft commander and the 61 year-old Co-pilot were both qualified to act as pilot in command of the accident aircraft and constituted the usual flight crew for the operator's only G-IV. They "customarily changed seats between flights" with the pilot occupying the left seat acting as the commander. The 61-year old Co-pilot was also the Chief Pilot and Director of Maintenance for SK Travel. Both had been flying the accident aircraft type since it had been acquired by the operation seven years previously and both had, earlier in their careers, logged time on business jets other than the accident type. Available evidence indicated that the aircraft commander had been PF.
The accident flight was being made to return the four passengers to Atlantic City from where they had arrived in the same aircraft flown by the same crew several hours earlier. The two pilots and the Cabin Attendant had remained with the aircraft. Once the passengers had boarded, the engines were started and the aircraft taxied to runway 11, reaching it about eleven minutes after the passengers had boarded. During this time, it was noted that the CVR had "recorded minimal verbal communication between the flight crewmembers, and there was no discussion or mention of checklists or takeoff planning" whereas the AFM "includes five checklists to be completed before takeoff: the Before Starting Engines checklist, the Starting Engines checklist, the After Starting Engines checklist, the Taxi/Before Takeoff checklist, and the Lineup checklist."
The 'Starting Engines' Checklist included the requirement to disengage the gust lock, achieved by moving the gust lock handle on the pedestal between the pilots (see picture below) to the OFF (down) position. The 'After Starting Engines' Checklist included a check of the flight controls which is achieved by one of the pilots moving the elevators, ailerons, and rudder from stop to stop to confirm that they move freely and correctly. FDR data showed that "the flight crew did not complete a flight control check after engine start or at any time thereafter".
As the aircraft turned on to the runway, the CVR recorded the pilots discussing the appearance of the “RUDDER LIMIT" caution on the Engine Indicating and Crew Alerting System (EICAS). Such an annunciation would be indicative of the rudder contacting its stops or "being prevented from reaching its commanded position, as would be the case if the rudder was commanded to move with the gust lock engaged". However, the AFM stated that this annunciation is advisory with no crew action required and it was noted that although it appeared because of the interaction of the engaged gust lock with the yaw damper system, it "is not normally used to alert flight crews to the status of the gust lock system" and did not do this.
Following brake release for take off, with a likely target EPR of 1.7, FDR data showed that the throttle levers were advanced manually to an EPR of "about 1.42 over a period of about 4 seconds". Then after an interval of about 5 seconds, the A/T was engaged and the EPRs began to increase again to reach their eventual maximum values of about 1.6 as the aircraft reached about 60 knots before beginning to drop to about 1.53 and stabilising as the PF said "couldn’t get (it manually any further).”
A review of the available QAR data from previous flights over the previous 9 months found that, out of the 176 recorded takeoff events - all made by the accident aircraft flight crew - only 2 had been preceded by a complete flight control check. Data from each take off was examined to note the airspeed at the time that the elevators (and therefore the control columns) had begun to move aft during the takeoff roll and in all cases except for the accident takeoff, this occurred at between 60 and 80 knots. The data for the immediately preceding 20 flights was also examined to see how the accident flight crew had advanced the thrust levers during take off and it was found that in each case, this action had been continuous to the same final position with in each case a corresponding resultant 1.7 EPR (maximum thrust takeoff) set before engaging the A/T.
It was noted that the AFM version of the 'Lineup' Checklist (although not the one recovered from the wreckage) is followed by a note that, at 60 knots during the take off roll, the pilot should confirm that "the elevators are free and the control yoke has moved aft from the full forward to the neutral position". The basis for this is that aft movement of the control yoke (column) should be evident as aerodynamic force on the elevators increases because of their 'at rest' position of about 13° trailing edge down. As the accident aircraft passed 60 knots, FDR data recorded the elevator (and therefore the control columns) as still at about the 'at rest' position.
Calls of 'eighty', 'V1' and 'rotate' followed from the PM. One second after the 'rotate' call, the PF said "...lock is on." and then repeated this six times over the next 13 seconds. Five seconds after the first of these calls, FDR evidence indicated that one of the pilots had moved the Flight Power Shutoff Valve (FPSOV) (see diagram below for its location) to the 'on' position, considered to have probably been an attempt to achieve unlocking of the flight controls whilst the crew were still unaware of the 'on' position of the gust lock lever. It was noted that whilst this action would have removed hydraulic pressure from spoiler and primary flight control actuators, it would not have had the desired effect on locked status of the elevator.
About 11 seconds after the 'rotate' call with the aircraft having reached its maximum groundspeed of about 162 knots and with about 420 metres of the 2140 metre-long runway remaining, the FDR recorded the left and right brake pressures beginning to rise. There was no verbal communication between the pilots about rejecting the takeoff and only four seconds after the brake pressure increase were the thrust levers retarded. Normally this latter action would have led to the ground spoilers being deployed to assist deceleration but the earlier movement of the FPSOV had disabled them due to the consequent loss of actuator hydraulic power. Five seconds after reaching its maximum ground speed and one second after the PF says “I can’t stop it.”, the thrust reversers were deployed as the aircraft exited the runway onto the paved RESA at just over 150 knots. One second after reaching the end of the 311 metres of RESA at 105 knots and going onto the grass, CVR data recorded the first impact sound and FDR data ended shortly afterwards with the aircraft still moving at about 90 knots. Having passed through the perimeter fence, the aircraft stopped spanning a small river valley about 570 metres beyond the end of the runway.
The ILS LOC antenna located about 200 metres beyond the end of the paved RESA was found to have been demolished along with a number of its support poles which had been severed 45cm above their bases. Blackened and charred grass between the LOC antenna and the finally-stopped position indicated that this impact had initiated the fire.
A contract pilot stated to the Investigation that although he had flown with the accident PF some years previously and regarded him as a 'good pilot', he habitually "did not use a formal item-by-item checklist". The same pilot also stated that "when the gust lock was not disengaged before starting the engines, some pilots occasionally used the FPSOV handle to momentarily remove hydraulic pressure from the flight controls, which allowed the gust lock to be disengaged without shutting down the engines" but this comment was expressly not attributed to a specific pilot or flight crew.
Examination of all the evidence available led the Investigation to conclude that the take off had been attempted with the gust lock lever (see the illustration below) in the 'on' position. A review of the crew actions noted that when the EPR did not reach the expected level because the throttle position was constrained by the resulting gust lock/throttle lever interlock, the PF had not immediately rejected the takeoff but had instead engaged the A/T. This had allowed the EPR to increase towards (but not reach) the target value and the thrust levers to move towards (but not reach) their normal position. Then, when the PF (must have) realised that the controls were locked, he "still did not immediately initiate a rejected takeoff". The Investigation concluded that "if the flight crew had initiated a rejected takeoff at the time of the PF’s first 'lock is on' comment or at any time up until about 11 seconds after this comment, the airplane could have been stopped on the paved surface". Instead, initiation of a rejected takeoff was delayed to the extent that the accident became unavoidable.
It was concluded that ineffective intra flight crew communication and non-use of the prescribed challenge and response Normal Checklists had been contributory to the failure to unlock the controls but also noted that oversight of the operation, including voluntary and independent assessments of operational safety had failed to disclose that such behaviour appeared to have been routine rather than exceptional for the crew concerned.
It was also concluded that the gust lock/throttle lever mechanism did not perform as it was required to under generic type certification requirements and that had it done so, it would have prevented the aircraft from achieving any significant acceleration which would have led to "an unmistakable warning that would most likely have prevented the accident".
The Probable Cause of the accident was determined to have been "the flight crew members’ failure to perform the flight control check before takeoff, their attempt to take off with the gust lock system engaged, and their delayed execution of a rejected takeoff after they became aware that the controls were locked".
Three Contributory Factors in the accident were determined as:
- the flight crew’s habitual noncompliance with checklists.
- Gulfstream Aerospace Corporation’s failure to ensure that the G-IV gust lock/throttle lever interlock system would prevent an attempted takeoff with the gust lock engaged.
- the Federal Aviation Administration’s failure to detect this inadequacy during the G-IV’s certification.
Safety issues identified during the Investigation were as follows:
- the need for flight crew use of the challenge and response checklist execution.
- the importance of using FOQA data analysis to define the scope of procedural noncompliance in business aviation.
- the need to ensure that any objects along the extended runway centreline up to the perimeter fence are frangible.
- the need to retrofit the gust lock system on all existing G-IV airplanes to comply with the existing certification requirement that the gust lock must limit the operation of the airplane so that the pilot receives an unmistakable warning if it is engaged at the start of take off.
- the need for guidance on the appropriate use and limitations of the use of engineering drawings in a design review carried out to show compliance with regulatory requirements for aircraft type certification.
Five Safety Recommendations were made:
- that the Federal Aviation Administration should identify non-frangible structures outside of a runway safety area during annual 14 Code of Federal Regulations Part 139 inspections and place increased emphasis on replacing non-frangible fittings of any objects along the extended runway centerline up to the perimeter fence with frangible fittings, wherever feasible, during the next routine maintenance cycle. (A-15-30)
- that the Federal Aviation Administration should, after Gulfstream Aerospace Corporation develops a modification of the G-IV gust lock/throttle lever interlock, require that the gust lock system on all existing G-IV airplanes be retrofitted to comply with the certification requirement that the gust lock physically limit the operation of the airplane so that the pilot receives an unmistakable warning at the start of takeoff. (A-15-31)
- that the Federal Aviation Administration should develop and issue guidance on the appropriate use and limitations of the review of engineering drawings in a design review performed as a means of showing compliance with certification regulations. (A-15-32)
- that the International Business Aviation Council should amend International Standard for Business Aircraft Operations auditing standards to include verifying that operators are complying with best practices for checklist execution, including the use of the challenge-verification-response format whenever possible. (A-15-33)
- that the National Business Aviation Association should work with existing business aviation flight operational quality assurance groups, such as the Corporate Flight Operational Quality Assurance Centerline Steering Committee, to analyze existing data for noncompliance with manufacturer-required routine flight control checks before takeoff and provide the results of this analysis to its members as part of its data-driven safety agenda for business aviation. (A-15-34)
The Final Report was adopted by the Board on 9 September 2015 and subsequently published.
- Runway Excursion
- Beyond the Runway End Safety Area
- Post Crash Fires
- Standard Operating Procedures (SOPs)
- Normal Checklists and Crew Coordination (OGHFA BN)
- Safety Management System
- Flight Data Monitoring (FDM)
- Flight Control Locks (Gust Locks)
- Crew Resource Management
- The Human Factors "Dirty Dozen"
- Continuation Bias