If you wish to contribute or participate in the discussions about articles you are invited to join SKYbrary as a registered user

 Actions

EC35, vicinity Glasgow City Heliport UK, 2013

From SKYbrary Wiki

Summary
On 29 November 2013, control of an Airbus Helicopters EC135 undertaking a night VMC night for policing purposes was lost after both engines flamed out following fuel starvation. The subsequent crash killed the three occupants and seven on the ground, seriously injuring eleven others. The Investigation found that although the pilot had acknowledged low fuel warnings after both fuel transfer pumps had been switched off, the helicopter had not then been landed within 10 minutes as required. No evidence of any relevant airworthiness defects was found and without FDR/CVR data, a full explanation of the accident circumstances was not possible.
Event Details
When November 2013
Actual or Potential
Event Type
Human Factors, Loss of Control
Day/Night Night
Flight Conditions VMC
Flight Details
Aircraft AIRBUS HELICOPTERS EC-135
Operator Babcock Mission Critical Services Onshore
Domicile United Kingdom
Type of Flight Aerial Work
Origin Glasgow City Heliport
Intended Destination Glasgow City Heliport
Take off Commenced Yes
Flight Airborne Yes
Flight Completed No
Flight Phase Manoeuvring
MNV
Location - Airport
Airport vicinity Glasgow City Heliport
General
Tag(s) Helicopter Involved
HF
Tag(s) Procedural non compliance
LOC
Tag(s) Loss of Engine Power
Outcome
Damage or injury Yes
Aircraft damage Hull loss
Non-aircraft damage Yes
Non-occupant casualties Yes (7)
Fatalities Most or all occupants (3)
Causal Factor Group(s)
Group(s) Aircraft Operation
Safety Recommendation(s)
Group(s) Aircraft Airworthiness
Investigation Type
Type Independent

Description

On 29 November 2013, an Airbus Helicopters (formerly Eurocopter) EC135 T2+ being operated for policing purposes by Bond Air Services from the Glasgow City Heliport at night in VMC crashed onto a bar in the city centre, partially demolishing it and resulting in the deaths of all three occupants and seven people who had been in or adjacent to the building. A total of 32 people on the ground were injured, 11 seriously. Although the helicopter was destroyed, the main fuel tank remained intact and there was no post crash fire. No indication of an abnormal or emergency situation on board was communicated prior to the accident.

Investigation

An Investigation was carried out by the UK AAIB. Special Bulletins S9/2013 and S2/2014 were published on 29 November 2013 and on 14 February 2014 respectively to advise on progress in the Investigation.

It was noted that the accident flight had taken place under a Police AOC and the accident flight had therefore been made using a 'State' aircraft. In respect of oversight by Safety Regulators, the effect of this was to make the operation subject to UK National rather than EASA Regulation. Operator Bond Air Services was noted to be a major provider of helicopters and pilots for both Police and Air Ambulance (HEMS) operational use in the UK, deploying a fleet of in excess of 26 aircraft including 23 EC135s. Two of these EC135s were based at the Glasgow City Heliport, one for Police Scotland (the accident aircraft) and the other for the Scottish Ambulance Service. The 51 year old pilot was found to have been an experienced ex-military helicopter pilot who had joined Bond Air Services and completed his EC135 type rating some five years earlier after leaving the Royal Air Force.

The helicopter was not fitted - nor was it required to be fitted - with either an FDR or a CVR. However relevant data was recovered from the Non Volatile Memory in some on-board equipment provided for engineering purposes and ATC radar and R/T recordings and CCTV recordings were available. The time-stamps on the radar recording were used to align the different data sources. Weather conditions throughout the flight were good with little or no cloud and excellent in flight visibility, although it was a moonless night.

It was established that the aircraft had left its base at the Glasgow City Heliport with the pilot and two police observers on board and 400kg of fuel in tanks. Various surveillance tasks were undertaken and all ATC and tactical communications normal. Then, a little over 20 minutes prior to the fuel starvation which led to the loss of control, and a few minutes after it was estimated that both fuel transfer pumps in the main fuel tank, which transfer fuel to the supply tank which feeds the engines had been switched off, 'LOW FUEL' Warnings for both engines began to occur. All LOW FUEL Warnings were accompanied by an aural attention-getter, which was cancelled after each annunciation, indicating that the pilot would have been aware of all of them. These warnings, initially intermittent, became permanent and continued until the end of the flight. ATC were not advised of "any fault with the helicopter or any other concern". These indications meant that there would have been approximately 32 kg / 28 kg of fuel remaining respectively in the left and right supply tanks, respectively which is why the AFM instructed pilots to ‘LAND WITHIN 10 MINUTES on receipt of such warnings.

Despite this, the helicopter had continued to complete three further 'routine surveillance tasks' over a period of 13 minutes before the pilot advised ATC that tasking was complete and the helicopter was returning to its base at Glasgow City Heliport. ATC responded by confirming the previously issued clearance to operate in the Glasgow CTZ not above 2000 feet QNH, at which point, the helicopter had been airborne for 93 minutes.

There were no further radio transmissions to ATC from the pilot and the helicopter proceeded on an approximately westerly track towards the heliport at around 1000 feet QNH making a ground speed of approximately 100 knots. Then, about 40 seconds prior to the loss of control and approximately a mile before the accident site (and 2.7nm from the destination heliport), the right engine flamed out. The memory action to select the failed engine off was not taken and 32 seconds later, the left engine also flamed out and the rate of descent which had followed the initial engine failure began to increase. At this point, the helicopter was estimated to have been between 500 and 700 feet agl which would have left relatively little time to transition to a successful autorotation. Memory items to achieve such an autorotation were not actioned and rotor N1 decreased below the minimum recoverable rotor rpm in autorotation, which it was considered was "likely to have been the result of collective pitch being applied". It was noted that, in the absence of action to select the SHED BUS to the Emergency position, both the RADALT and steerable landing light would have been inoperative. Overall, the evidence available did not make it possible to establish "why a more successful autorotation and landing was not achieved, albeit in particularly demanding circumstances".

Eye witnesses reported having heard "noises similar to a 'misfiring car' and a 'backfire', as well as numerous 'bangs' followed by silence as the helicopter descended very rapidly and in a level attitude before striking and penetrating the roof of a crowded single storey Public Bar. The roof of the building collapsed and the helicopter entered the space below with a calculated "decelerative force of 70g". There was no fire but as a result of the impact, the three occupants of the helicopter (the pilot and two police observers) were killed as were seven people in the bar. On the ground, a further 11 people sustained serious injuries and 20 others sustained minor injuries. A reconstruction of the flight path of the helicopter immediately prior to the accident is shown below.


Flight path at the end of the flight (Reproduced from the Official Report)


An extensive examination of the wreckage following its removal to AAIB facilities found no evidence of any airworthiness defects which might have had a bearing on what happened.

A total of 76 kg of fuel was recovered from the undamaged main fuel tank but the engine supply tank was found to have been empty at the time of impact. Wreckage examination and testing showed that both of the fuel transfer pumps in the main tank which are normally on in order to achieve fuel transfer to the supply tank "had been selected off for a sustained period before the accident, leaving the fuel in the main tank, unusable". It was noted that these switches "were of the simple unguarded toggle type" and also in close proximity to the engine prime pump switches, which were also unguarded and found after the crash to be in the 'ON' position whereas they would normally be selected OFF and remain so following engine start. It was concluded that whilst it was not possible to determine how the prime pump switches came to be in the 'ON' position, "unintentional selection of the inappropriate switches was possible.

It was found that the accident helicopter type had accumulated more than 3 million flying hours over about 20 years without any previously reported instances of fuel starvation. The Investigation "could not establish why a pilot with over 5,500 hours flying experience in military and civil helicopters, who had been a Qualified Helicopter Instructor and an Instrument Rating Examiner, with previous assessments as an above average pilot, did not complete the actions detailed in the Pilot’s Checklist Emergency and Malfunction Procedures for the LOW FUEL 1 and LOW FUEL 2 warnings". It was also considered that given that the applicable Final Reserve Fuel defined by the Operator was 85kg, "the pilot might have been expected to make a PAN call, upgrading it to a MAYDAY on reaching the Final Reserve Fuel IFR, if he was aware of the fuel state". However, as with other aspects of the Investigation, "due to the lack of evidence that might otherwise have been provided by cockpit voice and flight data recorders, the investigation was unable to determine the reasons for this apparent omission".

During the course of the Investigation, a number of issues relevant to the accuracy of fuel quantity indications but not central to the issues raised were identified and corresponding safety action was taken. The ancillary function of the suitably trained front seat observer as a pilot assistant during operations conducted under a Police AOC was noted, in particular that such observers "do not operate helicopter related system switches, unless in an emergency situation and at the request of the pilot". It was noted that at the time the Investigation was being written up, 32 helicopters were being operated under UK-issued Police AOCs".

Overall, the primary concern of the Investigation was the inability to establish with certainty an explanation of the sequence of events which had culminated in the accident. This was attributed to the lack of carriage of both an FDR and a CVR and the absence of any corresponding regulatory requirement. The potential value of Airborne Image Recorders (AIR) in the investigation of small aircraft accidents and the existence of a corresponding Technical Standards Document - EUROCAE ED-112A and the different arrangements in Europe for the safety regulation of HEMS and police helicopter operations using similar aircraft were highlighted.

The Investigation formally identified the following Causal Factors:

1) 73 kg of usable fuel in the main tank became unusable as a result of the fuel transfer pumps being switched off for unknown reasons.
2) It was calculated that the helicopter did not land within the 10-minute period specified in the Pilot’s Checklist Emergency and Malfunction Procedures, following continuous activation of the low fuel warnings, for unknown reasons.
3) Both engines flamed out sequentially while the helicopter was airborne, as a result of fuel starvation, due to depletion of the supply tank contents.
4) A successful autorotation and landing was not achieved, for unknown reasons.

In addition, the Investigation also identified the following three Contributory Factors:

1) Incorrect management of the fuel system allowed useable fuel to remain in the main tank while the contents in the supply tank become depleted.
2) The RADALT and steerable landing light were unpowered after the second engine flamed out, leading to a loss of height information and reduced visual cues.
3) Both engines flamed out when the helicopter was flying over a built-up area.

Seven Safety Recommendations, all but two relating to the need for flight recorders to be required on small helicopters such as the one involved in the accident, were made at the conclusion of the Investigation as follows:

  • that, when the European Aviation Safety Agency requires a radio altimeter to be fitted to a helicopter operating under an Air Operator’s Certificate, it also stipulates that the equipment is capable of being powered in all phases of flight, including emergency situations, without intervention by the crew.
[2015-030]
  • that, when the (UK) Civil Aviation Authority require a radio altimeter to be fitted to a helicopter operating under a Police Air Operator’s Certificate, it also stipulates that the equipment is capable of being powered in all phases of flight, including emergency situations, without intervention by the crew.
[2015-031]
  • that the (UK) Civil Aviation Authority requires all helicopters operating under a Police Air Operators Certificate, and first issued with an individual Certificate of Airworthiness before 1 January 2018, to be equipped with a recording capability that captures data, audio and images in crash‑survivable memory. They should, as far as reasonably practicable, record at least the parameters specified in The (UK) Air Navigation Order, Schedule 4, Scale SS(1) or SS(3) as appropriate. They should be capable of recording at least the last two hours of (a) communications by the crew, including Police Observers carried in support of the helicopter’s operation, and (b) images of the cockpit environment. The image recordings should have sufficient coverage, quality and frame rate characteristics to include actions by the crew, control selections and instrument displays that are not captured by the data recorder. The audio and image recorders should be capable of operating for at least 10 minutes after the loss of the normal electrical supply.
[2015-032]
  • that the (UK) Civil Aviation Authority requires all helicopters operating under a Police Air Operators Certificate, and first issued with an individual Certificate of Airworthiness on or after 1 January 2018, to be fitted with flight recorders that record data, audio and images in crash-survivable memory. These should record at least the parameters specified in The (UK) Air Navigation Order, Schedule 4, Scale SS(1) or SS(3), as appropriate. They should be capable of recording at least the last two hours of (a) communications by the crew, including Police Observers carried in support of the helicopter’s operation, and (b) cockpit image recordings. The image recordings should have sufficient coverage, quality and frame rate characteristics to include control selections and instrument displays that are not captured by the other data recorders. The audio and image recorders should be capable of operating for at least 10 minutes after the loss of the normal electrical supply.
[2015-033]
  • that the (UK) Civil Aviation Authority considers applying the requirements of AAIB Safety Recommendation 2015-032 and AAIB Safety Recommendation 2015-033 to State aircraft not already covered by these Safety Recommendations.
[2015-034]
  • that the European Aviation Safety Agency mandate the ICAO Annex 6 flight recorder requirements for all helicopter emergency medical service operations, regardless of aircraft weight. The last two hours of flight crew communications and cockpit area audio should be recorded. The cockpit area audio recording should continue for 10 minutes after the loss of normal electrical power.
[2015-035]
  • that the European Aviation Safety Agency mandate image flight recorder requirements for all helicopter emergency medical service operations, regardless of aircraft weight. The image recordings should have sufficient coverage, quality and frame rate characteristics to include actions by the crew, control selections and instrument displays that are not captured by a data recorder. The recording should be of the last two hours of operation, including at least 10 minutes after the loss of normal electrical power to the flight recorder.
[2015-036]

The Final Report of the Investigation was published on 23 October 2015.

Further Reading/Related Articles