Glasgow helicopter crash mystery: 76kg of fuel in tank

23 Oct 15 6 Comments

Today the British Air Accidents Investigation Branch (AAIB) released their report on the Eurocopter EC135 which crashed into Glasgow city centre two years ago.

On the 29th of November 2013, the helicopter crashed through the roof of the Clutha Vaults Bar while there were over a hundred people on the premises. The crash killed the pilot and two police officers on board as well as seven people in the bar. Another eleven people in the bar were seriously injured. The pub reopened in July of this year.

The helicopter was on Police Scotland operations. It did not have a flight recorder on board; however a fair amount of data was recovered from the non-volatile memory of the systems on board. This, along with radar, ATC and CCTV recordings allowed investigators to collect information about the flight.

The day-shift pilot had been flying all day, refuelling the helicopter three times, twice at Inverness and again when he returned the helicopter to Glasgow. He handed over the helicopter to the night-shift pilot with no abnormal indications and 400 kilogrammes of fuel on board.

The night-shift pilot had been trained to fly helicopters in the Royal Air Force and had 5,592 hours of flying experience with 646 on type. The helicopter was a Eurocopter EC125 T2+, a twin-engine light utility helicopter.

He reported for his shift at 19:30 that evening.

He departed Glasgow City Heliport at 20:44 with two police observers, who were carrying out surveillance tasks with night vision goggles. They initially flew towards Oatlands to assist in the search for a person believed to have been struck by a train. At 21:21, they requested clearance to route towards Dalkeith, about 42 nautical miles east of the heliport. At this stage, they had approximately 273 kilogrammes of fuel left.

They arrived at Dalkeith at 21:41 with an estimated 203 kilogrammes of fuel. They completed their task there at 21:45 and flew back towards Glasgow with about 192 kilogrammes of fuel.

At 21:59, the pilot contacted Glasgow ATC to inform them of his position and that he was now heading towards Bothwell, South Lanarkshire.

It was as they were approaching Bothwell that the first Low Fuel 1 warning occurred.

The warning unit is in the middle of the instrument panel and provides visual and audio indications when a warning is triggered. The warnings illuminate the warning in red and a repetitive gong sound is triggered, audible to the pilot and the observers through their headsets. The gong sound can be silenced using the audio reset button. When a new caution illuminates, any pre-existing warnings disappear temporarily and the new caution is highlighted by horizontal yellow flashing bars at the top and bottom of the screen.

The gong was cancelled (acknowledged) by the pilot pressing the reset switch. This stops the aural indication and clears the yellow flashing bars and the master caution.

The correct response to a low fuel warning is to follow the Pilot’s Checklist Emergency and Malfunction Procedures, which call for all transfer pumps to be selected on and check that the fuel transfer pump circuit breakers are in. If the warning(s) remain, the pilot should switch the air conditioning off and switch the bleed air off. If the warning remains on, the helicopter is to land within ten minutes.

The caution went off a second time and was again acknowledged by the pilot. The Low Fuel 2 warning illuminated and the gong sounded and was again acknowledged by the pilot. The Low Fuel 1 warning re-illuminated for a third time and was acknowledged. After this, both low fuel warnings remained illuminated for the rest of the flight.

The pilot made the expected radio calls and there was no indication of concern.

It’s worth taking a moment to look at the EC135 fuel system. The fuel is pumped from the main fuel tank by two transfer pumps into the supply tank which is divided into two cells. Each cell feeds an engine.

When the main and supply tanks are full of fuel, above the supply tank divider and overflow channel levels, fuel can flow freely between the main and supply tanks. However, to ensure a constant transfer of fuel into the supply tanks, there are two electrical transfer pumps fitted towards the forward and aft ends of the main tank. In normal operation, one or both of these transfer pumps should be running constantly, to deliver fuel via non-return valves into a common transfer manifold.

At 22:06 the helicopter arrived at Bothwell with an estimated 122 kilogrammes of fuel remaining. It carried out three surveillance tasks in the Bothwell area, without any radio transmissions made. At 22:14, the helicopter would have had approximately 100 kilogrammes of fuel left.

At 22:19, the pilot contacted ATC to say they had completed their tasks in the Bothwell area and were returning to Glasgow City Heliport. Glasgow ATC confirmed that the helicopter was clear to enter the Glasgow Control Zone and the clearance was acknowledged by the pilot. The police observers had continued conversations using the Airwave system. There was no indication in any of these conversations that they had noticed any fault with the helicopter.

I should point out a side issue. The EC135 helicopters had been reported by the operator as occasionally showing wrong fuel quantity indications and, in 2013, the accident helicopter itself had defect reports and maintenance to deal with this. Generally, the issue was solved by replacing the fuel sensor unit. After the accident, it was discovered that the problem may have been caused by water contamination brought about the cold washing operations and the manufacturer has now specified daily hot washing process instead of the cold chemical engine compressor-wash operations. There’s no evidence that the fuel system in the helicopter was indicating incorrectly that night; however it’s clear that erroneous indications need to be dealt with.

No further radio transmissions were received from the pilot. At this time, they had approximately 86 kilogrammes of fuel.

From the operator’s Operations Manual:

Company policy is that the aircraft should not land with less than 60kg of fuel in the tanks.

If it appears to the aircraft Commander that the Final Reserve Fuel may be required, a PAN call should be made. If the Final Reserve fuel is then subsequently reached, this should be upgraded to a MAYDAY.

The final Reserve Fuel for an IFR flight at the time was 85 kilogrammes (now 90). Even this close to Glasgow City Heliport, it seems clear they are going to use that fuel. At the very least, a PAN call would have been in order. With no flight data recorder and no cockpit voice recorder, it is impossible to know why this did not happen.

The helicopter flew at a speed of 100 knots and an altitude of about 1,000 feet above mean sea level. At 22:21, at about 2.7 nautical miles east of Glasgow City Heliport, the right engine flamed out.

The engine flameout should have been accompanied by eng fail, eng oil p, fuel press and gen discon caution indications and the instruments should have indicated a power loss. In this situation, the Emergency and Malfunction Procedures instruct the pilot to establish a OEI flight condition, identify the affected engine, carry out a single-engine emergency shutdown, if the situation allows for OEI flight, and land as soon as practicable.

An OEI flight condition is when one engine is inoperable. This procedure includes ensuring that the correct engine has been identified and switching it off. This does not appear to have happened.

Half a minute later, 1.8 nautical miles east of Glasgow City Heliport, the left engine flamed out. There is only one other recorded instance of an EC135 suffering a total loss of engine power.

The helicopter was at an altitude of somewhere between 700 and 500 feet.

The ROTOR RPM warning illuminated and sounded in the cockpit – this warning is that the speed of the rotor had decreased below 97%. The warning was extinguished. It re-illuminated and was extinguished again, which means that it was likely that the pilot was actively trying to maintain the rotor rpm. It re-illuminated a third time and stayed on for the remainder of the flight as the helicopter descended rapidly.

His only option was to attempt a forced landing at night in a built-up area of Glasgow.

In this situation, the Emergency and Malfunction Procedures instruct the pilot to enter autorotation (again, a memory item), using the collective to maintain the Nr within limits. The procedures recommend an airspeed of 75 kt, noting that 90 kt should be used to achieve the maximum range and 60 kt for the minimum rate of descent. Then, as memory items, the pilot should switch both eng main switches and both fadec switches off.

The switch selections were not made and no MAYDAY transmission was received.

In training, most autorotations are started between 1,000 and 2,000 feet above the ground. The helicopter was between 500 and 750 feet with very little time to establish a stable autorotation followed by a flare recovery and landing.

Between 650 and 500 feet, a battery discharge warning triggered, indicating that the battery was the only source of electrical power. This means that the pilot no longer had accurate height information and he did not have a landing light to help him see his landing point.

The last recorded radar position of 390 feet was taken at 22:22:19. A number of people saw the helicopter as it descended into the city. There were bangs “like a misfiring car” and then silence.

The helicopter crashed into the roof of the Clutha Vaults Bar. The main rotor blades and the tail rotor were not rotating, which is evidence that collective pitch had been applied during the final descent. The helicopter was falling down, not moving forward: evidence that a flare manoeuvre had been initiated. However, it was all too late. The pilot had only eight seconds between the second engine flame out and the crash.

The roof collapsed downwards, as the helicopter penetrated the structure, and numerous beam and roof boards were disrupted. This created a pile of debris on which the helicopter came to rest. The RSJ, which was slightly offset from the point of impact, was displaced from its supporting pillars. The disruption and roof collapse created large amounts of brick, mortar and plaster debris which covered everything in a very thick layer of dust.

When the helicopter crashed, the main fuel tank held 76 kilogrammes (95 litres) of fuel. The supply tank held 0.4 kilogrammes (half a litre) in the number 1 drain. No fuel was recovered from the number two supply tank drain. The airframe fuel shut-off valves were set to the open position for normal flight operation.

No fault could be found with the helicopter fuel system contents and there was no blockage, no sign of contaminates and, even after the accident, the two fuel transfer pumps and the supply tank prime pumps ran correctly when tested.

Both engines quite simply ran out of fuel.

It was estimated that, that, while the helicopter was in the latter stage of its cruise at 3,000 ft amsl, in the descent to 2,000 ft amsl or in the early stage of the cruise at 2,000 ft, the transfer of fuel from the main tank to the supply tank ceased, leaving only the fuel in the supply tank available to the engines. It was concluded that this was due to the fuel transfer pumps in the main fuel tank being switched off, either simultaneously or in sequence.

At some point while the helicopter was returning to Glasgow from Dalkeith, the fuel pumps were set to OFF, which meant that the only fuel available to the engines was the fuel already in the supply tanks.

The primary reason for turning off the pumps would be if there was a risk of the pumps running dry. While they were orbiting for their various tasks, it’s possible that the fuel transfer pump in the main tank could have run dry. If one of the caution messages had illuminated, and it’s impossible to know for sure, then switching off the pumps would be the correct response. However, it would have had to run dry for three minutes to illuminate a caution message, which is believed unlikely at Dalkeith based on the the flight pattern and the level of fuel in the tank. The caution message is more likely to have illuminated as they approached Bothwell, but that’s two minutes later than when the pumps were believed to have been switched off.

The EC135 fuel system does not have any indication for the crew to advise when an exposed pump becomes re-submerged in fuel.

There’s no clear explanation as to why both fuel transfer pumps were switched off but more important is the question of why they were not turned back on when the low fuel warning was triggered.

However, it was established that the low fuel 1 and 2 warnings were activated and acknowledged. Whether and how the pilot followed the relevant procedure in the Pilot’s Checklist Emergency and Malfunction Procedures could not be established. With a positive indication of fuel in the main tank, the procedure prompts the pilot to check that the fuel transfer pump switches are on, that the fuel transfer pump circuit breakers are in and, if the warning light remains on, to switch off the air conditioning (if installed) and switch off the bleed air if the outside air temperature is less than 5°C. It then states: ‘LAND WITHIN 10 MINUTES’.

The real mystery, then, is the actions after the initial low fuel warnings. The pumps were not turned back on and the pilot and police observers continued with their tasks. We have no idea what conversation took place in the helicopter, but it’s hard to understand why, under the circumstances, they carried on with another surveillance task on their route back to Glasgow City Heliport, after the low fuel warnings had been active for at least eight minutes.

One oddity is that in the wreckage, both fuel prime pump switches were found in the on position. They are not required to be on for any procedure other than for starting the engines before flight. These switches are not guarded or gated, so they might have been moved at impact or during the victim recovery operation. It’s not clear that they had been switched on prior to the crash. Selecting them to on would not have had an effect; however it’s possible that the wrong switches were switched on. The AAIB note this possibility without further analysis.

Not surprisingly, a key safety recommendation is that all helicopters operating under a Police Air Operator’s Certificate be equipped with flight data and cockpit voice recorders.

The causal factors are straight forward and depressing.

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.

The contributory factors are no more helpful in understanding this tragedy, although they at least allow us for some room for improvement.

1. Incorrect management of the fuel system allows 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.

Scottish First Minister Nicola Sturgeon has commented on the report:
ScottishGovernment – News – AAIB report on Clutha tragedy

Our thoughts continue to be with the families and friends of those who lost their lives in the Clutha tragedy. I had previously called for this report to be made public as soon as possible and I therefore welcome the fact that has now happened. However, it is deeply disappointing that after two years of investigation the report does not reach a clearer conclusion – indeed in some respects, it seems to raise more questions than it answers. I therefore share the disappointment of the families that it does not provide the closure they sought and hope that the Fatal Accident Inquiry that the Crown Office has now said will happen as soon as possible, can help the families get the answers they seek.

Except that sadly, there are no better answers to be had. This frustration even comes through in the investigators’ comments in a way that we don’t often see in AAIB reports:

Whilst the selection of both fuel transfer pumps to off has occurred before in other types of helicopter, this was the first fuel starvation accident involving an EC135 in more than three million flight hours of operation, over a period of about 20 years.

In summary, 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.

Yes, it’s deeply disappointing not to have the answers. Unfortunately, we’re unlikely ever to know.

Category: Accident Reports,

6 Comments

  • Human beings are 1 billion years old. Aircraft are advancing faster than we are. From reading the AAIB’S report, I can see that the crew were aware that they could be low on fuel for about 7-8 minutes yet continued on surveillance. Instead of running through the emergency checks in the pilot’s notes and protecting the safety of all, the crew appeared to have put loyalty to their colleagues first.

    The procedure for “low fuel warning”is:

    1.check both fuel pump transfer switches are on
    2.check fuel pump transfer circuit breakers are in.
    3.If warning light is still on, switch off air con if fitted.
    4.Switch off the bleed air if outside temp is less than 5 degrees centigrade
    5. Land within 10 minutes.

    • “At separate points in the helicopter’s final flight, circumstances existed that caused Captain Traill to switch off both fuel transfer pumps; and where each pump was properly switched off by Captain Traill. When switching off the second (i.e. the aft) fuel transfer pump, Captain Traill appears to have overlooked the fact that he had previously switched off the forward fuel transfer pump approximately 11 minutes earlier [224].”

      “By January 2013 the potential risk of the CAD displaying a higher fuel quantity compared to the actual quantity of fuel on-board had been identified; and pilots, including Captain Traill, had been made aware of that potential risk [280].

      The sheriff principal has determined that the following fact is also relevant to the circumstances of the deaths, namely, that it is more likely than not that the quantities of fuel displayed on the fuel quantity indication system of the helicopter contradicted the LOW FUEL warnings (see Finding 7.1). There are, however, two points that must be stressed in relation to this finding.

      Firstly, the contradictory fuel display is only of relevance until the illumination of the LOW FUEL warnings. At that point the actions set out in the Pilot’s Checklist should have been performed by Captain Traill. By not carrying out the actions set out in the Pilot’s Checklist, Captain Traill consciously took a risk in proceeding on the basis that the LOW FUEL warnings were in some way erroneous (when they were not). That decision had fatal consequences. There was no logical basis for preferring the (possibly erroneous) figures displayed on the CAD to the accurate LOW FUEL warnings, particularly in circumstances where the fuel transfer pumps were both switched off [312].”

  • It is clear that the a/c could very likely have safely landed at its base if its fuel had been correctly managed.
    However, it is also clear that it could not have so landed while observing the required MLA (because the pilot/crew had overstayed their mission(s)).
    A question so far unaddressed is what protocol the pilot is required to observe after landing below the MLA. Is he required to make a report? And are the service staff required to (separately) make a report if they find that, in refuelling, the amount that they deliver implies that a below-MLA landing must have been made?
    Would the pilot be subject to some reprimand (or even sanction) if he was obliged to make a report concerning a below-MLA landing (for a reason that he could not justify)?
    Imagining that the pilot might have believed that he would suffer some sanction (maybe severe) if he made a below-MLA landing then it offers a potential reason why he might try to preserve fuel in the main tank (so that he was able to avoid making a below-MLA report and the Refuelling staff would not detect a below-MLA landing). Perhaps it also offers a reason why (having tried to preserve fuel in the main tank) he did not make the PAN call (because that would reveal what he had done).
    Of course, it does not explain why he did not take the obvious remedial measures when matters got yet more desperate but, by then, it was all happening quite quickly and perhaps logical thought was in short supply.

    A further thought occurs:-
    let us imagine that the a/c had succeeded in making it to its base on the fuel in the supply tanks, but that these tanks were near empty; no below-MLA reports are made.
    The a/c is subsequently then fully refuelled into its main tank but, since its equipment is switched off, none of that fuel is transferred to the supply tanks (except by overflow? – does that obviate the issue?).
    Only when the a/c is started prior to its next flight is fuel pumped to the supply tanks.
    The result of this might be that the a/c goes out on its next mission with upto 60kg less fuel than it is presumed to contain. Thus, its mission-time capability might be upto 18 minutes less than expected.

    • Your point illustrates why any safety management system (SMS) that incorporates flight data monitoring (FDM) or Flight Operations Quality Assurance (FOQA) must be set up such that pilots need not fear reprisals. Safety problems can only be remedied if they’re visible.

Leave a Reply to Mendel

Your email address will not be published. Required fields are marked *

*
*
*

This site uses Akismet to reduce spam. Learn how your comment data is processed.