AAIB

The Ahmedabad crash: accidental or deliberate?

David Learmount Uncategorized , , , , , , , , , , , , , , , , , ,

The movement of two small switches on the aft end of the flight deck centre console, reachable easily by both pilots, appears to hold the key to what happened to the Air India Boeing 787-8 that crashed fatally just after take-off at Ahmedabad on 12 June.

It seems that one of the pilots selected these switches from “Run” to “Cut Off”, stopping the engines at a critical point just after the aircraft became airborne. The purpose of this article is to examine the arguments for and against deliberate action (compared with unintentional error) on the part of one of the Air India pilots.

On 11 July the Indian Air Accident Investigation Bureau published its preliminary factual report on Air India Flight AI171, a Boeing 787-8 registration VT-ANB. This has revealed the movement of the fuel control switches (FCS) mentioned above, and the resulting consequences of that movement. The data confirming this was derived from the two Enhanced Airborne Flight Recorders (EAFR) in the accident aircraft.

There are two FCSs, one for each engine in a 787. They have two settings: Cut Off and Run. The first act by any crew in starting a 787’s engines on the ground is to set the switches to Run. On the ground or in the air, setting the switches to Cut Off stops the fuel flow to the engines. (See photograph below, showing the switches just behind and below the engine power levers)

According to the AAIB report, just after take-off at Ahmedabad, these switches were moved from Run (up position) to Cut Off (down position). The left switch was moved first then, one second later, the right switch. This action cut off the fuel flow to both engines. There is no automatic function that could move these switches, so they must have been moved manually, or by something physically impacting them. Each switch has a locking mechanism so it cannot be moved accidentally, and there are guard brackets either side of the pair to deflect inadvertent contact by objects. To select the switches from one setting to the other, they must first be pulled out against a spring force to release a locking mechanism, then moved up or down.

Timeline (UTC):

08:07:37 VT-ANB begins take-off roll. 08:08:39 Lift-off at 155kt. 08:08:42 Max airspeed achieved 180kt, also No. 1 FCS switch was moved from Run to Cut Off, followed by the FCS for engine No. 2. 08:08:47 the Ram Air Turbine began supplying hydraulic power. 08:08:52 No 1 engine FCS moved from Cut Off to Run. 08:08:56 No 2 engine FCS moved from Cut Off to Run. 08:09:05 Mayday call transmitted. 08:09:11 EAFR recording stopped.

The report says: “In the cockpit voice recording, one of the pilots is heard asking the other why did he cutoff [sic]. The other pilot responded that he did not do so.” Each pilot is recorded on a separate channel, so the AAIB must know which pilot made each statement, but has decided not to release the information at this preliminary stage, The report confirms that the copilot was the pilot flying, the captain the pilot monitoring. So it seems that one of them, apparently, moved both FCS from Run to Cut Off (see Timeline above), and the other noticed him doing it. Then, about 10 seconds later, one of the pilots attempted to restart the engines by restoring both FCSs to Run.

The report explains the effect of restoring the FCSs to Run, first in 787s generally, then specifically what happened in this case: “When fuel control switches are moved from CUTOFF to RUN while the aircraft is inflight, each engines full authority dual engine control (FADEC) automatically manages a relight and thrust recovery sequence of ignition and fuel introduction. The EGT [exhaust gas temperature in VT-ANB] was observed to be rising for both engines indicating relight. Engine 1’s core deceleration stopped, reversed and started to progress to recovery. Engine 2 was able to relight but could not arrest core speed deceleration and re-introduced fuel repeatedly to increase core speed acceleration and recovery. The EAFR recording stopped at 08:09:11 UTC.”

Take-off and early climb is a period of intense concentration by both pilots, the joint task being to ensure the aircraft maintains a steady climb while allowing the airspeed to increase gradually in a controlled way.

Under normal circumstances, after unstick there is only one actionable task for the pilots to carry out quickly: to check that a positive rate of climb is confirmed by the flight instruments, then select the undercarriage up. This task is normally carried out by the pilot monitoring on orders from the pilot flying, and it would entail moving the undercarriage control lever – located on the forward instrument panel – manually upward. In this case, according to the report, no-one called for the gear to be retracted, and no-one selected it up.

Instead, at about the time the gear would normally have been retracted, the FCS were moved downward from Run to Cut Off, the left switch first, the right switch a second later.

It is difficult to imagine that a crew member would have made such a gross error as reaching down and slightly back to move two small switches downward, one after the other, as a substitute action for a well established routine which would have involved reaching forward to move a single lever upward. And there was no cueing request from the pilot flying to pull the gear up anyway.

Pilots have occasionally, however, carried out inadvertent gross errors that almost defy credibility. You can see here the description of how, in January 2023, a Yeti Airlines ATR72 scheduled passenger flight was inadvertently set up for disaster during a visual circling approach to land at Pokhara airport, Nepal. I wrote that linked piece based on the preliminary report, but when the final report was published by the Nepal authorities it gave the following verdict: “The most probable cause of the accident is determined to be the inadvertent movement of both condition levers to the feathered position in flight, which resulted in feathering of both propellers and subsequent loss of thrust, leading to an aerodynamic stall and collision with terrain.” The check pilot had been asked by the pilot flying to increase the flap setting from 15deg to 30deg, but instead of moving the flap lever, he moved the pair of engine condition levers (picture supplied in linked article) to the position that demands the propellers to feather and stop turning.

If one of the pilots of AI171 did know what he was doing when he moved the FCS, he must have known that his action would have more or less guaranteed the result the world has witnessed, because there was insufficient time to restore usable power once it had been cut. VT-ANB was airborne only 3 seconds before the first FCS was switched to Cut Off, followed a second later by the second FCS, then ten seconds after that the FCS were both switched back to Run. The total airborne time was 42 seconds before colliding with the buildings that began break-up of the aircraft.

As for the likelihood that professional pilots would want to cause the destruction of the aeroplane they are flying, history provides evidence that it happens from time to time.

This was the summary of the situation as presented in the FlightGlobal annual safety review for calendar year 2023, which points out that deliberate acts by pilots to bring down airliners have been carried out by aircrew from all regions and cultures: “Pilot suicide on commercial flights in the last three decades has not involved only Europeans and North Americans. A Japanese, a Moroccan, an Egyptian, a Mozambican, a Botswanan, and a Singaporean, among others, have all been involved. The Flight Safety Foundation’s Aviation Safety Network accident database shows that, in its records beginning the 1950s, there has been a marked acceleration in the numbers of flights brought down by pilot suicide since the beginning of the 1990s, and this acceleration has continued in the new century. It is undoubtedly a modern flight safety hazard.” Since that time, although China has not confirmed it, the rest of the world has reason to believe that the March 2022 loss of a China Eastern Airlines Boeing 737 was not an accident.

It is inevitable that deliberate action by flight crew should be considered when a disaster like AI171 occurs. The India Air Accident Investigation Bureau will undoubtely investigate this possibility. But just one part of the trajedy is that, when all the flight crew die, their intentions will never be known for certain.

How did that happen?

David Learmount Uncategorized , , , , , , , , , ,
How did that happen? A Boeing Stearman lands on a Cessna 182.

A Cessna 182 was nearing the end of its landing roll on runway 22 at Dunkeswell aerodrome, Devon, UK, when a Boeing Stearman landed on top of it.

Remarkably, both pilots survived, but the Cessna pilot was badly wounded when propeller blades cut through the top of his cockpit.

The Air Accident Investigation Branch report describes what the Cessna pilot experienced: “With around 100 m to go to the [runway] intersection [where he intended to turn off to the right], the pilot reported that he heard and felt what seemed like an ‘explosion’ and then became aware of propeller blades rotating in front of his face. He recalled that the cockpit was filled with debris from the shattered windshield, shards of metal and splintered wood.”

It was a cloudy day but the visibility was pretty good below the cloud-base, which was 1,200ft above ground level, with a few patches at 600ft to 800ft. The AAIB says Dunkeswell witnesses described the weather as “workable”.

The pilots of both aircraft say they made calls on the Dunkeswell Radio frequency, but only the Cessna pilot received a reply.

Both pilots were alone in their aircraft, and both of them were experienced aviators.

Neither was aware of the other aircraft’s presence in or near the Dunkeswell circuit, and at the point of collision neither had seen the other.

If I were to stop the story here, even experienced general aviation pilots would wonder how this situation could possibly have developed. The answer, as usual, is that lots of factors combined; but even after reading the AAIB’s report, some questions remain unanswered.

The assembled contributory factors – in no particular order – are these: Dunkeswell Radio maintained an imperfect listening watch for its air-ground communication service (not ATC as such), the duty officer having multiple other duties while on watch; the Cessna pilot (G-OMAG) intercepted final approach from the “dead side” without clearing the non-standard arrival; the Stearman pilot (N68427), based at Dunkeswell, states he made several calls to Dunkeswell Radio when returning to the circuit, but received no response; neither pilot heard any calls from the other, both believing they were alone in the circuit; forward visibility from the Stearman’s cockpit is limited, especially directly ahead and below the nose. Finally, the AAIB found that the airport operator was not aware that the main AIP (aviation information publication) entry for Dunkeswell did not contain reference to the requirement for inbound traffic to join the circuit via either the downwind or base legs of the active runway. Accordingly, commercial AIP providers did not do so either.

The sequence on the day (20 August 2021) goes something like this: the Cessna, having obtained prior permission to arrive at Dunkeswell, took off from Bodmin – about half an hour to the west – at 13:00. The Stearman had taken off from Dunkeswell at 13:05 to carry out some flying in the local area to the north, and returned later intending a couple of circuits.

At 13:26 the Cessna passed Exeter to its north, and at the same time the Stearman joined Dunkswell’s left hand circuit for runway 22 from the east, and carried out a touch-and-go at 13:29. The Stearman pilot said he made a radio call indicating his intentions but could not recall getting an acknowledgement. Following his touch-and-go, N68427 flew a left hand circuit (see illustration below), intending to land from a curved base leg into a short final approach, because the wing-down attitude provides better sight of the runway for a Stearman pilot.

At about 13:27 the Cessna pilot called Exeter and said he was continuing with Dunkeswell Radio, then called the latter and was informed runway 22 was active with a QFE of 986hPa.

The AAIB report says that when the Stearman entered the 22 left hand circuit from the touch-and-go, the Cessna was about a mile to the west of the airfield, the dead-side of the circuit. The Cessna pilot decided to turn north (see illustration above), then carry out a right hand loop onto the base leg, continuing the turn onto final approach from the dead side. At that point, says the Cessna pilot, he called “Golf alpha golf final 22”. There was no acknowledgement, and the pilot told the AAIB he thought his aircraft was alone in the circuit.

CCTV images of the two aircraft show the two aircraft on final approach, the Stearman slightly higher, initially curving in from the left of the runway extended centreline, the Cessna on the centreline. As they converged, the Stearman remained higher than the Cessna and slightly ahead of it, the latter overhauling the biplane because of its faster approach airspeed. When the Cessna touched down just beyond the displaced runway threshold, the Stearman was above and close behind it, with the Cessna in the pilot’s blind spot below the biplane’s nose.

As the Cessna slowed toward its intended turn-off, the Stearman “touched down” on top of it. The AAIB report provides the Stearman pilot’s perception of what was happening: “The pilot reported that the landing ‘didn’t feel right’ and that the aircraft was not responding to control inputs. He applied power to correct what he felt was a drift to the left, then reduced power to idle. The aircraft continued to swing further to the left, off the runway and onto the grass.” (see headline picture).

The Cessna pilot could not open his cockpit door, but the Stearman pilot managed to do it and helped him out. The airport fire and rescue service attended rapidly and doused the Cessna with foam because fuel was leaking from the left wing tank.

Because the airfield only provided an air-ground communication service, not ATC as such, communications are not logged, but the pilots and some airport witnesses report calls being made in this case. The mystery is: why didn’t either pilot hear the other’s calls?

The UK Civil Aviation Authority’s Skyway Code provides plenty of apposite advice for this serious incident. One piece is this: “If you believe the circuit is clear but are not sure, there is no harm in asking over the radio whether there is any other traffic – it is not unknown for pilots to stop making position calls if they believe they are alone in the circuit.” Another is to stick rigidly to circuit procedure, which would include not joining from the dead side.

Shoreham and the future of air shows

David Learmount Uncategorized , , , , , , , , , , ,

No person or organisation associated with the 2015 Shoreham flying display accident escaped criticism – implied or actual – in the Air Accident Investigation Branch’s final report.

When the aircraft that later crashed, the Hawker Hunter T7, took off from its North Weald, Essex base heading for Shoreham to fly the display, it had several time-expired or unserviceable components in it. In retrospect the AAIB report says it was not in compliance with its permit to fly, yet none of these faulty components caused the accident.

The Flying Display Director hired to manage show safety was fully qualified in terms of knowledge, experience and expertise to oversee all aspects of the flying display, but the report implies there were some things – like the exact display routine the Hunter was to fly – he should have risk-assessed manoeuvre by manoeuvre. Yet even if he had, the crash might still have happened.

The Hunter’s pilot was fully qualified in terms of experience, training, flying recency and medical fitness to carry out the display he had planned, but on the day he got one of the aerobatic manoeuvres badly wrong, making mistakes that are difficult to understand in somebody so experienced.

The mistakes meant that he himself was seriously injured when thrown clear of the aircraft on impact with the A27 highway next to the airfield, but 11 people on the road were killed.

By flying a similar Hunter through the same manoeuvres, the AAIB has determined that the pilot could not have pulled the aircraft out of his “bent loop” without hitting the ground once he had passed the apex too low and failed, at that point, to carry out an “escape manoeuvre” by rolling the aircraft upright.

If he had used his ejection seat during the high speed descent from the loop it would not have saved him, and the pilotless aircraft would have continued to impact with the ground, possibly in much the same place. The only way the pilot could have prevented killing the people on the A27, says the report, was to crash the aircraft into nearby fields, but during the last second or so he probably still hoped he could avoid harm to road-users by pulling up in time.

Why was he too low at the loop’s apex?

He should have entered the loop from a 500ft base, but he started at about 185ft. The height of the top of a loop compared with the entry height is a product of speed and engine power at entry. The aircraft should have entered at a minimum 350kt with full power selected, but the Hunter entered at 310kt with less than full power until well into the pull-up. The pilot should have aimed for a 4,000ft apex with 150kt indicated airspeed over the top, but in fact it got to less than 3,000ft with 105kt.

Unless the pilot recognised the lack of energy at that point and carried out the rolling escape manoeuvre, he and the aircraft were doomed.

Why a pilot with so much experience of teaching, let alone flying, aerobatic manoeuvres failed to heed these indicators that the loop was going wrong may never be known, because trauma has obliterated the details of the fatal flight from the pilot’s memory, according to the report.

Air shows involve risk. A study by the AAIB has recently quantified that risk, and my blog a year ago describes the findings in detail.

The final Shoreham report confirms the impressions given by the earlier AAIB bulletins on the subject. Because no-one in an on-site air display audience in UK has been killed since the early 1950s, such success appears to have led to complacency.

Not rampant complacency, but a relaxed belief that all the people involved are experts who know what they are doing, so they don’t need to be given the third degree before a show.

The sign that not all was well was the number of serious air display accidents, mostly fatal, that occurred just outside the area controlled by the display organisers – just like the Shoreham Hunter crash.

The AAIB found that 65% of all air show accidents came into that category, but almost always the only person harmed was the pilot. So nobody, including the CAA, raised the alarm, until now.

Meanwhile aerodromes used for decades as air show venues have suffered encroachment at their boundaries by expanding residential and industrial development. This affects the profiles aircraft are allowed to fly during a display, and flight display directors are bound to take this into account.

No longer are display lines, and entry and exit profiles dependent purely on where the display audience “crowd line” is, they have to take into account what each aircraft would have to do in the event of a technical or operational mishap during the display to avoid crashing into a nearby populated zone.

These are considerations that will affect air shows in the future. If a flying display stops being exciting, it might as well give up. Or go somewhere else more rural.

Coastal air displays will survive, because the escape route for aircraft in trouble is obvious.

The best example of the conundrum air show organisers face is what has happened to the traditional Red Arrows display at the biennial Farnborough International Air Show. When the Reds reviewed their Farnborough routine in detail following the tightened guidelines published in the early Shoreham bulletins, they found they had to curtail their display considerably.

In a statement following the release of the Shoreham final report, the CAA says: “We are fully committed to ensuring that all air shows take place safely, for the six million people who attend them each year in the UK and for the communities in which they take place.”

Let’s hope the CAA means what it says.