Battling a Black Swan Flight … and Winning

Editor’s Note: Last week, Portuguese investigators GPIAAF (Gabinete de Prevençäo e Investigaçäo de Acidentes com Aeronaves e de Acidentes Ferroviärios) released the  final report of the harrowing Air Astana ERJ190LR flight KC1388 accident of 11 November 2018. The plane was on a post-maintenance flight, carrying three Air Astana pilots and three technicians from the operator.

Shortly after takeoff from Alverca do Ribatejo airbase, Portugal, the regional jet became uncontrollable. The pilots not only issued a distress signal, fighter aircraft were scrambled to escort them to the sea, where the Air Astana crew anticipated ditching their plane.

This exclusive narrative is from a detailed conversation CAT Europe Editor Chris Long had with the Air Astana flight crew, presenting their in-the-cockpit perspective.

The first that Captain Vyacheslav Aushev learnt of the task was whilst he was in another aircraft, having arrived in Almaty, Kazakhstan, where he was instructed to call crew control during checkout. He was then advised that he was planned to collect an Embraer 190 from the military airport at Alverca, near Lisbon, and return it via a refuelling stop at Minsk, Belarus, to the home base at Nur Sultan, Kazakhstan’s capital. This with no fare-paying passengers, but with three maintenance engineers returning to base. The aircraft was due out of a “C” maintenance check but, because of repeated delays totalling 11 days, the original aircraft captain was no longer able to take the flight, hence the short notice.

On arrival at the maintenance base, Captain Aushev met up with the rest of the ferry crew – First Officer Bauyrzhan Karasholakov and the Safety Pilot (also qualified on type) FO Sergey Sokolov, and had a comprehensive handover from the initial captain.

The aircraft was not yet ready, so Captain Aushev divided up the tasks of preparing for the flight. He was to be in charge of checking and completing the comprehensive post-“C” check paperwork, fuel planning and check for fuel discrepancies; FO Karasholakov to plan all the FMS inputs; and FO Sokolov to complete the performance calculations, door checks and aircraft equipment.

One of the technical problems was that the flight control software needed to be reloaded, and that took some time. Whilst waiting, they completed as much planning as could be done in advance. A full night’s sleep and breakfast was then necessary before starting again the following day. Next morning there was a further delay, so the last-minute flight planning was verified, and the planned crew duty time started on leaving the hotel. The aircraft having been signed off as “C” Check Complete, a normal pre-flight series of checks was initiated.

Weather checks revealed that there were potentially difficult meteo conditions in Minsk – snow, low cloud base and reduced visibility. The plan was to fly a single approach to Minsk and then, if necessary, divert to Sheremetyevo airfield, Moscow.

On start-up there were two minor snags. A warning notified that the level of hydraulic fuel was low, and although the pressurisation and conditioning was showing as normal, there was no flow through the normal ventilators. Back to the hangar.

The tech issues having been signed off, once again the usual pre-flight checks started; this time the only anomaly was during the walk-around when it was noticed that the access panel for the hydraulic system was open – quickly corrected. However, it was also apparent that a lot of rain had fallen – puddles forced the recalculation for a “contaminated runway” take off. Change of configuration (flaps 4 – close to a landing configuration), and full thrust. Finally, and by no means least – there were changes of runway between 22 and 04 – each change, of course, needing adjustment to the takeoff parameters.

None of this was desperate – we have all experienced pre-flight technical issues, pressure of weather, runway changes, and tight flight crew duty time limitations, but it always flags up a heightened situational awareness! The good news was that the team building between the three crew had already been created – a critical aspect in what was to come.

Accident flight path perspective. Image credit: flyback Embraer

The Trouble Started …

Although particular attention was paid to the possibility of a rejected take off, nothing unusual was noted during the take off and rotate. The trouble started about 10 seconds after becoming airborne, when small, unexpected oscillations in roll alerted the crew to a problem. The oscillations gradually built up, but the aircraft continued to climb (full thrust already selected). The crew experienced further undemanded roll and yaw inputs, resulting in big changes to aircraft attitude – the aircraft was effectively uncontrollable. With the higher angles of bank, the nose dropped, and pitching to recover the altitude only helped as the aircraft rolled through a nearly wings-level attitude. Captain Aushev said that it felt as if there was a “complete disconnect between the control yoke and the aircraft”.

Throughout the event there were no visible/audio warnings of system failure, but later on the safety warnings for stall, TCAS and GPWS all operated in the appropriate way. Return to the airfield was considered but, not having established control over the aircraft, the training at Air Astana kicked in – climb to get time and to make a plan. Even a less-than-perfect plan is better than none.

There were no indications as to the source of the problem, but the fact that the flight control software had been reloaded led to the initial conclusion that it was perhaps this which created the difficulty. So first, at 1000 ft after takeoff, the autopilot was selected to see whether it could manage the situation. It was not able to do so. Then the crew coordination kicked in even more deliberately as the flap was raised by increments – each confirmed by the Safety Pilot, FO Sokolov, before selection. The least destabilising configuration was found to be with the flaps set at 1. It was then decided to selectively and separately disengage the control channels – first roll, then yaw, then pitch. The sequence was for confirmation that the two pilots were ready – then the Safety Pilot deselected the notified channel – there was a pause to see whether it made any difference. Again, this did not solve the issue.

In the meantime, there were extreme excursions from level flight resulting in major changes to altitude and heading. The tendency was always to turn to the left. Given that the cloudbase on takeoff was 300 ft there was no realistic chance to descend to VMC under control, so the clawing for altitude continued, hoping to get VMC on top. It was also decided that, given that an uncontrolled crash into terrain was a distinct possibility, the marginally better bet was to go for a ditching – an aspect that had been recently covered in training, with all the hazards that that option presented.

ATC Support and F16 Escorts

The radio handover to Lisbon Approach immediately after takeoff (due to the proximity of commercial traffic from the busy Lisbon airport) resulted in a very effective ATC input. The crew had nothing but praise for the cool way that Lisbon ATC offered help only when it was requested. Having decided on ditching as the least dangerous way of getting down, repeated requests for a heading to the sea were required, the problem being that, with an aircraft which could not keep a steady heading “for more than one second” there was no way that such a course could be maintained.

VMC was finally discovered at about 18,000 ft, but a frightening plunge had the aircraft in a near vertical descent, the altimeter unwinding so fast that it was unreadable, airspeed in excess of 350 kts IAS, loud aerodynamic noise and, as they pulled out as firmly as they dared – over the strident calls from the GPWS, a creaking sound from the airframe. As they pulled out, still IMC, the Safety Pilot saw a glimpse of a hill in front of them but decided it would not help the team to call that as they were doing the maximum to recover anyway.

ATC then offered the help of a fighter escort from the Portuguese Air Force to help to navigate to the sea, and two F16s were scrambled to get to them. The join up was a little dramatic, in that the TCAS reacted vigorously to the presence of other aircraft, until the fighters modified their approach.

The crew continued to discuss solutions, however wild, to try to get some resolution. This was always a group discussion and collectively decided on. For instance, it was suggested that one of the control yokes be disconnected to isolate false inputs. That idea was discarded, as it need the combined work of both pilots to manage what little control there was. Now more than an hour into the flight, there was need for water for everybody – both to drink but also to face-splash to try to refresh. One of the travelling engineers was called forward for that, and another idea emerged. Exactly what work had been carried out during the “C” check? The engineers came forward with their list of tasks carried out, and as the crew moved down the list they came to the “Aileron Cables Changed”. Suddenly there was a light bulb moment.

If that was the core problem, then moving the aileron in the opposite direction should work. The engineers were dispatched to the cabin to report exact aileron deflection as a result of cockpit roll input. There was quick confirmation that this was the crux of the problem.

It is, of course, one thing to define the problem, quite another to resolve it. Years of learning to fly instinctively now needed to be unlearned. Think about the challenge of riding a bicycle with hands crossed over on the handlebars.

Once again the crew cooperation came to the fore. Each time a control input was being considered all three pilots cross-checked and confirmed before and during the control deflection. Difficult, but achievable with minor deviations from the desired flight path.

By now not only were the F16s helping with navigation but, with the aircraft under (partial) control at an optimum speed of 200 kts, the decision was to try for a suitable airfield with VMC and a long runway. The military air base of Beja was identified by ATC, and a diversion initiated, with the F16s leading and updated vectors from ATC. After completing all usual descent and landing checks, the choice was for a long straight-in approach to runway 19R. During the descent on finals the aircraft started rolling again as the natural reversion to traditional control inputs cut in. The decision was to go around (with no complicating change of configuration) rather than simply throw the aircraft at the airfield. At this point Captain Aushev realised that his personal tank was almost empty, and that FO Sokolov was fresher than the two pilots in the operating seats. In addition, he was the one who (with his extra capacity on the jump seat) had the clearest perception of what was needed to fly the aircraft.

Pictured from left to right: Capt. Vyacheslav Aushev (left), FO Bauyrzhan Karasholakov (centre), and FO Sergey Sokolov (right).

The Captain Makes a Call

The role of captaincy could be defined as using all available resources to best effect to ensure the safe conclusion of a flight. It is here that Captain Aushev demonstrated that principle to the ultimate. Recognising his own fatigue, he allocated the seat and landing to the Safety Pilot, FO Sokolov. Although this individual had been subject to the same emotional stress as the others in the crew, he had not had the same physical workload. He was entirely familiar with the situation and understood what was required to execute a successful landing. The aircraft was still difficult to control, and, in fact, a second go-around had to be initiated as another unstable approach was recognised. It was then decided to use rudder only for directional control from very short final to touchdown.

For the next attempt an F16 was used to lead to a 3-mile short final. At a low height, due to wind drift, the aircraft was better set up for the unofficial runway 19L (a former taxiway) which was immediately in front of the aircraft. Permission to land on that was requested at short notice, and immediately authorised by the military controller, and a successful landing was accomplished.

Experience from Training

What is clear from the debrief is that the crew relied heavily on training experience. The emphasis on detailed planning and SOPs helped in the generic, but the Air Astana philosophy throughout training is that the first resource is to buy time, make a plan, stay with the SOPs, and overlay all of that with some common sense. In this case, because there were no faults signalled by the aircraft systems, the SOPs did not solve the issue, but they did allow a systematic approach to the problem solving. Climbing to buy time to make a plan worked, but the overriding tool was the very close crew cooperation. Interestingly, their work in CRM sessions had emphasised the eye-to-eye contact, and they deliberately used that during the flight to reinforce confidence in the joint decisions.

Keeping calm, and mutual encouragement in that, together with a “Never Give Up” mantra resulted in a successful conclusion to a very tough situation.

There are many technical and engineering issues to be considered, and that is properly left to the formal enquiry. The purpose of this narrative is simply to record the impressions and recollections of the operating crew – and of a job well done!

A final thought for the pilots amongst you – how would you have coped with a real Black Swan event like this - what have you learnt from this?

READ MORE. Chris Long profiled Air Astana’s training programme in CAT Issue 4/2019. To read Air Astana – the Kazakh Solution click here.