Taiwan's accident investigators have taken the unusual step of publishing part of the flight data recorder printout for the crashed ATR 72-600 almost as soon as it was available to them. There are no rules or protocols saying they must do so, and none saying they should not.
The printout they released concerns only the data for the engines. It is a series of graphical lines describing the state of 12 different engine parameters against a timeline, with barometric altitude also displayed. The graphs provide numerical values for some of those parameters; others just show whether a switch is on or off – like the fuel shut-off valve for example.
This data tells the investigators, in great detail, what happened, but still not – at this stage – why it happened.
The graph shows that the crew certainly suffered the engine "flame-out" they reported in a Mayday call to ATC: the turbine temperature for No 2 engine (the right-hand one) dropped, power was lost and the propeller auto-feathered.
But then, in the stressful situation prevailing from that point, the crew carried out the shutdown drill for the working engine, so it stopped too.
Why would the investigator release this information so soon without knowing the cause?
The investigators knew the information about this critical mistake would soon have to be released, and it looks as if they believed it would be better to publish the cold data that shows what occurred, rather than to make a statement – without releasing the data – that could be interpreted as a premature judgement about the human factors of this case.
Perhaps the most famous previous case in which a disaster occurred because an engine failed and then the crew mistakenly shut down the good engine (rather than the damaged one) was the British Midland Boeing 737-400 crash at Kegworth, UK, in 1989. In that case 47 of the 126 people on board died.
In the TransAsia case the total airborne time for flight GE235 was 2min 40s.
All was going well for 45s after take-off, but as the aircraft was climbing through about 1,200ft (pressure altitude) the turbine temperature for the No 2 engine dropped and the engine auto-feathered. It is not clear why. The aircraft continued to climb on the power from the remaining engine, reaching a maximum height of about 1,650ft.
But during that short period the crew allowed the power lever (throttle) of the failed No 2 engine to stay where it was, and started slowly pulling back the power lever of engine No 1 (the working left engine). When it had been reduced almost to idle setting, the fuel was shut off and the left engine also feathered. Just before they shut off fuel to No 1, they advanced the throttle of the failed No 2 engine as far as they could, as if it would provide them with additional power.
At that point there would have been a total absence of engine and propeller noise, but lots of alarms going off as systems lost their electrical power. From that time onward the crew had, as it turns out, 1min 15s of gliding time before hitting the surface. That is not really long enough to go through a successful engine re-start drill, but they did begin an attempt to re-light No 1 about 15s before impact.