A Troubling Report from Rostov

Reconstruction of the FlyDubai 737-800 wreckage

The Russian Interstate Aviation Committee has published an interim report on their investigation into the crash of FlyDubai FZ981. The report is quite troubling as it suggests that the aircraft did not suffer a low speed event or stall, but rather hit the ground with flying airspeed in an extreme nose down attitude. The report is here.

The report starts with the facts that are already known: the aircraft attempted one approach which was aborted followed by an extended time in a holding pattern. Nearly two hours later, a second approach was attempted and a second go around was initiated at 721 ft above ground level (AGL).

The report states that a possible reason the second go around was initiated was due to a sudden increase in indicated airspeed of 20 knots to 176 knots. This is entirely plausible as a gust of this velocity would cause the flaps to automatically retract to avoid an overspeed. Should this happen while on final approach under 1000 ft AGL, the correct decision would be to go around.

Windshear

There has been some speculation that windshear might have been the primary cause of the crash. I think it is important to differentiate between the windshear which might be generated due to convective activity (a thunderstorm) as opposed to gusts found in frontal activity. The weather was consistently poor for many hours preceding the crash with gusty winds but no reported thunderstorms.

The winds were reported as 20 degrees off of runway heading at 27 knots gusting to 42 knots. These winds would certainly make for a difficult approach and landing as the plane would be bucking like a bronco, but they would not be a challenge to staying airborne. Windshear found in thunderstorms can threaten an aircraft on approach but there is no indication such conditions existed here.

Anomalies on the Go Around

It was on the second go around that trouble started. The crew set the flaps to 15 and retracted the gear which is normal procedure. At a height of 1900 ft, the pilot flying pushed on the control column which decreased pitch and caused airspeed to increase. It is at this time that the flaps would normally be retracted so as to not overspeed them. The report states that the flaps automatically retracted to position 10 while the speed increased to over 200 knots.

The flaps blow-up protection on the 737-800 is designed to prevent damage to the flaps due to unintentional overspeeds. Called the Flap Load Relief system, it will retract the flaps from 15 to 10 when the airspeed indicates 201 knots. The flaps will automatically re-extend to the selected position of 15 when the airspeed falls below 196 knots.

It was at this time that the thrust was reduced slightly and the flaps auto-extended themselves back to 15. The thrust was then increased back to full power while pitch and airspeed increased. The flaps once again auto-retracted to 10 where they remained until impact.

The aircraft continued on a rather steep climb of 3150 feet/minute until reaching an altitude of nearly 3000 feet. A vertical velocity of 3000 feet/minute on a go around is steep but not necessarily a problem. The aircraft is by now light having burned its holding fuel, and is not carrying a full load. Pilots can choose to use somewhat less than full thrust in these types of situations but using full thrust is not procedurally wrong. It just means you might have a greater likelihood of overshooting an altitude or causing a flap overspeed which is what happened.

Pitch Over

It was at this time that the aircraft pitched over and began its fatal dive. The report states that the FDR recorded a simultaneous push on the control column accompanied by 12 seconds of forward stabilizer movement. Let's take these one at a time.

Pitching over to stop a steep climb is completely normal. Pitching over which results in a -1g acceleration is not. Normal gravity is 1g. Anytime you feel light in your seat, you're at something less than 1g. A zero g pushover would mean everything in the cabin would float. You'd come off your seat slightly. It would be uncomfortable and rarely happens. A negative 1g acceleration is effectively the same as if your seat was upside down and you were hanging from the seat belt. It would be extremely uncomfortable and short of extreme turbulence just doesn't happen. And it appears to have been caused by a combination of flight control and trim input.

Stabilizer trim is designed to compensate for airspeed changes affecting how the airplane flies. As an airplane increases speed, the nose naturally wants to come up to attempt to maintain the same airspeed as before. Pilots would have to keep pushing on the control column when accelerating to maintain level flight. Trim repositions the horizontal stabilizer to relieve this force. It is controlled by a thumb switch on the yoke which activates an electric motor to actually move the entire horizontal stabilizer and align it with the slipstream. 

On the 737 there is a large wheel next to the throttles which also is connected to the trim motor. This wheel allows manual positioning of the trim if the motor fails and also provides visual feedback to the pilots when the trim motor is running. Normal trim technique is to trim in bursts of one or two seconds. The trim motor on the 737 has two speeds for use depending on whether the flaps are extended or not. The high speed setting is active when the flaps are down. 

A 12 second run of the trim motor, especially in high speed mode, would never be needed during normal operations. This might lead to some speculation that the aircraft suffered what's known as "runaway trim" where the switch might stick and run the trim uncommanded. Boeing has provided several safeguards to prevent this from happening. There is an electric stab trim cutout switch located on the center control stand which removes electric power from the system. 

In addition, power is removed from the system any time pitch inputs don't match the control column inputs. This means the motor cannot trim forward when a pilot is pulling aft and vice versa. Lastly, the trim wheel is designed to be grasped and held when running if all else fails. With these safeguards, a runaway trim problem seems unlikely.

Final Dive

The report states that after the forward control column and trim inputs, the aircraft entered a dive after reaching a peak altitude of just under 3300 ft and subsequently hit the ground in a 50 degree nose low attitude at a speed of over 320 knots. Normally an airliner doesn't see a descending pitch attitude of more than 5 or 10 degrees nose low.

I honestly don't know what to make of this report. The crew flew a somewhat sloppy go around allowing the speed to increase where the flaps blew up, but that by itself wasn't gross or necessarily dangerous. The flap load relief system functioned as designed.

The question to be answered is why the pilots initiated the pushover and simultaneous forward trim? The airplane essentially dove into the ground. I am sure various mechanical or structural failure and other possible scenarios are being considered. The investigators have their work cut out for them.

This is just a preliminary report and the CVR tapes, while referenced in the report, have not been released to the public and may provide more context.