China Eastern 737-800 MU5735 accident March 2022
I plotted a GE 3D trajectory when the data became available (similar to the one subsequently posted). I didn't post it because the data is sparse and it didn't strike me as particularly informative ...
Join Date: Jun 2011
Location: Costa Rica
Age: 55
Posts: 48
Likes: 0
Received 0 Likes
on
0 Posts
My problem with the suicide theory is timing. They were at top of decent which tells me all pilots should be in the cockpit. I doubt even a Chinese captain could come up with an excuse good enough to get the other two pilots out of the cockpit just as they are about to begin the decent. Perhaps I'm wrong and they do things differently in China but I can't imagine any professional crew leaving one pilot alone during such a critical time.
Join Date: Aug 1998
Location: UK
Posts: 18
Likes: 0
Received 0 Likes
on
0 Posts
Sudden Altimeter Change
I remember years ago descending into CDG around 15,000 ft in a 737, when both altimeters suddenly ran down to zero and then back to the correct altitude within about 2-3 seconds. Unfortunately during this apparent descent the autopilot “captured the cleared descent altitude” we had preset (somewhere around 6,000 ft) … and then once we were “back at 15,000 ft” the started a rather aggressive dive back down towards the captured altitude, which I was only able to stop quickly by disconnecting the automatics … before resetting various (altitude & modes) and re-engaging autopilot & A/T.
Then had the same thing happen about 2 weeks later.
Although ASR’d at the time I don’t remember ever getting to the bottom of what caused it although my suspicion has always been that it had something to do with mobile phones being turned on (give away “nya-nya-nya” chirp radio interference around the same time).
Then had the same thing happen about 2 weeks later.
Although ASR’d at the time I don’t remember ever getting to the bottom of what caused it although my suspicion has always been that it had something to do with mobile phones being turned on (give away “nya-nya-nya” chirp radio interference around the same time).
flight) than it lost altitude. The following points also show the same lateral displacement. It was not a position glitch.
Many other (except we two), only did look at the course that seem to not have changed and from this they probably did draw the not correct conclusion that the aircraft did continue straight ahead with a nose down input. Which it did not.
The lateral displacement to the left need a (steep) left bank to take place. This part you will not see on a 2D dive profile from the side.
The SJ182 accident have the same lateral displacement (also 1600m) but that ADS-B track do not have a gap of 15-20s during the turn so for SJ182 those points show the left roll leading to the inverted dive/barrel roll.
Without 3D, or at least looking at this with 2d from both the side and from above, you wint notice the lateral displacement to the left that tell us that a left roll had a dominant part of the upset.
MU5735 initiated the upset by a left roll with positive G, probably just what we would see if a pilot get disoriented and fly a about 1G seat of the pants pitch during the roll.
The dive following the lateral displacement is probably inverted initially and the about 1G barrel roll continue until the aircraft is almost right on the wings again.
Engine failure, Auto throttle failure or something that caused the initial part of the roll followed by spatial disorientation or a IRS failure maybe.
Last edited by T28B; 13th Apr 2022 at 19:27. Reason: clean up spaces
When you did your 3D trajectory check, you did see a parallel movement from the last point before the upset until the first point with altitude loss that is in the order of 1600m. During these 15-20s the aircraft moved much more to the left (seen in the direction of
flight) than it lost altitude. The following points also show the same lateral displacement. It was not a position glitch.
Many other (except we two), only did look at the course that seem to not have changed and from this they probably did draw the not correct conclusion that the aircraft did continue straight ahead with a nose down input. Which it did not.
The lateral displacement to the left need a (steep) left bank to take place. This part you will not see on a 2D dive profile from the side.
The SJ182 accident have the same lateral displacement (also 1600m) but that ADS-B track do not have a gap of 15-20s during the turn so for SJ182 those points show the left roll leading to the inverted dive/barrel roll.
Without 3D, or at least looking at this with 2d from both the side and from above, you won't notice the lateral displacement to the left that tell us that a left roll had a dominant part of the upset.
MU5735 initiated the upset by a left roll with positive G, probably just what we would see if a pilot get disoriented and fly a about 1G seat of the pants pitch during the roll.
The dive following the lateral displacement is probably inverted initially and the about 1G barrel roll continue until the aircraft is almost right on the wings again.
Engine failure, Auto throttle failure or something that caused the initial part of the roll followed by spatial disorientation or a IRS failure maybe.
flight) than it lost altitude. The following points also show the same lateral displacement. It was not a position glitch.
Many other (except we two), only did look at the course that seem to not have changed and from this they probably did draw the not correct conclusion that the aircraft did continue straight ahead with a nose down input. Which it did not.
The lateral displacement to the left need a (steep) left bank to take place. This part you will not see on a 2D dive profile from the side.
The SJ182 accident have the same lateral displacement (also 1600m) but that ADS-B track do not have a gap of 15-20s during the turn so for SJ182 those points show the left roll leading to the inverted dive/barrel roll.
Without 3D, or at least looking at this with 2d from both the side and from above, you won't notice the lateral displacement to the left that tell us that a left roll had a dominant part of the upset.
MU5735 initiated the upset by a left roll with positive G, probably just what we would see if a pilot get disoriented and fly a about 1G seat of the pants pitch during the roll.
The dive following the lateral displacement is probably inverted initially and the about 1G barrel roll continue until the aircraft is almost right on the wings again.
Engine failure, Auto throttle failure or something that caused the initial part of the roll followed by spatial disorientation or a IRS failure maybe.
I'd want to see at least 5x as many data points before attempting to draw any conclusions about the detailed trajectory during the upset.
The FR24 "granular data" contains only 18 data points covering the last 2 minutes of flight, from FL290 until contact lost, so about 7 seconds apart on average.
I'd want to see at least 5x as many data points before attempting to draw any conclusions about
the detailed trajectory during the upset.
I'd want to see at least 5x as many data points before attempting to draw any conclusions about
the detailed trajectory during the upset.
no position.
But there is at least 100 points on the straight and level flight before the upset of which everyone is on a perfectly straight line viewed in 3D. No hickups or glitches. Not even a minor imperfection.
There is a 16s gap between the last normal point and the first in the upset. From that mostly 5s gap and I’m sure 5s is close enough when the existing point describe a spiral, to see how the trajectory is in between these 5s gap. For example in a loop you could understand the pull up in the end of a loop with points 5s apart, as the only logic way is the rounded trajectory in between. What we see is a ”about” 1G barrel roll.
Maybe not exactly but not very far from this.
There is a couple of 10s gaps, produced by the lack of a single data point( loss of one per five seconds create a 10s gap between points) but the position of these lost datapoints do not really affect the possibility to understand what happened.
Why, is not that easy though.
The SJ182 initial upset is a carbon copy, the parallel displacement is the same, the initial altitude loss is about the same so the trajectory of SJ182 can be used to understand how Mu5735 got from the fine cruise on level to the first datapoint in the upset.
So we more or less know “what” but we do not have the “why”.
Last edited by T28B; 13th Apr 2022 at 19:28. Reason: spacebloat
Join Date: Dec 2019
Location: OnScreen
Posts: 416
Likes: 0
Received 0 Likes
on
0 Posts
......
The SJ182 initial upset is a carbon copy, the parallel displacement is the same, the initial altitude loss is about the same so the trajectory of SJ182 can be used to understand how Mu5735 got from the fine cruise on level to the first datapoint in the upset.
So we more or less know “what” but we do not have the “why”.
The SJ182 initial upset is a carbon copy, the parallel displacement is the same, the initial altitude loss is about the same so the trajectory of SJ182 can be used to understand how Mu5735 got from the fine cruise on level to the first datapoint in the upset.
So we more or less know “what” but we do not have the “why”.
Implying, significant g-forces are required to accomplish a 3D-360 within something like 12 seconds, immediately followed by another 40 seconds with good directional stability. Easy to accomplish, when being completely shaken by a probably completely uncoordinated 3D-360 and stiff controls, due to the high speed, the aircraft obtained.
Also, it does not explain, WHY this diving stage happened again, this time, without the S-curve (or assumed 3D-360).
From earlier experiences, we do know that rudder hard-over is a difficult to control item at low speed, not at high speed. Unpleasant, but manageable at high speed.
From earlier experiences, we do know, that aerodynamic engine damage does not (have to) lead to control issues. Again, unpleasant, but manageable.
We do know, the aircraft got "stable" on course again (and even recovered from the initial dive), so the aerodynamic damage to the aircraft isn't that big (at that moment).
Before the upset, MU5735 was flying nearly double the ground speed of SJ182. Ground speed is what counts for the kinetic and subsequent rotational energy of an object, implying nearly 4 (OK, maybe closer to 3.5) times the energy for MU5735. That increase in energy implies the same 3.5 times increase for the mass related forces involved and 2^3 (=8, round down to 6-7 for the effect of not completely the double ground speed) times for the aerodynamic forces involved.
1) The parallel displacement do not aquire any more G-force to move an object for example 1600m at 16 seconds if the speed is higher. Of course the trajectory will be spread out on a longer distance. Just as in the case of MU5735.
2) Rotational energy? There will be nothing that hinders a aircraft to roll for example 90 degrees in X seconds just because the groundspeed/airspeed is higher. The amount or roll per travelled distance will be lower though. Just as in the case of MU5735.
During the left roll, the diving barrel roll and the pull upp there is a clear understandable trajectory, and the "low resolution" doesnt really matter to see the plot. After the pull up, the plot is not that easy to se and stable flight is not what I see.
If anything is to be said about the after pull up, it is that maybe the aircraft did continue to barrel roll and in that case continued to roll inverted and that the final part of the dive was inverted. There is another parallel displacement to the left,which might come from another half barrel roll. But there is a another 16s gap about that position so it is hard to "know" the trajectory at that point.
Last edited by AAKEE; 13th Apr 2022 at 22:58.
This is the SJ182 equivalents of the MU5735 above. A few points hidden during the initial upset to create a similar gap.
Upper picture in both cases is from straight behind, viewed from the trajectory of the normal flight. The most right dot is several 3d points lining up in a perfect line, because of this we only see on dot.
We can see that SJ182 struck the water about in the middle of the original track and the 0.9Nm maximum displacement. Looking at the MU5735 we can see that when the altitude loss was similar to the SJ182 sturking the water, MU5735 also was about in the middle between the 0.9Nm displacement and the original track.
Join Date: Jan 2020
Location: Bucharest
Posts: 93
Likes: 0
Received 0 Likes
on
0 Posts
Now imagine you have a small drone, and you fly it inside that train. Again, it would be irrelevant if the train were moving at 100km/h or 200km/h relative to the ground, or if it were stationary. It would have no effect on the flight of the drone inside the train, and the G forces and aerodynamic effects it experiences.
Join Date: Dec 2019
Location: Copenhagen
Posts: 3
Likes: 0
Received 0 Likes
on
0 Posts
Wait, what? Speed is relative. As a result, kinetic energy is relative as well. Let's say you are sitting down in a train moving at a constant speed of 100 km/h relative to the ground, in a straight line. You as a passenger would have zero kinetic energy relative to the train frame of reference. And if the train wouldn't have any windows (and assuming you can't rely on noise either), you wouldn't even be able to tell if the train is stationary or is moving.
Now imagine you have a small drone, and you fly it inside that train. Again, it would be irrelevant if the train were moving at 100km/h or 200km/h relative to the ground, or if it were stationary. It would have no effect on the flight of the drone inside the train, and the G forces and aerodynamic effects it experiences.
Now imagine you have a small drone, and you fly it inside that train. Again, it would be irrelevant if the train were moving at 100km/h or 200km/h relative to the ground, or if it were stationary. It would have no effect on the flight of the drone inside the train, and the G forces and aerodynamic effects it experiences.
Join Date: Jan 2008
Location: Netherlands
Age: 46
Posts: 343
Likes: 0
Received 0 Likes
on
0 Posts
Wait, what? Speed is relative. As a result, kinetic energy is relative as well. Let's say you are sitting down in a train moving at a constant speed of 100 km/h relative to the ground, in a straight line. You as a passenger would have zero kinetic energy relative to the train frame of reference. And if the train wouldn't have any windows (and assuming you can't rely on noise either), you wouldn't even be able to tell if the train is stationary or is moving.
Now imagine you have a small drone, and you fly it inside that train. Again, it would be irrelevant if the train were moving at 100km/h or 200km/h relative to the ground, or if it were stationary. It would have no effect on the flight of the drone inside the train, and the G forces and aerodynamic effects it experiences.
Now imagine you have a small drone, and you fly it inside that train. Again, it would be irrelevant if the train were moving at 100km/h or 200km/h relative to the ground, or if it were stationary. It would have no effect on the flight of the drone inside the train, and the G forces and aerodynamic effects it experiences.
Join Date: Apr 2000
Location: the netherlands
Posts: 15
Likes: 0
Received 0 Likes
on
0 Posts
Well it seems that they have restarted using the 737-800 in China, so I guess even though no announcements have been made they must have discovered something that caused the accident,
Join Date: Dec 2019
Location: OnScreen
Posts: 416
Likes: 0
Received 0 Likes
on
0 Posts
Intriguing is, we don't hear anything about the CVR/FDR results. Given the MU B737-800 are still grounded, unless the CVR and FDR are not readable (unlikely), this accident does not seem to be a deliberate crew action.
Join Date: Dec 2019
Location: OnScreen
Posts: 416
Likes: 0
Received 0 Likes
on
0 Posts
Wait, what? Speed is relative. As a result, kinetic energy is relative as well. Let's say you are sitting down in a train moving at a constant speed of 100 km/h relative to the ground, in a straight line. You as a passenger would have zero kinetic energy relative to the train frame of reference. And if the train wouldn't have any windows (and assuming you can't rely on noise either), you wouldn't even be able to tell if the train is stationary or is moving.
....
....
When you (with mass m) jump off the train with speed v-train, you are going to dissipate: 0.5*m*(v-train)^2.
When you run on the train, with v-run, and subsequently jump off, you are going to dissipate: 0.5*m*(v-train+v-run)^2 (and not 0.5*m*(v-train)^2 + 0.5*m*(v-run)^2).
When you run on the train, backwards, where v-run= - v-train and jump off, you are going to dissipate: 0.
So, debunked (and I leave out the rest of your reasoning, since that qualifies for the same).
Only half a speed-brake
Answer why keep them grounded might be more simple than anything - the flight operation is so severely cut CEAir may be even keeping other types parked these weeks due lack of work.
Join Date: Dec 2019
Location: OnScreen
Posts: 416
Likes: 0
Received 0 Likes
on
0 Posts
Add "Canceled" and "Flown" and you get roughly to the 2021 levels. So, there was a plan.
Shanghai and many other big China cities got into lock-downs or under travel-limitations outside the own environment, around 1 month ago.
So, yeah, these items made it easy for MU to swap out the B737-800 against other airplanes, for example the B737-700 and the A32x. Especially the A32x, MU does have quite a lot of these.
Though, effectively, MU still does not use the B737-800, whereas it can be expected that at least part of the routes could be served with this type, on busy days. But it isn't.
We just have to wait what is going to happen. Assuming the CVR/FDR are read-out by now and no other measures are announced (or taken by MU), more or less implies, it is unclear what happened. Or better, why this accident happened.
Shanghai and many other big China cities got into lock-downs or under travel-limitations outside the own environment, around 1 month ago.
So, yeah, these items made it easy for MU to swap out the B737-800 against other airplanes, for example the B737-700 and the A32x. Especially the A32x, MU does have quite a lot of these.
Though, effectively, MU still does not use the B737-800, whereas it can be expected that at least part of the routes could be served with this type, on busy days. But it isn't.
We just have to wait what is going to happen. Assuming the CVR/FDR are read-out by now and no other measures are announced (or taken by MU), more or less implies, it is unclear what happened. Or better, why this accident happened.
Only half a speed-brake
To avoid confusing someone, the above graph is the nation-wide domestic sectors, not only MU. Cutoff-point between not-planned against planned but cancelled is not clear.
The total planned count matches the volume of 2021 APR-JUL which was the same as 2019 BTW.
Online sources suggest before the crash around 40 of the -800 were in service compared to 60 stored, accompanied by 25ish -700 all in active service.
Whereas the single-aisle Airbus fleet had 300 active planes flying, above those parked.
40+30+400 = 470
470 / 3 (optimistic, double the real today's sector count) = around 150 planes needed and 320 left unused beyond those mothballed.
Keeping the 800s grounded altogether, with 40 hulls, is an irreleveant decision that costs nothing and has some marketing potential for damage control.
The total planned count matches the volume of 2021 APR-JUL which was the same as 2019 BTW.
Online sources suggest before the crash around 40 of the -800 were in service compared to 60 stored, accompanied by 25ish -700 all in active service.
Whereas the single-aisle Airbus fleet had 300 active planes flying, above those parked.
40+30+
Keeping the 800s grounded altogether, with 40 hulls, is an irreleveant decision that costs nothing and has some marketing potential for damage control.
Last edited by FlightDetent; 17th Apr 2022 at 23:41.