FlyDubai 737 crash in Russia
03-20-2016, 07:14 PM
#51
Banned
Joined APC: Jun 2008
Posts: 8,350
Simply a consideration, but what I read indicated the low level jet was about 5000 feet and they began a missed approach about 1700 feet, climbed to about 5500 and control was lost to nose down pitch. If they were at slow speed reconfiguring to a clean wing and encountered a rapid and strong windshift (especially with any bank angle due to turbulence), it MAY have exacerbated a stall and loss of control, not unlike that on the back side of a microburst. Also, it's possible icing could have played a factor and it's not known what ice protection systems were in operation. The 738 has surprisingly little airfoil ice protections systems with only a few slats on the wing and nothing on the tail. A suddenly clean wing (slats retracted) just when hitting a strong shear at minimum speed with bank, ice or both and..........? An ice related tail stall might also be a possibility if the right conditions were met.
It will be interesting to see what the boxes reveal.
03-20-2016, 08:50 PM
#52
Gets Weekends Off
Joined APC: Jul 2007
Position: B737 /FO
Posts: 345
Nice explanation, eaglefly. I never encountered a low level jet stream before, nor seen one reported. It seems very possible they could have been exposed to significant ice followed by a go around in shear conditions.
Just got a picture back for the black boxes:
Russia plane crash: Rostov-on-Don black boxes 'damaged' - BBC News
Just got a picture back for the black boxes:
Russia plane crash: Rostov-on-Don black boxes 'damaged' - BBC News
03-21-2016, 01:37 AM
#53
Banned
Joined APC: Jun 2008
Posts: 8,350
Nice explanation, eaglefly. I never encountered a low level jet stream before, nor seen one reported. It seems very possible they could have been exposed to significant ice followed by a go around in shear conditions.
Just got a picture back for the black boxes:
Russia plane crash: Rostov-on-Don black boxes 'damaged' - BBC News
Just got a picture back for the black boxes:
Russia plane crash: Rostov-on-Don black boxes 'damaged' - BBC News
03-21-2016, 03:37 PM
#54
Somewhere in Europe
Joined APC: Jan 2010
Position: A330 FO
Posts: 117
03-21-2016, 04:12 PM
#55
Gets Weekends Off
Joined APC: Jun 2014
Posts: 923
03-21-2016, 05:19 PM
#56
New Hire
Joined APC: Mar 2016
Position: Retired
Posts: 7
Hi all,
I'm new here - been on pprune for a decade but got fed up with the discussion (un-)culture there. I've been lurking as an anonymous spectator for some time - this is my first contribution.
Taking the FR24 data and calculating accelerations and rates for altitude shows an interesting picture:
The blue points are the actual data points, the blue line a spline fit (so not necessarily physical, remember that). The red points and line are two data points averaged. The x axis is time in seconds before the last data point.
You can see that from the time of the go around onwards, the climb rate keeps increasing for almost 20 seconds at an increasing acceleration. To produce that, one would have to pull back on the yoke for 20 seconds with increasing force, or not counteract a TOGA thrust induced pitch moment (sufficiently).
The peak climb rate achieved is 4700fpm at 44s, this drops to 2500fpm at 25s, and increases again to 3500fpm by 19s, before following an essentially -1G parabolic trajectory into the ground.
The last 10 seconds are essentially spent in free fall - accelerating at about -1G (-0.93G = -9.1 m/s^2). The only way an airliner will fly that kind of profile on its own is if it is in/near a 90 degree bank, or with pitch trim full forward perhaps?
From the videos we know that the wings were attached and that the aircraft was in a steep bank. Can anyone with experience on the type and making some mass assumptions figure out how much energy would be left at the top of the rather steep GA climb? And how long would it take to bring pitch trip fully forward?
I'm new here - been on pprune for a decade but got fed up with the discussion (un-)culture there. I've been lurking as an anonymous spectator for some time - this is my first contribution.
Taking the FR24 data and calculating accelerations and rates for altitude shows an interesting picture:
The blue points are the actual data points, the blue line a spline fit (so not necessarily physical, remember that). The red points and line are two data points averaged. The x axis is time in seconds before the last data point.
You can see that from the time of the go around onwards, the climb rate keeps increasing for almost 20 seconds at an increasing acceleration. To produce that, one would have to pull back on the yoke for 20 seconds with increasing force, or not counteract a TOGA thrust induced pitch moment (sufficiently).
The peak climb rate achieved is 4700fpm at 44s, this drops to 2500fpm at 25s, and increases again to 3500fpm by 19s, before following an essentially -1G parabolic trajectory into the ground.
The last 10 seconds are essentially spent in free fall - accelerating at about -1G (-0.93G = -9.1 m/s^2). The only way an airliner will fly that kind of profile on its own is if it is in/near a 90 degree bank, or with pitch trim full forward perhaps?
From the videos we know that the wings were attached and that the aircraft was in a steep bank. Can anyone with experience on the type and making some mass assumptions figure out how much energy would be left at the top of the rather steep GA climb? And how long would it take to bring pitch trip fully forward?
03-21-2016, 09:10 PM
#57
Gets Weekends Off
Joined APC: Jul 2007
Position: 767 Captain
Posts: 3,989
I get the x-axis is time on both graphs, but please explain the y-axis. Pretend some of us don't know the significance of a "spline fit". 
What are the units? Time = 0 and the vertical rate starts at -90....what?
If time zero is the start of the go around, how is it that both y-axis values in each graph are starting in negative territory? If the y-values are the graph titles (vertical rate and vertical acceleration): How could either of those components remain negative values for what looks like the first 15 seconds of an aircraft go around. What data source did you use to determine this?
What are the units? Time = 0 and the vertical rate starts at -90....what?If time zero is the start of the go around, how is it that both y-axis values in each graph are starting in negative territory? If the y-values are the graph titles (vertical rate and vertical acceleration): How could either of those components remain negative values for what looks like the first 15 seconds of an aircraft go around. What data source did you use to determine this?
03-21-2016, 09:17 PM
#58
New Hire
Joined APC: Mar 2016
Position: Retired
Posts: 7
Sorry for not being more clear: time = 0 for the last datapoint, time = 60 at the initiation of the go around. So the time axis is more like a countdown.
To read the plots in correct time order, read them from right to left: 120 seconds are shown, so the last 2 minutes of the flight. Starting with a glideslope intercept at 120s to the right.
The unit in the Y axis in each plot is in square brackets in the title: meters per second (20 m/s ~ 40 kts ~ 4000fpm) for the vertical rate, and acceleration (i.e. the rate of change of the vertical velocity or climb rate) is of course given in m/s per second, or m/s^2.
A spline fit is simply a curve fit based on a square (or cube) function: https://en.wikipedia.org/wiki/Spline_interpolation
It interpolates (i.e. calculates missing points inbetween known points, under the assumption that the behaviour of the data is smooth between the points)
So on the example of the vertical rate plot: Start at the right at 120s, vertical rate is 0, then at 110s they intercept the GS and vert rate goes to about -3m/s (~-600fpm), this is maintained until go around at 60s where the rate increases to 4700fpm.
Oh and the last edit: the data source are the flightradar24 points.
I hope this helps...
To read the plots in correct time order, read them from right to left: 120 seconds are shown, so the last 2 minutes of the flight. Starting with a glideslope intercept at 120s to the right.
The unit in the Y axis in each plot is in square brackets in the title: meters per second (20 m/s ~ 40 kts ~ 4000fpm) for the vertical rate, and acceleration (i.e. the rate of change of the vertical velocity or climb rate) is of course given in m/s per second, or m/s^2.
A spline fit is simply a curve fit based on a square (or cube) function: https://en.wikipedia.org/wiki/Spline_interpolation
It interpolates (i.e. calculates missing points inbetween known points, under the assumption that the behaviour of the data is smooth between the points)
So on the example of the vertical rate plot: Start at the right at 120s, vertical rate is 0, then at 110s they intercept the GS and vert rate goes to about -3m/s (~-600fpm), this is maintained until go around at 60s where the rate increases to 4700fpm.
Oh and the last edit: the data source are the flightradar24 points.
I hope this helps...
Last edited by itsphysics; 03-21-2016 at 09:32 PM.
03-22-2016, 06:46 AM
#59
Banned
Joined APC: Jun 2008
Posts: 8,350
Interesting.
The 738 has 2 "TOGA" (go-around thrust requests via auto-throttle) selections normal and full. Full isn't used often, but if it was selected (two applications of the throttle TOGA buttons) and the aircraft is trimmed for approach speed (go-around flown manually without autopilot), it could very well pitch itself up into a very steep climb gradient and possibly into a stall if not managed as it is trimmed for a low airspeed (most swept-wing stalls also result in falling off on a wing to a steep bank angle as noted also in the Bagram 747 crash). Configuration changes could aggravate this situation. Turbulence and wind shear were possibilities so perhaps they went for full TOGA and it got away from them on an IMC go-around in poor conditions ?
I still also wonder about icing and definitely fatigue.
The 738 has 2 "TOGA" (go-around thrust requests via auto-throttle) selections normal and full. Full isn't used often, but if it was selected (two applications of the throttle TOGA buttons) and the aircraft is trimmed for approach speed (go-around flown manually without autopilot), it could very well pitch itself up into a very steep climb gradient and possibly into a stall if not managed as it is trimmed for a low airspeed (most swept-wing stalls also result in falling off on a wing to a steep bank angle as noted also in the Bagram 747 crash). Configuration changes could aggravate this situation. Turbulence and wind shear were possibilities so perhaps they went for full TOGA and it got away from them on an IMC go-around in poor conditions ?
I still also wonder about icing and definitely fatigue.
03-22-2016, 12:33 PM
#60
Banned
Joined APC: Dec 2015
Position: Aeroflot
Posts: 179
[QUOTE=itsphysics;2093776]Sorry for not being more clear: time = 0 for the last datapoint, time = 60 at the initiation of the go around. So the time axis is more like a countdown.
Using FR24 is hardly an accurate source for any flight data. Everyone here is an armchair expert. No one has a clue what happened.
Using FR24 is hardly an accurate source for any flight data. Everyone here is an armchair expert. No one has a clue what happened.
Thread
Thread Starter
Forum
Replies
Last Post
