Boeing simulations of MH370 at end-of-flight. (Click on image to enlarge.)
Introduction
With the failure of the recent search by Ocean Infinity to locate the debris field along the 7th arc at latitudes as far north as 25S and at widths equal to or exceeding +/- 22 NM, we consider whether it is possible that with the assumption of uncontrolled flight, the plane could have impacted the ocean farther than 22 NM from the 7th arc, and was missed because the search along the arc was too narrow.
At each phase in the subsea search for MH370, the search area was defined by following parameters:
- A range of latitudes along the 7th arc
- A width inwards and outwards from the 7th arc
The range of latitudes for the search was determined by analysis of the satellite data, aircraft performance, and drift models. Collectively, these data sets constrain the position that MH370 crossed the 7th arc. Unfortunately, due to the imprecise nature of these data sets, official and independent investigators have proposed a fairly large range of latitudes, and this continues to be the subject of debate.
In a parallel effort by the ATSB, the width inwards and outwards from the arc was estimated by the distance the aircraft could have glided after fuel exhaustion. With the assumption that there were no pilot inputs after fuel exhaustion, this distance was determined by end-of-flight simulations conducted by Boeing based on aircraft conditions requested by the ATSB. The simulations were completed in Boeing’s engineering simulator, which offers high fidelity of the aircraft’s performance. Some of the results of these simulations were presented in an ATSB report entitled MH370 Search and Debris Update, released in November 2016.
Simulation Results
Because of the importance of the simulations in establishing the search width, I asked the ATSB for more details about the conditions modeled and the results of the ten simulations. (Because Boeing conducted the simulations based on conditions requested by the ATSB in support of the search, they are an ATSB product.) Although legal restrictions prevent the ATSB from providing specific details about each of the ten simulations (referred to as Cases 1-10), the ATSB did reveal that the initial speeds varied between M0.75 and M0.83, the starting altitudes were either FL350 or FL400, the initial headings were either 178°, 184° or 190°, and turbulence was either light or moderate. Winds at various altitudes were included in the simulations.
In addition to these general descriptions, the ATSB did provide me with numerical results of the ten simulation runs in the form of X, Y, and altitude in one second increments, where X is the E-W position in NM, Y is the N-S position in NM, and the altitude is in feet. (The ATSB deliberately did not include the latitude and longitude values in these files so that there was no association with a particular crossing of the 7th arc.) The ATSB has permitted me to share these results on this blog so that a broader group of investigators can independently analyze the results. To my knowledge, this is the first time these results have been released to the public.
Although not included in the files, derived quantities such as groundspeed, Mach number, track, vertical speed, vertical acceleration, bank angle, and wing loading are all calculable, although some parameters such as Mach number and calibrated airspeed need assumptions about the wind and temperature fields.
A presentation of the detailed technical analyses of all the simulations would be lengthy and not appropriate for a blog post. However, some generalized observations and results are summarized below:
- There were two groups of simulations. In the first group (Cases 1,2,5,7,8,9), the autopilot was lost after the flameout of the second (left) engine, which corresponds to a “normal” configuration of the electrical system.
- In the second group (Cases 3,4,6,10), the electrical configuration was in an “alternate” configuration in which the left generator and left backup generator were isolated with switches in the overhead panel. In this case, when the right engine is running, power is supplied to the left bus from the right bus through the bus tie breaker. However, after the first (right) engine flames out, both main AC busses and both transfer busses lose power, which causes the flight control mode to transition to “secondary” and the autopilot and autothrottle disengage, even as the left engine continues to provide thrust.
- The SATCOM’s log-on request at 00:19 is assumed to occur two minutes after the loss of the autopilot. (The two minute interval assumes the APU would take one minute to start supplying power, and the SDU would subsequently take one minute to power up and request a log-on to the satellite.) The ATSB has adjusted this position in all the simulations so that the two-minute mark occurs at the same X,Y values, and the paths can be compared. The two-minute mark serves as a reference position (labeled with “2 mins” in the figure above).
- For all ten simulations, the paths stay within a distance of 32 NM from the reference position. However, since all paths are curved, the farthest impact point from the 7th arc is only about 15 NM, and depends on the track angle between the initial path and the arc tangent.
- There were five simulations (Cases 3,4,5,6,10) in which the rate of descent exceeded 15,000 fpm and the downward acceleration exceeded 0.67g, which are the values indicated by the two final values of BFO. This included all four cases with the “alternate” electrical configuration, and one with normal configuration. The maximum downward acceleration for these cases ranged between 0.87g and 1.30g.
- For the five simulations (Cases 3,4,5,6,10) with high descent rates, the impact occurs within 14 NM from the reference position, and about the same distance from the 7th arc, depending on the track angle between the initial path and the arc tangent.
- For the five simulations (Cases 3,4,5,6,10) with high descent rates, the descent rate of 15,000 fpm and the downward acceleration of 0.67g occur at different times, and are not predicted to occur at the time of the log-on. If the impact distance is measured from the point at which the descent rate first exceeds 15,000 fpm, the distance ranges between 4.7 NM and 7.9 NM.
- In none of the simulations did the plane fly straight with level wings after the autopilot was disengaged. Ultimately, the magnitude and direction of the bank that develops is the net effect of a many factors, including thrust asymmetry, TAC, manual rudder input, weight imbalance, aerodynamic asymmetry, and turbulence, with the dihedral effect of the wings and center-of-mass tending to restore the bank to zero.
Simulation of Alternate Electrical Configuration
Case 6: A simulated end-of-flight path for the left generators turned off before fuel exhaustion. (Click on image to enlarge.)
The four simulations (Cases 3,4,6,10) with the alternate electrical configuration are interesting because in all cases, the descent rates and downwards accelerations exceeded the values suggested by the final BFOs, and the high speed descent is achieved within minutes of loss of the autopilot. To further illustrate the sequence of events, we consider Case 6, with initial conditions of a groundspeed of 425 knots at an altitude of 35,000 ft, and shown in the figure above.
When there is no more fuel in the right tank, the right engine coasts down and the left engine speed increases to full thrust. The loss of AC power causes a transition to secondary control mode, and the autopilot and the autothrust are disengaged. As the right engine coasts down, the automatic thrust asymmetry compensation (TAC) will apply left rudder, but the amount of rudder does not change after the flight control mode degrades to secondary.
After the right engine flames out, there is a slight turn to the right, which means the rudder position does not completely balance the yaw from the resulting thrust asymmetry. It is possible that the transition to secondary control mode occurs before the TAC can apply sufficient rudder input to fully compensate for the thrust asymmetry of no right engine thrust and full thrust of the left.
At some point, the remaining fuel in the left tank is exhausted, and the left engine shuts down. With both engines producing no thrust, the left rudder position now causes a steep turn to the left, and the plane descends. At the point it reaches 27,000 ft, the descent rate is 15,000 fpm, and it has achieved a downward acceleration of 0.3g over the preceding eight seconds. The plane levels off at about 19,000 ft, and then begins to again rapidly descend, reaching a descent rate of 15,000 fpm at about 18,000 ft, and impacting the sea about 3.3 NM from this point. The impact distance from the earlier point of reaching a descent rate of 15,000 fpm is about 7.9 NM.
Discussion
[Phrases in italics were added on Aug 21, 2018 for clarity.]
If we consider that the end-of-flight Boeing simulations were representative of the actual conditions of MH370, we have three possibilities:
- If the flight was uncontrolled after fuel exhaustion, and if we ignore the final BFO values, the plane impacted no farther than about 15 NM from the 7th arc.
- If the flight was uncontrolled after fuel exhaustion, and if the descent rates suggested by the final BFO values are valid, then the aircraft impacted no more than about 8 NM from the 7th arc.
- If the flight was controlled after fuel exhaustion, an efficient, stable glide starting at about 40,000 ft could have achieved a distance of 120+ NM from the 7th arc.
Possibilities (1) and (2) suggest the next search should occur along the 7th arc north of 25S latitude at a width of about +/-25 NM. The +/-25 NM is an estimate that includes uncertainty in the simulation models and uncertainty in the final BTO values.
Because possibility (3) requires pilot inputs after fuel exhaustion, if this occurred, we have to assume that pilot inputs also occurred during the powered part of the flight after 19:41. As such, there would be no reason to believe that paths reconstructed with the assumption of automated flight with no pilot inputs are representative of the path actually flown. The large width (+/- 120 NM) and the difficulty in objectively constraining the crossing latitude of the 7th arc would make it difficult to define a manageable search area without additional information.
Until additional information or insights become available, it will be difficult to establish a new search area.
@TimR: If you could provide us with more specifics or proof of your assertions regarding the intentions of the captain, it would help justify a search further north of 25S. As of today, most people treat it as an unsubstantiated rumor, of which there have been many.
Perhaps, in drawing a pause on the radar/ADS-B information this is an update for VATM’s upgrade progress in 2013.
Given the ADS-B mandates by Singapore and Australia prior to 2014, MAS will have been operating ADS-B compliant transponders on the B777 fleet during 2014. I’d like to confirm if the transponder included “Aircraft Derived Data” (ADD) in its broadcast, the data that meets enhanced surveillance requirements. If so, any data logged from the VATM Con Son receiver may be significant.
Today, I noted one paragraph in the SIR, under Synopsis, page xv:
The investigation [..] comprised an Investigator-in-Charge (IIC) and three main Committees, comprising: Airworthiness; Flight Operations; and Medical/Human Factors.
No indication that a Structures committee was convened as aircraft debris began to be recovered.
Forget about SIO. Very interesting CSIRO study about marine debris. Modelled marine litter usually enter the Indian Ocean from Southeast Asia within one to two years (Figure 20).
https://image.ibb.co/hLh9az/Particles_Southeast_Asia.jpg
Study here https://publications.csiro.au/rpr/download?pid=csiro:EP159268&dsid=DS3
@Victor Iannello
“The SATCOM’s log-on request at 00:19 is assumed to occur two minutes after the loss of the autopilot. (The two minute interval assumes the APU would take one minute to start supplying power, and the SDU would subsequently take one minute to power up and request a log-on to the satellite.)”
How long do you assume the ADIRU would take to boot? Having lost it’s reference heading whilst in-flight, it may may take even longer for the output to be reliable.
If the AES is sucking the bus, can the initial data be relied upon?
@Victor. Nice that you got more info finally and thank you for the work in your analysis and presentation.
Some off the top of the head starter observations and questions:
• Achieving the descent rates is much dependent on the right engine power loss configuration. If the left (‘normal configuration’) then but 1 of 6 did.
• You indicate the timings do not match the BFOs. Have you calculated how far off 2mins 7secs were the times to that descent rate?
• If achieving the descent rates qualitatively is a go/no go, that indicates the right is the only acceptable. That suggests that if unmanned the right IDG would have been switched on at 18:24 to restore the SDU.
• However your anlysis shows that within the limitations of the simulations, achieving those descent rates does not matter much if there is a 25NM limit.
• On the limitations, did they say what the assumption was as to the time difference between engine flame outs; was there a spread; and was the increased consumption due to the right providing all electrical power for 6 hrs included in those?
• They did not simulate right or left engine relights or the effect of APU start. On the latter I am thinking of the APU restoring AC, though not of course the autopilot with that.
• Statistically the numbers are small. “the farthest impact point from the 7th arc is only about 15 NM” is an example. Aside from a bit of a swoop in at the end it may not be representative of the true maximum.
• I think some of these points need to be clarified before a general conclusion can be made about ±25NM, though your analysis certainly helps with that.
• For those who find Cases 6 and 9 and even 10 hard to make out I assume the ATSB original to be the same?
An observation on the paths vs attitude to I-3F1.
At the 2 minute mark, after loss of generator power, paths 4, 3, 6, and 10 put the aircraft in a good position for line of sight abeam from an antenna aperture to I-3F1. Paths 1, 2, 5, 7, 8, and 8 put the aircraft longitudinal orientation close to a radial line from I-3F1. In other words, the satellite is directly behind the aircraft and approx 40º above its position (antenna look angle depends on pitch).
If the initial LOR burst from the aircraft was ‘missed’ by the GES, add 12-19 seconds before a second LOR occurs. If that was to have occurred, the path terminus for all but 8 and 4 shift closer to the arc.
@flatpack
The ADIRU operation is not interrupted, its power supply redundancy is assured by the main battery.
In reading about the FlightRadar24 data reports, I see several squawk codes containing the numerals 8 and 9. These are invalid values as squawk codes are octal (ranging from 0 to 7) as has been mentioned. If the reports from Paul Sladen do indeed show these 8’s and 9’s, it calls into question the reliability of the all the reports.
@Don Thompson
Thanks.
Slightly off topic but are there any non battery-backed inputs to the ADIRU that could cause an erroneous output. I was thinking along the lines of AF447 and pitot tube heaters but they would presumably require a longer power interruption to ice-up.
@Bruce Robertson: The (octal) squawk code for MH370 was 2157. I discovered that if this is (incorrectly) interpreted as hex, it would convert to 8535 decimal, which is the squawk code listed in a graphic supplied by FR24. I think the conversion was performed incorrectly by whomever prepared this graphic. On the other hand, the squawk code in the CSV file from FR24 correctly lists the squawk code as 2157. I don’t think the incorrect squawk code indicates a problem with the data.
@David: You’re asking lots of questions. I appreciate your interest, and I’ll answer some, but be aware that I asked permission from the ATSB to share the CSV files for each simulation so that others can perform their own calculations.
Have you calculated how far off 2mins 7secs were the times to that descent rate?
I’ve done it for some, but not all. It’s easily calculated for each case.
That suggests that if unmanned the right IDG would have been switched on at 18:24 to restore the SDU.
The left bus could have been isolated by opening the breakers for the left bus tie, the left IDG, and (optionally) the left backup generator. Power at 18:24 could have been restored by restoring the left tie bus breaker to AUTO. There was no need to ever switch the right IDG.
On the limitations, did they say what the assumption was as to the time difference between engine flame outs; was there a spread; and was the increased consumption due to the right providing all electrical power for 6 hrs included in those?
I do not know the fuel levels in the tanks at the start of the simulations.
On the latter I am thinking of the APU restoring AC, though not of course the autopilot with that.
I do not believe this effect on the flight dynamics, if any, was modeled.
Statistically the numbers are small. “the farthest impact point from the 7th arc is only about 15 NM” is an example. Aside from a bit of a swoop in at the end it may not be representative of the true maximum.
If we believe the descent rates that we derived from the final BFOs, the distance from the 7th arc is less than 8 NM with no pilot inputs. Once the aircraft achieves high rates of descent and high downward accelerations, the flight dynamics becomes increasingly unstable.
For those who find Cases 6 and 9 and even 10 hard to make out I assume the ATSB original to be the same?
I’m not sure what you are asking. If you are having problems distinguishing each path in the first figure, you can use the files I provided to obtain the precise paths.
@Don Thompson
RE: “Given the ADS-B mandates by Singapore and Australia prior to 2014, MAS will have been operating ADS-B compliant transponders on the B777 fleet during 2014. I’d like to confirm if the transponder included “Aircraft Derived Data” (ADD) in its broadcast, the data that meets enhanced surveillance requirements. If so, any data logged from the VATM Con Son receiver may be significant.”
Perhaps I’m mistaken, but doesn’t the DF17 message frame that’s broadcast via the Mode S extended squitter (ie ADS-B) only include items such as aircraft identification, position, altitude & emergency status? Other ADD that meets the Mode S enhanced surveillance requirements (ie selected altitude, roll angle, IAS, magnetic heading, etc) would only be sent in response to a Mode S interrogation.
@Don Thompson
I’ll have to check, but I recall the Mode S enhanced surveillance requirements were implemented for operations in Europe some time around 2005. I don’t have confirmation, but given that Malaysian’s B777s were used on its routes to Europe, I’d say the aircraft must have been compliant.
@Victor. Thank you. A couple of comments then I will leave this alone for a while.
Yes, “If we believe the descent rates that we derived from the final BFOs, the distance from the 7th arc is less than 8 NM with no pilot inputs.” But also we must believe that the right engine bore all the electrical load.
In that case the 18:24 restoration of power, later unmanned, would have entailed the right IDG being selected if both had been selected off earlier, or loaded through a bus tie which had been opened earlier when the left IDG was selected off as you say. I think most likely the left back up gen would have been selected off already in that hypothetical, for some reason, that reason going a-begging and raising the possibility that the right B/U had been selected off earlier too. Even harder to explain would be how the left B/U would have been isolated at 18:24 or later I would suggest.
It might be better to forego the right engine sub-set that comes with a BFO realisation requirement since 25 NM seems good enough.
That wider margin also could allow short glide distances eg for a pilot who glided for a while before plunging.
As to the cases hard to make out I meant in your top diagram some are indistinct.
@Victor Iannello,
For the alternate power configuration, it seems rather likely that the left engine would continue operating some 8-10 minutes after the right engine failed (causing the autopilot to be disconnected and the SPU to auto-start). There are 3 main contributors to the run time difference, assuming no fuel was transferred after17:07. First, there was less fuel in the right tank than in the left tank at 17:07. Second, the right engine consumed about 1.5% more fuel. Third, the right engine had twice the electrical load as normal, increasing its fuel consumption, while the left engine had no electrical load.
Do you know, or have you inferred from the ATSB data, how long the engine run time difference was assumed to be in the simulations using the alternate power configuration? I don’t see how less than 2 minutes is possible.
@Victor Iannello,
Perhaps I missed it, but was it the case that the several simulations for each power configuration only differed by the degree of residual rudder mis-trim?
@ Victor Iannello
Don Thompson wrote “The final F24 record, timed at 17:21:03, records only the 4 digit ATC assigned octal code (2195).” What squawk code is this? I haven’t been able to convert it to anything reasonable.
@flatpack asked “are there any non battery-backed inputs to the ADIRU that could cause an erroneous output“.
Yes, some inputs to the Air Data side of the ADIRU may be unreliable. However, the IRU function isn’t affected. The data words used by the AES originate in the IRU function.
@Bruce Robertson asked “ATC assigned octal code (2195). What squawk code is this?”
It’s simply a unique code, from a defined pool of numbers, assigned by ATC to the a/c crew before departure that enables the a/c to be identified by SSR interrogation. Or, while enroute, a code is assigned by the ATC managing an FIR the aircraft enters.
2100-2177 is a range assigned to Malaysia and designated for international flights.
Victor,
Are you saying Power(SDU) at 18.24 was restored automatically? Timeline: Take off at 16.41, at 17.07 ACARS Failure, after 17.19 (last ATC), Communication Failure, after 17.22 Turn Back to KLIA, Navigation Failure, at 18.03 SDU Failure, and 7h38 after take off at 00.19 Fuel Exhaustion or Aviation Failure in the SIO. A lot of Failures. With all these Electrical Failures, the Malacca Strait diversion is not to be believed, nor the impossible fluctuating altitudes indicated in the previous and the more recent reports, nor a diversion of THREE hours until 19.41 after take off. MH370 was only 38 minutes into its flight and could have made it back to KLIA in near enough the same time after dumping fuel. So what stopped them from achieving that? Catastrophic depressurization at 35000 feet leading to the Ghost Flight is my reckoning. 2-3 Minutes only if O2 didn’t or wasn’t deployed successfully.
Victor and Don,
Can you please clarify which squawk code was assigned to MH370. 2195 (which falls outside the range assigned to Malaysia’s International Flights), or 2157 (which does not), or are we talking two different aircraft?
Not relevant but interesting:-
https://nakedsecurity.sophos.com/2018/08/13/in-flight-satellite-comms-vulnerable-to-remote-attack-researcher-finds/
@Air Life: I am not aware of code 2195, which is not even an octal number.
No, I didn’t mean to imply that the power to the left bus was restored automatically. Likely, it was from manually resetting one or more switches in the overhead panel.
@DrB asked: Do you know, or have you inferred from the ATSB data, how long the engine run time difference was assumed to be in the simulations using the alternate power configuration? I don’t see how less than 2 minutes is possible.
Looking at Case 6, the right engine fails at 119 s based on the drop in speed and the start of the descent starts seconds later. The maximum bank to the right (afterwards a bank to the left starts) is about 10 deg, and occurs at 162 s, which appears to be due to the loss of left thrust. That greatly compresses the time between loss or right thrust and loss of left thrust.
When I first analyzed this case, I believed the loss of left thrust was the autothrottle reducing thrust during the descent as the speed increased passed 250 KIAS, which is approximately the starting speed. This scenario had the left engine running but at idle thrust for the remainder of the flight. However, Andrew explained that with no A/C power, the servo motors on the autothrottle are not powered, so I can’t explain the loss of left thrust other than a flame out, if the loss of left thrust did indeed cause the bank to the left.
Are we sure that there is no other power source for the autothrottle servo motors? If the autothrottle somehow stayed engaged, it would explain a lot.
I think it will take more queries to the ATSB to determine the exact sequence of events. On the other hand, in every sequence with a steep descent (>15,000 fpm), there was an impact soon after. Long glides with level wings and oscillating high descent rates don’t occur.
It’s also possible that the steep descent within 2 minutes of loss of autopilot is an indication the there were pilot inputs.
@DrB: Here’s another possibility. When the servo motors to the autothrottle lose power, the autothrottle does not disengage. When the APU comes online, power is restored to the servo motors, and the functionality of the autothrottle is restored. In this scenario, the left engine could be at idle thrust for many minutes after the right engine flameout.
@Victor, Air Life, Bruce Robertson
My apologies, I erroneously quoted a transponder code 2195 in an earlier comment. That was a simple error.
According to Malaysia’s Factual Information the ‘Lumpur Delivery’ ATC unit assigned transponder code 2157 at 16:26:41UTC, the various SSR logs are consistent in showing octal code 2157.
@Andrew
DF 17 defines the Downlink Format for ADS-B extended squitter, which contains the EM, Extended Squitter Message. Five message types were originally defined for the EM: 05/Airborne Pos, 06/Surface Pos, 08/Aircraft ID, 09/Airborne Vel, and 0A/the event driven message. A sixth was defined in DO-260A, EM 1D/Target State and Status.
I agree that the Malaysian 777 fleet would have been compliant with the European requirements, so I expect that it was broadcasting the extended squitters listed above while its transponder remained active. If VATM recorded all received DF17 EMs, that log may have contained useful information, but only if it was accessed and reviewed.
@Don Thompson said: I agree that the Malaysian 777 fleet would have been compliant with the European requirements, so I expect that it was broadcasting the extended squitters listed above while its transponder remained active.
If there were no extended squitters, we would have no ADS-B data from FR24, right?
@DrB, @Andrew: According to the FCOM:
Autothrottle disconnect occurs automatically:
* if a fault in the active autothrottle is detected
* when either reverse thrust lever is raised to reverse idle
* if the thrust levers are overridden during a manual landing, after the autothrottle has begun to retard the thrust levers to idle
* when both engines are shutdown
If the loss of power to the autothrottle servo motors is not a “detected fault”, then the autothrottle remains connected, and the functionality will be restored when the APU comes on line.
@Victor,
ADS-B comprises extended squitter messages (EMs). If a transponder wasn’t broadcasting ADS-B then FR24’s log would contain no records for that transponder.
FR24, and the other public web trackers, don’t require all the EMs to identify and plot the track of an aircraft. It’s not clear if the plethora of available feeder units and software relay all received EMs to the web trackers.
I’d expect an ANSP’s receiver station, or network, to maintain a comprehensive log of EMs received.
@Don Thompson: Yes, I agree. We are trying to use FR24 in a way it was not intended to be used, and some are trying to draw definitive conclusions on ambiguous data. (The low precision of the latitudes and longitudes are a tell-tale sign that the data was processed, as the lower precision values are not consistent with the Compact Position Reporting format used by ADS-B.) Access to the raw data received by the Con Son receiver would be helpful. My guess is that data will be difficult to obtain, and I also question whether there was any data after 17:20:31.
@Victor
“…Here’s another possibility…the left engine could be at idle thrust for many minutes after the right engine flameout.”
Interesting, yes that’s similar to what I am worried about. I have guesstimated with FS9 that maybe 250-nm glide is possible with 2 engines on idle. This compares to your quote of 125-nm with no engine power.
@TBill: A glide of 250 NM with no thrust is impossible, if that’s what you are implying. The (L/D) of the wings is not sufficiently high to permit such a shallow flight path angle.
@Victor
I was saying idle engine (min. fuel flow) thrust is 250-nm…are you saying that idle engine is zero contribution to distance?
@TBill: I believe at idle engine speed, the drag introduced by the engine is balanced by the thrust produced by the engine, so the net force is zero. If the engine is shutdown, there is additional drag due to the windmilling of the engine rotors.
@Victor
…so could I go farther with 2 egines in idle in the 70-knot winds at 30 South? I get I can maybe reach Dordrecht Hole from Arc7, but it’s FS9 so I am not claiming it is reality (yet).
@TBill: FS9 should do fine predicting the glide path with idle engines and a wind. But you’ll have to accurately model the wind field. At 30S,98E, the wind is towards the east. It’s about 70 knots at FL340, but it’s down to around 30 knots at FL210 and around 12 knots at FL100.
@Don Thompson
RE: ”DF 17 defines the Downlink Format for ADS-B extended squitter, which contains the EM, Extended Squitter Message. Five message types were originally defined for the EM: 05/Airborne Pos, 06/Surface Pos, 08/Aircraft ID, 09/Airborne Vel, and 0A/the event driven message. A sixth was defined in DO-260A, EM 1D/Target State and Status.”
Thanks, that’s what I thought. I wasn’t sure which parameters you meant when you mentioned ADD in relation to ‘enhanced surveillance requirements’,
@Victor
@TBill
RE: ”If the loss of power to the autothrottle servo motors is not a “detected fault”, then the autothrottle remains connected, and the functionality will be restored when the APU comes on line.”
The TMCS monitors the ASM response to commanded thrust changes. A fault is detected if the ASM fails to respond (eg due to a power loss), causing the autothrottle to disconnect. The autothrottle function would need to be reset after power restoration, otherwise it would remain disconnected.
@Andrew: After the right engine fails, there could be a period of time in which the commanded left engine thrust is full, and the thrust lever is already in the full position. Perhaps in this case, there would be no detected error in commanded thrust, and the autothrottle would remain connected. If the APU comes online before the commanded thrust is less than full, perhaps the autothrottle’s functionality returns before it is ever disconnected.
@Victor Iannello,
1. Thank you for responding to my question about the time delay. It seems that at the moment we don’t know exactly what was assumed for the time difference in engine fuel exhaustions. This time difference is important.
2. Using the X-Y data, here are what I figured as the initial conditions for the 10 cases:
Case# Alt(ft) Horiz.GS(kts) True Track (35S)(deg) Elect. Config
1 40,000 450 179.5 Normal
2 40,000 450 179.5 Normal
3 40,000 450 179.5 Alternate
4 40,000 450 179.5 Alternate
5 40,000 450 174.1 Normal
6 35,000 425 174.2 Alternate
7 40,000 455 184.8 Normal
8 35,000 407 174.1 Normal
9 40,000 450 179.5 Normal
10 40,000 450 179.5 Alternate
3. The true track angles I derived using an approximate method can be compared to the values you said were given by ATSB: “. . . the initial headings were either 178°, 184° or 190°. . .” The differences will be due to the significant westerly cross winds, since the ATSB quotes headings (I assume true, not magnetic), whereas I computed true tracks. Assuming a 35S approximate latitude, I get roughly 180, 174, and 185 degrees for the true track values at the start of the X-Y data. That implies that the headings from ATSB were about 4-5 degrees larger than the true tracks I calculated, and this is consistent with ~32 knot right crosswinds, which appears reasonable for that general location and time.
4. From the table above, it appears that Cases 1-4,9,10 are identical (at 40,000 ft, 450 kts, 180 deg true track) except for perhaps two parameters. Cases 1,2,9 used normal electrical and 3,4,10 used alternate electrical. I suspect there were 3 values of rudder mis-trim used for each of the two electrical configurations. Do you think this was the case?
5. The Case #6 example indicates that the left engine was assumed to stop only about 40 seconds after the right engine was fuel exhausted (causing the beginning of a turn to the left and a phugoidal oscillation with generally higher ground speed). It appears this timing was assumed to try to match the RODs inferred from the 00:19 BFOs. However, it does not quite succeed in matching the BFOs at the proper time, being about a minute late. In addition, there is no plausible rationale, in my opinion, for the MEFE time difference to be less than one minute. It is much more likely to be about 8-10 minutes. To get down to one minute, approximately 200 kg of fuel would need to be transferred from the left tank to the right tank prior to fuel exhaustion. I don’t think that a pilot would make such a minor fuel transfer, so I don’t think that is likely to have happened. If a fuel transfer didn’t occur, then I don’t see how the assumed < 1 minute fuel exhaustion time difference could have happened.
6. On the other hand, if the L-R fuel exhaustion time difference is really 8-10 minutes, then it does not seem likely that the BFOs can also be matched at 00:19. So, based on these few simulations, it does not seem possible to match both the BFOs and the expected fuel exhaustion time difference. That may mean that the alternate electrical configuration simply cannot explain the data we have.
7. Looking at Case #5, the only normal electrical configuration that achieves 15,000 fpm ROD but (at a very late 1810 seconds after simulation start), the ground speed begins to decline at 120 seconds. I assume this corresponds to the right engine running out of fuel two minutes into the simulation. I'm not sure when the left engine stops thrusting. A phugoid oscillation appears at about 854 seconds, and a turn to the left begins near 867 seconds. These two events are about 12 minutes after R engine failure. Is that when the L engine is assumed to stop – 12 minutes after R engine failure?
8. It seems that none of the simulations is consistent with the BFOs at 00:19 (using the ATSB's APU/SDU power-up timeline), regardless of electrical configuration. Furthermore, perhaps the agreement might be improved by adjustment of the rudder mis-trim, but this is not obvious to me.
9. I can't think of any way the BFOs could be very far off, but either that happened or the 10 simulations run by Boeing did not include the correct fuel/electrical configuration, or a pilot flew part of the final descent.
@Victor
“At 30S,98E, the wind is towards the east. It’s about 70 knots at FL340, but it’s down to around 30 knots at FL210 and around 12 knots at FL100.”
I know. In one active pilot scenario, I think descent is favored to get out of the winds and sun, on the other hand, another scenario maybe the high winds carry MH370 almost all the way to Arc6 by the 23:14 sat call.
@DrB: Regarding Case 6, perhaps you missed my previous comments regarding idle thrust of the left engine with the autothrottle connected, which could explain the timing of the bank to the left occurring well before fuel exhaustion of the left engine.
If you calculate the air speeds, you’ll see the reversal of bank from right to left occurs suspiciously close to when the calibrated airspeed increases past the starting value (around 250 KIAS). I am still exploring whether the autothrottle was connected at this point.
Considering variability of initial conditions, APU start up times, and SATCOM log-on times, I think the timing of the high speed descent could match the BFO data.
I don’t what rudder trim was assumed for any of the cases, although in the past I have proposed that different values of trim caused different behavior for the “alternate configuration” cases.
Timing of Case 5 doesn’t seem close to what was observed. I think the “alternate” electrical configuration is still the best candidate for uncontrolled flight.
Of course, pilot inputs cannot be dismissed.
@victor,
I think we need to consider a slight variation of the electrical ‘alternate’ scenario. One where the right IDG is powering the Both main buses, however the L BU GEN is working and is able to power a Transfer bus when the first eng fails.
In this set up, when the right eng fails, the autopilot should remain engaged and only fails when the left eng fails.
@Mick Gilbert. As to how the ATSB (and others) came to sign off on the Malaysian report as accredited representative, here is an ATSB quote from today’s ‘The Australian’ about its role in the investigation of an accident by South Africa involving two Australian pilots.
‘The Australian Transport Safety Bureau has provided an “accredited representative” to assist the SACAA with its investigation, and help obtain information requested by the South African investigators from involved parties in Australia. “It is not the role of ATSB to review any reports of this investigation,” an ATSB spokesman said.’
I wonder still if the ICAO could require more work on the Malaysian report.
@Victor
RE: “After the right engine fails, there could be a period of time in which the commanded left engine thrust is full, and the thrust lever is already in the full position. Perhaps in this case, there would be no detected error in commanded thrust, and the autothrottle would remain connected.”
Perhaps, but as I mentioned in a previous comment, I don’t think the thrust lever would have reached the max thrust position. The autothrottle would have been engaged in SPD mode and the speed would have needed to start decreasing before the autothrottle commanded any thrust increase. However, in this scenario, the electrical power would have failed after about 7-9 seconds, removing power from the ASMs. I doubt the ASMs would have had time to move the thrust levers to the max thrust position before the power failed.
@Andrew: Thanks. At around 123 s, the speed decreases and the plane begins slowly rolling to the right. At around 160 s, the bank reaches about 10 deg and starts rolling to the left, is level at around 194 s, and continues to steadily roll left. If the roll to the left is not from a reduction of left thrust with constant rudder trim, I am not sure what caused it.
@David
Thanks for that reference, David. My reading of Annex 13, specifically 5.23, 5.25 and 6.3 as they relate to Accredited Representatives is that unless you are an AR from the State of Registry, the State of the Operator, the State of Design or the State of Manufacture you don’t get a say in the reporting. It’s look but don’t touch, speak only when spoken to. So all this hot air from the arm-waving, pot-bangers (Bailey, Keane et al) about the ATSB drafting a dissenting report on MH370 is just that, hot air. The only Non-Malaysian AR entitled to submit a dissenting report on MH370 would be the NTSB.
Given the nature of international organisations like ICAO I suspect that they would be reluctant to start ‘marking’ investigations and calling for ‘resubmits’. That’s not to say I don’t think that the final report needs more work, it does, in its current form it’s far more confusing than anything that went before it (eg the Factual Information).
@Victor
I haven’t yet caught up on the commentary over the last few days, so perhaps I missed something. At what point is the left engine assumed to have failed?
@Mick Gilbert. That clarifies it.
So the NTSB, inc Boeing and all, must have been satisfied with the report.
@Mick Gilbert
Para 6.3 of Annex 13 was substantially reworded in 11th edition, issued in 2016. It now reads (amended parts in bold):
6.3 The State conducting the investigation shall send a copy of the draft Final Report to the following States inviting their significant and substantiated comments on the report as soon as possible:
a) the State that instituted the investigation;
b) the State of Registry;
c) the State of the Operator;
d) the State of Design;
e) the State of Manufacture; and
f) any State that participated in the investigation as per Chapter 5.
If the State conducting the investigation receives comments within sixty days of the date of the transmittal letter, it shall either amend the draft Final Report to include the substance of the comments received or, if desired by the State that provided comments, append the comments to the Final Report. If the State conducting the investigation receives no comments within sixty days of the date of the first transmittal letter, it shall issue the Final Report in accordance with 6.4, unless an extension of that period has been agreed by the States concerned.
@David
The UK’s AAIB probably also had sign-off rights on the basis of the RR Trents.
@Andrew
Thanks for that. I’m wondering if that change would apply to an investigation commenced under the old framework?
@Mick Gilbert
I don’t know, but given the significant assistance Australia provided during the search and investigation, I find it difficult to believe that a significant ‘protest’ would simply be ignored. Perhaps I’m being naive.
@Victor Iannello,
@Andrew,
Combining suggestions to create a more detailed scenario for Case #6, let’s assume the right engine failed at 120 seconds elapsed time in the simulation. The left engine is still running at normal cruise thrust, with the auto-throttle servo being controlled in SPD mode. Then, within a few seconds, the aircraft begins to slow down (at 123 seconds), and more thrust is desired from the left engine, but there is only a small increase ( or perhaps even no adjustment at all) before the AC power is lost to the servo motor at about 8 seconds after right engine fuel exhaustion (I.e., circa 128 seconds). Thus 128 seconds in the simulation corresponds to circa 00:17:30 UTC. Now the aircraft has roughly cruise thrust from the left engine and no thrust from the right engine. The TAC will begin to adjust the rudder to the left during the interval from 120-128 seconds. We don’t know if the rudder adjustment will be completed within this time period. Perhaps the rudder adjustment was still in progress when the AC power was lost circa 128 seconds, and the TAC is only partial. If partial, then the aircraft will begin to turn to the right at 123 seconds, and this will continue through 128 seconds and afterward until there is a reduction in left engine thrust (or impact occurs). The TAC in the rudder is now fixed, and it won’t change after the APU starts.
When the AC power is lost at 128 seconds, the PFCS reverts to secondary mode. Does anyone know if secondary mode commands a particular thrust level, such as idle? For this exercise, I will assume that is the case. Next, let’s assume the APU starts up and delivers AC power in 60 seconds after loss of both Transfer Buses, which occurs at 128 seconds (the same time as loss of the Main Buses). So the APU delivers AC power beginning circa 188 seconds. If the APU power can cause a reduction in the left engine throttle setting, say down to idle, then the aircraft would stop turning to the right and begin turning left at 188 seconds. This is what the simulator Case #6 shows, except near 160 seconds, not at 188 seconds. However, I don’t know if an in-flight APU start can occur in only 32 seconds, which is needed to match the left turn beginning at 160 seconds in Case #6. I think it takes about 30 seconds for the inlet door to open. ATSB has used 60 seconds for the APU auto-start.
Continuing and using 60 seconds for the APU start time, the APU would start circa 188 seconds (00:18:30) and supply power to the SDU and to the auto-throttle servo motors. IF the servo motors did not exhibit a fault during the power outage, they may still be functional. In addition, IF secondary mode commands idle thrust, then the left engine may have been reduced to idle thrust at 00:18:30 (188 s). That would cause the right turn to stop (perhaps at 20 degrees or so of right bank) and a left turn to begin, because the TAC set the rudder for near-cruise thrust from the left engine, and the reduced left engine thrust would start a left turn.
That turn/bank to the left would increase with time, passing through wings-level and then banking to the left, causing an increasing ROD about a minute later at 247 seconds (00:10:29), when the SDU transmitted a log-on request after booting up, followed by the log-on acknowledge message 8 seconds later. The two transmissions created the two final BFOs at high RODs.
This scenario has two main features. First, it uses the alternate electrical configuration that causes loss of all AC Buses and APU auto-start upon right engine fuel exhaustion. Second, it has a means (still unproven) for the left engine to have a reduced thrust upon APU power-up. The left engine, in this case, would continue running until impact, perhaps some few minutes later, unless a pilot took control in the interim. This scenario seems to offer the possibility to match the BFOs at the correct time. However, the possible means to produce the reduced left engine thrust by servo control upon APU power-up rather than by fuel exhaustion are not well understood and require further investigation to determine their feasibility.
@Andrew, Mick Gilbert. What the ATSB was quoted as saying re the South African accident looks inconsistent with the 2016 amendment.
There is a misunderstanding somewhere.
To add to the muddle, in the 1997 Silk Air investigation the ATSB made comment but that was not accepted by the Indonesians on the grounds they were not accredited, as distinct from not being one of the then Annex 13 nominees.
To add to that the Annex 13 amendment and probably that which went before it are concerned with ‘Final’ reports, which the Malaysian’s wasn’t (to my regret, though not the relatives of those lost).
I have asked the ATSB whether they were free to comment formally on the Malaysian Safety Investigation report.
@DrB
RE: “When the AC power is lost at 128 seconds, the PFCS reverts to secondary mode. Does anyone know if secondary mode commands a particular thrust level, such as idle?”
PFCS reversion to secondary mode does not cause the thrust to reduce to idle. The autothrottle is controlled by the thrust management computing function (TMCF) within AIMS and is entirely separate to the PFCS.
@David
I wonder if the ATSB was quoted correctly in the report. As an accredited representative, the ATSB has the right, under Annex 13 and the South African Civil Aviation Act, to comment on the draft final report.
In the Silk Air case, I think BASI (as the ATSB was then called) was an ‘advisor’ not an accredited representative. As such, they had no Annex 13 entitlement to comment on the draft final report.
I’m not sure what you mean by your second last paragraph. Although the contents are disappointing (to put it mildly), the MH370 ‘Safety Investigation Report’ is the Annex 13 ‘Final Report’. The accredited representatives to the investigation should have had the right to comment, under the current and previous amendments to Annex 13. Para 6.3 in the earlier amendment read (my bolding):
“6.3 The State conducting the investigation shall send a copy of the draft Final Report to the State that instituted the investigation and to all States that participated in the investigation, inviting their significant and substantiated comments on the report as soon as possible…”
Victor and Don,
Thank you for your replies. My point is in order to even discuss end of flight simulations, there has to be clarity on which course MH370 took, i.e. a right turn back to KLIA after IGARI or a left turn after IGARI and diversion across Malaysia and up the Malacca Strait to the NW. The two won’t intersect on the 7th arc. There is reasonable doubt as to the left diversion. The fluctuating flight levels in the latest report would indicate there are two aircraft involved in the military radar. One at a higher altitude (possibly military) and the other at a lower altitude(possibly a departing flight to the west or North Africa). So far all the searches based on the left diversion and 7th arc have come up empty. Isn’t it time to re-evaluate the military radar data, which I believe has never been scrutinized by an independent body or radar specialists, apart from re-evaluating Inmarsat’s BTO/BFO values and Timeline. I’m more inclined to only work with the Timeline as I belief it to be more accurate and the BTO/BFO values to be superfluous in the right turn back to KLIA case. Ater 4 and a half years, isn’t it time to look the other way? The way MH370 turned first, right, before it’s course was turned left. Or, are we going to continue to beat a dead, left diversion horse to ad infinitum? Nobody is errorless and that includes Inmarsat and the the Military, but turning a blind eye to a feasible option, in the absence of conclusive proof, is unforgiveable.
@Air Life,
I would not categorise the case for the left turning diversion as subject to ‘reasonable doubt’, nor a ‘dead horse’. The accuracy of the path, as a 4D route, is questionable but 3 evidence, in both illustration and written narrative, is that three separate primary surveillance radars recorded 9M-MRO’s path after diversion. It would be useful to see the machine recorded data from which the track depictions were created.
Those three assets include DCA’s Kota Bharu-WMKC terminal area control radar, the RMAF ATC unit’s terminal area control radar at WMKB, and (at least) the RMAF Sqdn 310 air defence surveillance radar high on Western Hill, Penang.
It’s been noted to me that a right turn at IGARI, and a diversion path back towards WMKK would have put 9M-MRO into the view of Singapore’s Air Defence Radar at Bukit Gombak. That system is operated 24/7.
The ‘fluctuating flight levels’ are not untypical of long range 3D primary surveillance radars such as that sited at Western Hill.
But, please, work away on your exploration of a right diversionary turn and establishing at least as much verification as exists for the left turn.
@Andrew, “I wonder if the ATSB was quoted correctly in the report.”
Maybe not. I hope their response will answer that: I have quoted it.
“…the MH370 ‘Safety Investigation Report’ is the Annex 13 ‘Final Report’”
My opinion was based on the their statement to the families and relatives that it wasn’t, to their reported relief. Is a Safety Investigation synonomous with the Final? I note the Annex 13 terminology.
Thank you for clarifying what the pre-amendment version said as I understood the 11th amendment to alter the sense of who could comment.
Annex 13 reports.
Comments may or may not be submitted. A comment might state, “great work on use of spell check”. Comments may not be accepted, refer to NTSB’s work on MS990.
The investigating body is not bound to publish comments, this one hasn’t.
Yesterday, I reviewed contemporary reports for the establishment of the investigation team.
In response to the event, the MY government’s cabinet set up three committees: technical, next of kin liaison, asset deployment.
The technical committee was responsible for establishing the Annex 13 investigation team.
On 5th April 2014, the remit of the Annex 13 investigation was announced. The remit?! Someone obviously didn’t bother to read Annex 13 or related Doc 9756/Manual before making that line up.
On 29th April came the announcement that an IIC had been appointed. Perhaps the first assignment for the investigation team was attendance on air accident investigator training.
The reports of the Investigating Team have never produced anything of substance. That the contributions of other bodies are provided as appendices at least differentiates them from the main body of the report.
A fundamental weakness of Annex 13 is that each contracting state has the option of undertaking an investigation, regardless of the capabilities of that state’s administration. Annex 13 mentions ‘regional accident and incident investigation organisations’ (RAIO), that’s a good suggestion.
@David
RE: “My opinion was based on the their statement to the families and relatives that it wasn’t…”
I missed that statement. Do you have a link? Annex 13 refers to the final report as the ‘Final Report’, but I don’t think there’s any onus on States to name it that way. In the UK, USA, Hong Kong, South Korea and Greece the final report is called an ‘Aircraft Accident Report; in Taiwan it’s an ‘Aviation Occurrence Report’; in Japan it’s an ‘Aircraft Accident Investigation Report’; and in Canada it’s an ‘Aviation Investigation Report’. In the MH370 3rd Interim Statement, the Malaysians referred to a ‘Final Report’ being released following the suspension of the underwater search. However, the 4th Interim Statement referred to it as a ‘Safety Report’. I wonder if the name changed because of the inconclusive nature of the report and the sensitivities of the families, etc.
RE: “Thank you for clarifying what the pre-amendment version said as I understood the 11th amendment to alter the sense of who could comment.”
I think the 11th amendment merely clarifies who can comment. The first part of the earlier version of para 6.3 says that ‘all States that participated in the investigation’ can comment, but then refers only to the States of Registry, Operator, Design and Manufacture. My understanding, however, is that the intent has always been that any State that has an accredited representative in an investigation is entitled to comment.
@David, Andrew
Manual of Aircraft Accident and Incident Investigation, Part IV, Reporting:
“The Final Report of an aircraft accident investigation is the foundation for initiating the safety actions which are necessary to prevent further accidents from similar causes. Therefore, the Final Report on an accident must establish in detail what happened, how it happened and why it happened. The findings, causes and/or contributing factors of the Final Report should lead to safety recommendations so that appropriate preventive measures can be taken.“
@Don Thompson
RE: “The investigating body is not bound to publish comments, this one hasn’t.”
I think that’s a weakness of the ICAO system and, more generally, the UN. Annex 13 requires the investigating authority to ‘either amend the draft Final Report to include the substance of the comments received, or if desired by the State that provided comments, append the comments to the Final Report’. Nevertheless, some States get away with not playing by the ‘rules’.
RE: “Manual of Aircraft Accident and Incident Investigation, Part IV, Reporting”
Clearly, the MH370 Safety Investigation Report fails to meet that standard.
@Andrew said: I wonder if the name changed because of the inconclusive nature of the report and the sensitivities of the families, etc.
Without a doubt. There was uproar when the report was called “final”.
@Victor
Believe it or not, the boxed EICAS message in the following image says ‘AUTOTHROTTLE DISC’:
https://www.dropbox.com/s/jr0bd2sg6vdifwb/AUTOTHROTTLE%20DISC.png?dl=0
The image is a screen shot from one of ALSM’s simulator trial videos. In that video, the right engine failed first and the autothrottle advanced the left thrust lever to CLB thrust. The left engine subsequently failed and the AUTOTHROTTLE DISC message appeared when electrical power was lost. The message remained after the APU started and restored power.
Don, thank you for your reply. I have been working on the left and right turns for four and a half years, but have much supporting material in my arsenal for the right turn, shared with officials, too. As MH370’s transponder was non-operational and not expected to divert or even to show up in certain locals in the early hours of the morning on the 8th March 2014, I am of the opinion MH370 could have been missed or even misinterpreted as something else. As such, the right turn remains a very strong possibility and one I can’t ignore or exclude.
@David
“So the NTSB, inc Boeing and all, must have been satisfied with the report.”
In that case I may have a bone to pick with NTSB. In general though I am expecting a defensive approach, giving Malaysia freedom and sovereign rights to say what they want as long as they do not unfairly blame USA/Boeing for, for exmaple, unsubstantiated mechanical failure etc.
@Andrew: I looked at those simulations, and I watched whether the A/T button light was illuminated. Unfortunately, in those cases, the transition to secondary mode occurs after the failure of the second engine. We have no doubt the A/T is disconnected in this case, as those simulations show.
The “alternate” electrical configuration is different. After the failure of the right engine, the A/T would be initially commanding full thrust. I am still questioning whether the APU could have come online before there was a disagreement between the commanded thrust and the thrust lever position that remained after the ASMs were unpowered.
@DrB: As Andrew said, there is no reason for either the APU start up or a transition to secondary mode to cause the thrust to reduce.
@Andrew, @DrB: Here’s a plot showing the relationship between groundspeed, altitude, and estimated bank angle. We can attribute the initial roll to the right as partial compensation for the thrust asymmetry after the right engine flames out. The roll to the left starting at 160 s is harder to explain. The only real possibilities are a reduction of left thrust (which I attribute to the autothrottle reducing the left thrust when the speed increases), or a change in position of one or more flight control surfaces (which I can’t explain).
@Air Life,
I don’t doubt you hold the opinion that a right turn may have occurred, if any officials have shared anything back to corroborate your opinion that’d be very interesting indeed.
The traffic count through the KL FIR was not high on the night 7/8th March. A combined picture of SSR identified plots vs PSR unidentified plots would have readily differentiated 9M-MRO’s post diversion track. The Malaysian reports of 8th March 2015 and 30th July 2018 describe that they made a conclusion for the post-diversion track, using radar playbacks, that 9M-MRO turned left.
To turn right, not be detected by any PSR, and make the line of position described by the 18:25 Log On, implies that the aircraft would have actively evaded PSRs. Those PSRs would include DCA TARs at WMKC, WMSA, WMKK, WMKJ, and WMKL; the RMAF operated TARs at WMKB and WMKD; and the long-range air defence radars at Western Hill-Penang, Kuantan-Phanang, Bukit Lunchu-Johor, and Bukit Gombak-Singapore.
@Andrew,
@Victor Iannello,
Andrew’s picture shows the autothrottle disconnecting upon loss of AC power and not reconnecting upon APU start. Nice find! For Case #6 this means the left engine thrust should remain somewhere between cruise and climb after loss of power due to right engine fuel exhaustion. That seems to eliminate reduced left engine thrust as the cause of the left bank starting about 40 seconds after the right engine fails. If not that, what else could be the cause? We must be missing something, since the Boeing Level-D simulation clearly shows it happening. Could the rudder move then? Could the TAC be reset?
Update on Aug 21, 2018
I have added some comments to the discussion section (in italics) to improve the clarity. Unlike what has been reported, I never intended this post to be interpreted as a recommendation for a new search area.
@DrB: I am still questioning whether the A/T ever disconnected before the APU started. If the autothrottle is commanding full left thrust, and the thrust position was at the full position at the time that power was lost to the ASMs, perhaps there was no discrepancy between commanded and actual thrust, the TMCS did not detect a fault, and there is no A/T disconnect. When the APU starts supplying AC power, the A/T remains connected.
@DrB, @Andrew: Thinking about this more, the roll rate to the left is nearly constant from 160 s to 290 s. Yet, the speed and altitude are nearly constant from 210 s to 250 s. The constant speed and altitude indicates the left engine is producing thrust. If the left thrust was varying from a low value at 160 s to a high value at 210 s, the roll rate would also change.
So, I am now leaning towards the left engine producing constant thrust with no autothrottle, and a change in position of one or more control surfaces at 160 s.
The change in control surface position could be related to either the APU or the RAT coming online.
@Victor
Many thanks once again for an enlightening post. I have waited until I had completed a first analysis of all the data sets before commenting. In our BFO analysis we concluded that the ROC was around -4,000 fpm at 00:19:29 UTC and around -15,000 fpm at 00:19:37 UTC, only 8 seconds later. Many said that was impossible!
My key take away is that Case 3, 4, 5, 6 and 10 all have a phase were the ROC goes from around -4,000 fpm to around -15,000 fpm in around 8 seconds and this happens on average 27 seconds before impact.
Case 3 at 541 secs, ROC is -3,480 fpm and at 552 secs, ROC is -15,780 fpm in 11 seconds, TTG 25 seconds.
Case 4 at 410 secs, ROC is -4,080 fpm and at 418 secs, ROC is -16,380 fpm in 8 seconds, TTG 35 seconds.
Case 5 at 1868 secs, ROC is -4,260 fpm and at 1876 secs, ROC is -15,960 fpm in 8 seconds, TTG 15 seconds.
Case 6 at 349 secs, ROC is -3,360 fpm and at 356 secs, ROC is -17,100 fpm in 7 seconds, TTG 28 seconds.
Case 10 at 436 secs, ROC is -3,660 fpm and at 443 secs, ROC is -17,340 fpm in 7 seconds, TTG 30 seconds.
The gradual descent shown by Case 1, 2, 5, 7, 8 and 9 only exhibits such decent rates in Case 5 were the turbulence is medium, the altitude and speed lower. Case 1, 2, 7, 8 and 9 do not exhibit descent rates over 10,000 fpm.
As you point out the alternate configuration Cases 3, 4, 6 and 10 are indeed fascinating for their fit to the BFO data.
Case 6 is also an exception amongst all the cases, in that the Ground Speed never drops below 95% of the start speed.
Boeing in their analysis in the Malaysian Safety Report Appendix 1.6E on page 1294 in Table 4 show 22 options of different altitude and speed combinations, that they considered, from low and slow (FL030, 235 knots) to high and fast (FL400, 494 knots). The MH370 end points considered range from 10°S to 40°S.
It is interesting that the ATSB only selected 3 altitude and speed combinations and all were high and fast. None of the options correspond to either LRC or MRC modes. All the ATSB selections reach fuel exhaustion between 36.1°S and 37.3°S.
Best regards, I allow myself these contributions:
ANNEX 13: The State in charge of the investigation, always has the last word in the content of the report, even if a State with an Accredited Representative forms part of the process and recommendations, not necessarily these will be accepted or should be accepted by the State responsible for the investigation. process. Everything will depend on what the State in charge determined to do according to Article 37 or 38 of the 1944 Chicago Convention, with what is stated in Annex 13 of ICAO.
Remember that a Final Report can be Issued, if we consider the ICAO Accident Concept in point or separate “C” that indicates that the Aircraft is inaccessible or is missing. In the case of the MH370 some remains were located so the research process is guided by the DOC. 9756 and it is therefore possible to issue a Final Report that does not have to reach conclusions. But attention can also be opened again the Investigation (reopening) according to Chapter 5 of Annex 13, so that report is not the last word.
Search and Rescue (SAR): In search operations, the tracking area is estimated based on the worst scenario. In this case it is that the one that had the control of the aircraft maintained it until the aircraft fell to the sea.
Understanding the 7th arc as a point of reference, it is necessary to estimate that the aircraft deliberately planned and calculated the maximum distance from it that was already referred by Mr. Iannello.
Of course what we are trying to do now is reduce the search area, but it turns out that the maximum search area was not applied and I think that’s what Mr. Iannello refers to, since it is clear that we do not know the MH370 count, if more to the north of the 7th arc or more to the south. But there is also a problem the tracking system with the AUV was with inertial navigator and this generates an interesting margin of 650 meters for each hour of navigation, on the other hand we used a sonar SSS (Side Sweep) and no opening SAS synthetic that today is much more efficient, among other sensors, therefore the area sought by Ocean Infinity can not be discarded, there is necessarily a need to reevaluate it and look again there and in the remaining others if necessary with new technology, it is complex very difficult area.
However, to determine how much the aircraft could fly, it is necessary to be clear about the amount of fuel onboard the 9M-MRO (MH370), although the different reports talk about it, there is a doubt about the amount of fuel remaining on the last flight of that 9M-MRO aircraft, because for its destination to China it did not need to be fully filled and this information is not clear in the different reports (amount of fuel remaining before filling it for its flight to China).
An important fact is to determine with precision the turning point of the MH370 in the area of the Gulf of Malacca, because there is no doubt that the control course had the SUR course (we suppose, but it may not be that way). In the final report, this point is not clearly defined and is a limitation in the search operations. A colleague of this forum, it was clear you have to review ALL the radar information and I say you’re absolutely right, but you have to check everything, including cell phones, etc. For example, I ask: Did the other satellite to the east copy the hand grips of the ACARS system? This is not clear, the secret and the reservation, do not allow it and it is not good.
It is also necessary to review the simulated flights carried out by the pilot in command of the MH370 at his home, because this information will undoubtedly illustrate the possible direction that he took, because if he was responsible for the seizure of the 9M-MRO (we do not know , but we do not doubt it), in those flights there is information of interest. He will try to avoid the radars and more to the West (W) where there is a US military base and that must be kept in mind.
In Search and Rescue Operations (SAR), according to Annex 12 of the ICAO and the IAMSAR Manual (Doc. 9731), there SHOULD NOT BE SECRETS, nor Reservation of information. The comment by Mr. Iannello, about the secrets of ATSB of Australia, sadly is a real proof of why the MH370 does not appear and another real proof of that great failure to have secrets, is this forum that shows that there are many people in the world prepared and that can provide constructive and technical ideas. Of course there are also inventions, but they are few.
In short, more information is needed to have an estimated place of the accident, but what Mr. Iannello does today, is a positive contribution, to have another vision of what could happen, and that is VERY GOOD.
All contributions here are valuable, constructive and positive therefore the Secret in the handling of the MH370 information does so much damage, that’s why we did not find it.
My respects to all.
Enrique Martín
Organización Rescate Humboldt (ORH)
Caracas – Venezuela.
Email. sarrescate@gmail.com
Wsp:+58412209023
[Spanish version of this comment was deleted.]
@Victor
@DrB
RE: ”I am still questioning whether the APU could have come online before there was a disagreement between the commanded thrust and the thrust lever position that remained after the ASMs were unpowered.”
The APU auto-start sequence won’t commence until after electrical power is lost, some 7-9 seconds after the failure of the right engine. How could the APU come online before an autothrottle disagreement occurred? The point of my comment was to show that an AUTOTHROTTLE DISC message will be generated when the power fails, even if the thrust lever reaches the CLB thrust position. The message will not be removed when power is subsequently restored by the APU.
RE: ”The change in control surface position could be related to either the APU or the RAT coming online.”
It’s 01:49 am here and way past my bed-time. I’ll think about it some more later today.
@Richard Godfrey: Thank you for compiling the RoD data on all the flights.
Your comment reminded me of another aspect. If MH370 was in the “alternate” electrical configuration, then the log-on to the Inmarsat network occurred when there was still several more minutes of fuel in the left tank. In the past, we considered that the IFE log-on was missing because either the APU ran out of fuel or the plane had already impacted the ocean. For the “alternate” configuration, since there should have been enough APU fuel to last past the expected IFE log-on, it seems more likely that the plane had already impacted the water when the IFE log-on was expected.
@Victor
“Case 3, 4, 5, 6 and 10 all have a phase were the ROC goes from around -4,000 fpm to around -15,000 fpm in around 8 seconds and this happens on average 27 seconds before impact.”
The maximum lateral distance from the point of reaching -15,000 fpm to impact is 3.76 nm (Case 4).
The minimum lateral distance from the point of reaching -15,000 fpm to impact is 0.30 nm (Case 3).
@Andrew asked How could the APU come online before an autothrottle disagreement occurred?
I was proposing a situation in which the A/T is commanding full thrust and the thrust lever is already at the full position, such that there is no disagreement between target and actual thrust, both being “full”.
That said, I am leaning towards a change in control surface position, as I don’t think the speed, altitude, and roll rate are consistent with changes in thrust. Still thinking…
@Victor Iannello
How long does it take for the RAT to deploy? The RAT produces drag on the left side of the aircraft. Thus the RAT itself would induce a slight left turning of the aircraft. Suppose the left engine thrust went up at right engine flame-out, and the TAC compensated almost all of the thrust asymmetry. The result would be a slowly declining speed for a while and a slow right turn. Then when the RAT deploys, there would be a left turning action from the RAT drag. Maybe this was slightly larger than the effect of the TAC imbalance, and the result was a slow left turn. This scenario doesn’t require any restarting of the autothrottle or movements of flight control surfaces. Could it be that the drag caused by RAT deployment begins to have an effect at 40 seconds after right engine loss, or 32 seconds after loss of AC power?
@Victor,
When we previously worked through EOF scenarios the loss of actuator pressure and subsequent fairing effect of the L flaperon was discounted due to residual hydraulic pressure from the windmilling engines. Low speed windmilling would sustain some RPM in the Engine Driven Pumps (EDPs).
At the time I suggested that the hydraulic system configuration could be readily configured so as to rely on Main AC Bus electrical power alone. Primary L and R ENG pumps to off, Demand C1 and C2 Air pumps to off. Effectively, the RAT hydraulic configuration as landing gear & flaps are not expected to be deployed.
@DrB: RAT drag is interesting to consider. Years ago, I assumed a certain mechanical and electrical efficiency for the RAT power conversion, and knowing the true air speed and rated power of the RAT, the drag force can be determined. In turn, knowing the moment arm, the yaw moment can be calculated, and compared with the yaw moment from an engine. My recollection is the RAT moment was small compared to the engine moment. It would take some time to dig up those calculations. But as you say, even if small compared to the moment from the thrust asymmetry, the TAC compensation is also relatively large, so the net values might be small and comparable to the “RAT moment”.
On the other hand, as Richard G. observes, the Boeing simulations reaching high descent rates and high downward accelerations all impact the ocean close to the point where those values occur, in which case the details of the flight dynamics really don’t matter, and the plane should be found close to the 7th arc.
@Don: IF I understand you correctly, you are considering the case in which the hydraulic system is put in an alternate configuration. However, as far as we know, that was not done in the Boeing simulations.
There is also this blurb from the AMM that you found:
Training Information Point
When the RAT is extended and hydraulics off, the airplane rolls left. Two to three units of right control wheel rotation are necessary to hold the wings level.
In the end, I think we decided that the AMM was referring to a condition in which the hydraulics were selected off.
Victor – thanks once again for a highly illuminating post. I think it’s fair to say that these simulations lend credence to a flight terminus very close (horizontal distance) to the point of final power loss. The ATSB’s “bet” on a narrow search swathe looks entirely justifiable.
However, something serious doesn’t add up. Out of 6 conventional configuration simulations, only one produced ROD greater than 15,000fpm. The others produced maximum, transient ROD of -5000 to -7000. Even Case 5 which produced the higher ROD did so about 15 minutes after second flame out – 13 minutes too late to align with the BFOs. If that weren’t enough, the high ROD was only sustained for a period of 14 seconds, so you would have to be “lucky” to catch it on the BFO.
The “alternate” scenarios do succeed in producing ROD in excess of the -15000fpm required but again the timing is “wrong” – occurring 3 minutes to 5 minutes after first engine flameout & AC power loss. In any case, I thought the “alternate” electrical configuration was the purview of scarce “Accidentalists” like myself. Pijackers don’t need to shut down and isolate the entire left side power to achieve their devious intent, do they?
So it seems to me that these simulations have failed spectacularly to reproduce the ROD at the rate, acceleration and timing required for the 0019xx BFOs. The simulations might be wrong (initial conditions, fidelity or small sample), but it doesn’t change the fact that these simulations do NOT support conventional BFO interpretation.
CORRECTION I miss the word error of 650 meters per hour of navigation. Excuse me, please.
Of course what we are trying to do now is reduce the search area, but it turns out that the maximum search area was not applied and I think that’s what Mr. Iannello refers to, since it is clear that we do not know the MH370 count, if more to the North of the 7th arc or more to the South. But there is also a problem the tracking system with the AUV was with inertial navigator and this generates an interesting margin of error of 650 meters for each hour of navigation, on the other hand a sonar SSS (Scan Side Sonar) was used and not SAS synthetic opening that is much more efficient today, among other sensors, therefore the area sought by Ocean Infinity can not be discarded, there is necessarily to reevaluate and re-search there and in the remaining others if necessary with new technology, it is complex is a very difficult area.
My respects to all.
Enrique Martín
Organización Rescate Humboldt (ORH)
Caracas – Venezuela.
Email. sarrescate@gmail.com
Wsp:+584122090239
@Victor
“Case 3 at 541 secs, ROC is -3,480 fpm and at 552 secs, ROC is -15,780 fpm in 11 seconds, TTG 25 seconds.
Case 4 at 410 secs, ROC is -4,080 fpm and at 418 secs, ROC is -16,380 fpm in 8 seconds, TTG 35 seconds.
Case 6 at 349 secs, ROC is -3,360 fpm and at 356 secs, ROC is -17,100 fpm in 7 seconds, TTG 28 seconds.
Case 10 at 436 secs, ROC is -3,660 fpm and at 443 secs, ROC is -17,340 fpm in 7 seconds, TTG 30 seconds.”
The average time to the final -15,000 fpm point is 442 secs for cases 3, 4, 6 and 10.
If this point aligns to 00:19:37 UTC, then the simulations start at 00:12:15 UTC on average and 00:10:25 UTC at the earliest.
@Paul Smithson: I think there is some variability in the time for the APU to spin up, for the RAT to deploy, and for SATCOM to successfully request a log-on, all of which could affect the timing. I don’t think the calculated timings for the alternate configuration are all that different than what we expect from the satellite data.
We suspect that the SATCOM was deliberately shutdown by flipping at least the bus tie and left IDG switches, and then restoring one or both switches at 18:24. (That can be accomplished without leaving the cockpit.) I don’t think it’s a stretch consider that after 18:24, the switch configuration was with the left IDG and backup generator remaining off, and the left bus powered off the right bus through the bus tie. Why this configuration was chosen as compared to the normal configuration, I don’t know.
@Enrique Martin: Welcome to the forum, and thank for your comments. English is the common language here, so there is no need to repeat your comments in Spanish.
I assure you that the navigational error of the AUVs was carefully considered by OI. During each mission, the onboard inertial systems were re-calibrated with the host vessel to ensure the drift was kept within acceptable limits as dictated by the scanning overlap and other considerations.
I don’t understand your comments about uncertainty in the fuel level. We have records showing the amount of fuel loaded at KLIA, and we also have the fuel level transmitted in the ACARS performance reports up until 17:07 UTC. It is all consistent with the flight plan that was filed. We have also checked whether there were any abnormalities in what was filed, and the answer is the loaded fuel is consistent with company requirements.
As for the analysis of the pilot simulator data, with the limited data we have, we analyzed this closely and have tried to incorporate this into our thinking. I have posted an entire article on this topic, including links to other papers co-authored by me and others. Whether the Malaysians have performed similar or more detailed analyses is unknown, but in the end, they have reported that this evidence does not show intent of the captain. Many of us disagree, of course.
We agree that Malaysia should release all the radar data (including the military radar data), all cell phone registrations, and whatever other information they have.
@Victor
Below is a link to a table with a timeline for Cases 3, 4, 6 and 10:
https://www.dropbox.com/s/ll0pgqqba1a5o3u/End%20of%20Flight%20Simulations%20Timing.png?dl=0
I have assumed that the final time when a ROD of 15,000 fpm occurred was at 00:19:37 UTC.
The time when both engines flame out is marked by the speed starting to drop, followed 4 or 5 seconds later by the altitude starting to drop. This event appears to happen a little earlier than the ATSB assumption of 00:17:30 UTC.
@Victor
As you pointed out earlier, in each of the Cases 3, 4, 6 and 10, the time of impact is before the IFE setup at 00:21:06 UTC.
@All
Just in case this hasn’t been discussed before, has anyone got any views about the picture featured in the article below:-
https://www.businessinsider.com.au/the-search-for-debris-from-malaysia-flight-370-2014-3
The pictue is titled; “A member of staff at satellite communications company Inmarsat works in front of a screen showing subscribers using their service throughout the world, at their headquarters in London March 25, 2014.”
Clicking on the picture brings up the hi-def version (3500 x 2191).
The coverage pictured is obviously Europe, the Middle-East and Africa and is not Inmarsat-3 F1 but could be Inmarsat-3 F5, or more likely Inmarsat-4A F4 given that the highest area ID number is 188 and each I-4 satellite provides around 200 narrow spot beams.
Apparently each I-3 satellite provides four to six regional spot beams, so I wondered whether Classic Aero utilised them and whether MH370 had transited a beam boundary.
If only a single global beam were used it would be interesting to know how symmetrical the directive pattern of the antenna is (along with the signal strength data from MH370).
@Victor Iannello,
Above you said: “On the other hand, as Richard G. observes, the Boeing simulations reaching high descent rates and high downward accelerations all impact the ocean close to the point where those values occur, in which case the details of the flight dynamics really don’t matter, and the plane should be found close to the 7th arc.”
You didn’t mention the usual disclaimer “. . . unless a pilot glided the aircraft afterwards, in which case the aircraft may have impacted as far as ~100 NM from the 7th Arc.”
@Andrew. Final Report. You said, “I wonder if the name changed because of the inconclusive nature of the report and the sensitivities of the families, etc” Victor says its name was changed from that after “uproar” which is what I remember the press reporting.
To the pedant in me if one writes ‘final report’, in lower case, it is a descriptor. With capitals, ‘Final Report’ as in Annex 13 indicates that is to be its title. There are the defined expectations of a Final Report which for example include damage assessments. There is a fullsome guide as to how to the sorts of things be expected. The DGA flaperon report is the only one on a recovered item which is thorough, with scanning electron microscopic examination of fractures for example. The converse is the Malaysians making no mention in their report of the vortex generator. That appears to be languishing uncollected in Madagascar still and unexamined even as to its origin, though the press reported it that it has been available for collection since June.
The suspicious bit in me broke confinement on learning that this was not to be called a Final Report, for that appeared to relieve Malaysia of the formal obligations.
I am ever hopeful that the ICAO will ask for more work on the report, not of course just its lack of damage assessment. The ICAO is the fullback if those accredited did not or could not see to something with fewer shortfalls. Based on the precedents you cite though the title should make no difference to the way the ICAO receives it so that provides some reassurance as to the formal situation.
I hope we learn of its reaction.
@TBill. Yes it would be interesting to learn of the NTSB’s position noting it has been forceful in the past.
@all
I am half way complete on a new book about MH370 that focusses on the evolution of analytical thought relative to the dissappearance.
Title: “MH370 Analytics – From Occam to Metaphysics”.
@Victor, You quoted the AMM, “When the RAT is extended and hydraulics off, the airplane rolls left. Two to three units of right control wheel rotation are necessary to hold the wings level.”
Perhaps they mean the RAT drag due to its power generation, which has an upwards vector, would roll the aircraft left though as you observed the moment arm is short.
Also about the notion of the hydraulics being off you say, “We decided that the AMM was referring to a condition in which the hydraulics were selected off.” There is provision for the RAT to deploy automatically when hydraulics drop off but surely they do not have in mind that there could be good reason why someone would select them off?
To me, more likely a total loss of all hydraulics is envisaged, like the DC 10 after an engine failure ruptured all hydraulic lines. Also though, while we came to the conclusion that windmilling engines would provide hydraulics, that is unconfirmed and, particularly, would be the way at to slow indicated airspeeds.
One other general observation about drawing conclusions from Boeing simulations is that as you are aware, the data might be from outside the flight envelope or derived from aircraft behaviour outside the simulator data base. This could apply particularly to those descents which exhibit high descent rates.
Related, I would like to know whether in any of those the aircraft structural limits were exceeded but I doubt the data could be relied on with those caveats in mind.
My particular reason is that I expect to post a revised and comprehensive study of the flaperon separation, starting with the French DGA report on that. Its conclusion will be that on the balance of probabilities, in-flight separation in a high speed spiral was the cause.
Thanks for the welcome Victor Iannello, a pleasure to share with everyone here and I hope to learn and help a little.
I am happy that the navigation error of the AUV was considered by OI, it is correct. But little comment about it and that is not so good. On the other hand there is a report that 2 of your AUV were lost, I could not verify, but if true, this allows me to have reasonable doubts about the inertial navigation system. If you know something about it please, I’d appreciate a comment from Victor.
About fuel, I hope everything is clear, but I will explain in cases:
Suppose that return of a flight and its remaining fuel were 1000 LBS. For the flight to China it required 2000 LBS and that was placed, charging 3000 LBS accordingly, but they reported only 2000 LBS. That’s what I mean, because nowhere in the report is the amount of fuel remaining, but the total amount on board. Hence my doubt. But if everything is clear, then there is no doubt about it.
And if I share the need to release data, it is necessary to be able to have everything clear.
I think there is a lot of valuable information that can certainly help us a lot in this case, but there is no way to review it, thank goodness. You esteemed Victor achieved access to a material, which can help to see another phase of the accident from a different constructive point of view , as well expressed in his writing, may be north or south of the seventh arc, everything depends on the scenario, therefore all are valid today, because we do not know what happened in that cockpit.
I have no doubt this is a complex SAR case and what I see, is that the least there are experienced SAR experts, starting with the USCG, which at the beginning of the case towards the Indian Ocean were not consulates. I speak of people with real experience in SAR, not military who looked for 2 or 4 planes at some point in their career, but years looking for aircraft and boats, because the prospects of that experience will make a big difference. Hopefully this change is necessary for many to analyze everything and think about it, there is no other to add and work as a team, as today by your initiative Victor we are here, ALL from what I could read with the best desire to contribute and I join Everybody.
Sorry to put it in Spanish, it’s an old constubre.
Thank you for allowing me to be here with you.
Enrique Martín
Rescue Organization Humboldt (ORH)
Caracas Venezuela
WEB: http: //www.rescate,com
WSP: + 58-412-209-0239.
EMAIL: sar_rescate@hotmail.com – sarrescate@gmail.com
Dear @ flatpack and that I mean with the satellite.
On the east side of the satellite that received the hand grips, there is another satellite to avoid losing links, that is, in case the plane in flight leaves the coverage of a satellite, another one takes its signal.
It’s like cell phone, but the phone system or in this case the link automatically determines which satellite has better signal reception and definitely links the plane’s system to that satellite.
But what happens, it is possible that at some point in the navigation of the plane, the other satellite that is east of your route, try to link or connect with the plane’s system. This possibility of having been generated would have been registered in the company’s system (although, after a long time, these records may no longer exist, there is no doubt that it is true that it escapes), that is what happens with cell phones and SAR operations. Cellular tracking allows us to define a search area and define the area.
Hence my question and that is why your doubts are mine.
Enrique Martín
Organización Rescate Humboldt (ORH)
Caracas – Venezuela
WEB: http://www.rescate,com
WSP: +58-412-209-0239.
EMAIL: sar_rescate@hotmail.com – sarrescate@gmail.com
@flatpack
That image looks like Alphasat at E24.5º.
Inmarsat-3 craft provide 7 spot beams. Classic Aero may use a spot beam, but I understand that would be more typical of an Aero-I installation (I for intermediate gain antenna). 9M-MRO was very definitely using Aero-H, with its high gain antenna, and the SATCOM traffic log shows that the global beam was used exclusively. I have had sight of other logs for Classic Aero traffic, the global beam was exclusively used for the traffic detailed in those logs.
These satellites carry much more traffic than just Classic Aero: other higher data rate services command higher charges to subscribers, spot beams do enable higher data rate services to small portable terminals.
@Dr B said: You didn’t mention the usual disclaimer “. . . unless a pilot glided the aircraft afterwards, in which case the aircraft may have impacted as far as ~100 NM from the 7th Arc.”
We’ve been talking about the Boeing simulations, which included no pilot inputs. I thought that was understood. In the discussion section of the post, I specifically cited the possibility of pilot inputs and long glide (Possibility 3).
If there were pilot inputs at fuel exhaustion, then likely there were pilot inputs between 19:41 and fuel exhaustion, and in addition to a large search width, it will be very difficult to constrain the latitude range. I’m not seeing a way forward with that assumption.
@DennisW: Sounds like an interesting read. I’m sure many here will read it.
@David: I agree that parts of the simulation were likely beyond the range of Mach number and angle of attack that have been verified for model accuracy. However, if the aircraft dynamics had achieved such an extreme state, it is unlikely the aircraft would have flown much further. It doesn’t matter much if the impact distance from the point of high rate of descent was 5 or 6 NM.
We’d be very interested to read your report discussing the inflight separation of the flaperon. Please share it when you are ready.
@David,
The effect described when “the RAT is extended and hydraulics off, the airplane rolls left. Two to three units of right control wheel rotation are necessary to hold the wings level” was deemed due to the left flaperon PCU being set into bypass mode.
Without pressure in the L and R hydraulic systems this PCU is set into bypass, when in bypass the flaperon fairs free in the slipsteam with its T/E up causing a left wing down roll.
The Training Point noted hydraulics OFF: that was ultimately understood to mean no pressure in the L and R systems as a consequence of the flight deck EDP switches set of OFF. By setting an EDP switch to off the pump output is blocked and the pump is ineffective whereas a windmilling engine with EDP switch set ON would continue to provide flow and pressure.
@Victor
RE: ‘I was proposing a situation in which the A/T is commanding full thrust and the thrust lever is already at the full position, such that there is no disagreement between target and actual thrust, both being “full”.’
The ‘disagreement’ is a disagreement between the commanded ASM rate and the actual ASM rate; it is not a disagreement between the target thrust and the actual thrust. The AMM reads:
‘When the A/T engages, the control law sends rate commands to the ASMs. The ASMs send rate feedback (low speed tachometer data) to the TMCFs. The servo loop monitor compares these two signals and sets the inhibit logic when they do not agree. This causes the ASMs to disconnect.’
ALSM’s video shows that an AUTOTHROTTLE DISC message is generated when the electrical power fails, even if the thrust lever has reached the commanded position. The message is not removed when power is subsequently restored by the APU.
RE: ‘The change in control surface position could be related to either the APU or the RAT coming online.”
My understanding is that the Boeing simulations did not model the APU autostart. There may be a case for the RAT deployment causing a change in the right flaperon position if the C hydraulic system initially lost pressure after the R engine flameout.
The right flaperon actuators are powered by the R & C hydraulic systems, controlled by the L2 and C ACEs respectively. When electrical power was lost after the R engine flameout, the L2 ACE would ‘fail’ and the C hydraulic system primary pumps would also ‘fail’. If the C hydraulic system lost pressure, the R flaperon’s actuators would both be in bypass mode and the flaperon would float upward.
Reasons for the C hydraulic system to lose pressure include an abnormal hydraulic or bleed system configuration, such as:
– C1 & C2 demand pumps selected OFF, or
– L bleed air isolation valve selected OFF, or
– L engine bleed switch selected OFF
In that state, the R flaperon would float upward until the RAT deployed and started providing hydraulic pressure to the C system flight controls. At that point, the R flaperon’s C system actuator would go back to normal mode and the flaperon would be centred. That said, I don’t think there was any indication the Boeing simulation’s hydraulic or bleed systems were configured abnormally??
@DrB
RE: “How long does it take for the RAT to deploy? The RAT produces drag on the left side of the aircraft. Thus the RAT itself would induce a slight left turning of the aircraft.”
The AMM states that ELMS triggers the RAT to deploy when both transfer buses lose power for more than 15 seconds. The RAT then deploys in about 2 seconds, so the total time for deployment is a minimum of 17 seconds after the loss of electrical power. The RAT should spin up immediately upon deployment.
The RAT is located on the right hand side of the aircraft, as shown in the following images:
https://aeroblogbrasil.files.wordpress.com/2008/06/dsc_02722.jpg
http://www.boeingimages.com/Docs/BOE/Media/TR3_WATERMARKED/4/d/d/1/BI42131.jpg
https://i.ytimg.com/vi/BYVl8Wk-fCw/maxresdefault.jpg
The drag caused by the RAT would tend to make the aircraft yaw (and the secondary effect of roll) to the right. Given the relatively short distance between the RAT and the fuselage centreline, I think the yaw effect would be minor.
@Andrew: There is no servo rate error if the target position of the lever is the same as the measured position. Nor am I proposing that a disconnected autothrottle would automatically connect after the APU restores power.
We usually do well, but we’re not converging on this one. I don’t think it’s productive to continue discussing a hypothetical scenario which I now believe probably did not occur.
Why do you believe that Boeing did not include APU autostart functionality in the simulations? My understanding is that what was not modeled is residual fuel in the APU fuel line when there is no fuel in the left tank. However, for the alternate configuration, there is fuel remaining in the left tank when the autostart would occur.
@Victor
Fair enough; I’m afraid I don’t understand the scenario you proposed. If I might be permitted one more comment on the subject: the system is not measuring the target position of the lever vs the measured position. It is measuring the commanded servo motor rate vs the actual servo motor rate. For example, the ASM has a clutch that allows the pilot to override the thrust lever movement. If the pilot does override the thrust lever movement, the ASM continues trying to move the thrust lever against the pilot. In that scenario, the autothrottle does not disconnect because the ASM is still operating correctly. The disconnect logic does not depend on thrust lever position.
RE: “Why do you believe that Boeing did not include APU autostart functionality in the simulations?”
My mistake, I was thinking of the simulations where the configuration was normal.
@all @DennisW
DennisW dismissed suicide but raised the issue of political motive in an earlier post. There were some indications that Zaharie had concerns with the government. Assume that Zaharie wanted to grossly embarrass the government of Malaysia by making MH370 disappear and never be found. He will die as a consequence but this is a different objective than a pure suicide. If this was his objective he was extremely successful. Best ever!!!
He depressurized the aircraft early to humanely put the PAX to sleep and then die. He flew the plane into the sea. I do not know how you would fly the last few minutes to minimize surface debris. For example, maybe you dive to low altitude, bleed off airspeed, and then loop backward to enter a vertical dive but at low speed. If the SEA Q400 ramp worker could invert the aircraft and pull though a half loop, Zaharie had great skills and a simulator to practice. I am not suggesting this maneuver, but there must be some best way not have a breakup on the sea surface. My guess is a low speed vertical would be better than breaking up in high waves with horizontal entry. Ditching in rough sea state is often the best way to scatter wreckage all over the surface – not like the placid Hudson river.
Because he likely had no knowledge of the SATCOM pings and the concept of arcs, there is no reason to extend range by gliding after flameout. He would just perform the final entry maneuvers. The RAT would provide FBW control for the primary controls (but not flaps) so Zahaire would be able to have good control capability for a clean aircraft.
He does need to decouple the max range path from the max endurance path which I stated in an earlier post. He would perform periodic 360 turns of 15 to 30 minutes duration or S-turns. These would pull the arc 7 point north and maybe above 25 south. These would really mess with the arc to arc probabilities – not that he would know.
Everyone seems happy with the ATSB/DSTG particle filter which centered the hot spot at 38 south. I think the model was flawed. Their filter was designed to fly straight airline routes and verified for these. So for a ghost ship on AP, the model might be just fine. But if Zaharie flew this to sea entry and intentionally engaged in 360 turns or broad S-turns, the BTO/BFO analysis would show different hot spots. A single 360 could have a significant impact on the arc to arc probabilities. It would have been interesting to see a simulation where the required one or more random 360 degree turns after the turn south – very different map would result.
It looks like much thought went into the turn back at IGARI, the route to Penang (which is the best diversion airport), the flight up the center of the Strait of Malacca. So why wouldn’t significant planning have gone into the flight path to water entry? It possible that a mechanical failure caused the crew to immediate divert toward Penang and then died and then everything else was a ghost flight. The IGARA to MEKAR path may have all just been random and only looks planned after the fact.
Clearly nobody in Malaysia or ATSB wanted to consider a planned mission to hide the plane under the ocean so never considered anything other than a ghost flight and unpiloted final descent. Boeing did what they were told by ATSB.
So maybe Zaharie planned and executed his last mission very well and died with his passengers and crew, and created the worlds greatest aircraft mystery and showed the incompetence of the Malaysian government.
@All. That’s amazing. So 5/6 “normal” configuration simulations don’t even come close to producing -15,000fpm ROD at ANY stage [max is a little over a third of that]. 1/6 manages to do so, but fully 13 minutes after the timing “predicted” by our BFO. Should we not be concerned? It certainly doesn’t look to me like an affirmative piece of evidence! These simulations do NOT succeed in replicating the high ROD at ~2 mins after power loss required by the BFO.
To then seize upon the “alternate” configuration scenarios that did produce high ROD and say, effectively, “it must have been like that” is disingenuous. Nobody except the Accidentalists would predict that an alternate configuration might be in use. There is no reason for the Pijackers to invoke it.
@Paul Smithson
Once again you ignore the data, because it does not fit your scenario.
Even worse, you then determine that anyone following the data, is not sincere, deceitful and dishonest.
Fact is, the Cases 4, 5, 6 and 10 fit the BFO data and a timeline from dual flameout to impact.
https://www.dropbox.com/s/ll0pgqqba1a5o3u/End%20of%20Flight%20Simulations%20Timing.png?dl=0
Can you please forget the polemic and start dealing with the data.
@Andrew
I would like to address this question to You, since You are a B777-pilot:
If a trained pilot needs to ditch a B777, how much fuel/flight time would he chose to have in the tanks, when initiating the the first attempt? Of course assumed he has that choice.
In my humble view it seems clear, low fuel is fine, but he would like to have some reserve to have the chance for 2 or 3 runs or so.
Thank You very much fro Your time.
Michael
Well said Paul,
We now know the IGARI turn was flown manually(why it took 4 years to release that?).
It seems the flight across the Peninsular was an autopilot-off meandering flight.
The electrical configuration seems to have been severely compromised ( with something like only the R IDG and L BU generator remaining).
This all points to serious mechanical failure followed by pilot incapacitation. None of what we know fits with a pijack scenario.
@RG.
Case 5 (conventional) does NOT fit the BFO data and timeline since the required ROD occurs >10 minutes too late.
Cases 3,4,6,10 DO produce the required ROD and the timing is not bad. But none of these have anything to do with the conventional EOF narrative per ATSB reports (loss of power when left engine goes down). Instead they require an unusual, alternate electrical configuration to be in place – with all AC power sourced from right engine. Anyone who now wishes to depend upon these (alternate) cases as evidence of BFO veracity also needs to provide a good reason why such an alternate configuration should be in place. What’s so unreasonable about that?
@Victor
For the Cases 3, 4, 6 and 10, the maximum vertical acceleration g-force experienced was between 0.902g (Case 4) and 1.462g (Case 3). Case 6 was 1.213g and Case 10 was 1.400g.
For the Cases 3, 4, 6 and 10, the maximum lateral ground speed experienced was between 576.962 knots (Case 3) and 612.368 knots (Case 4). Case 6 was 599.093 knots and Case 10 was 609.260 knots.
A maximum ground speed of 612 knots gives a TAS of 578 knots (around 0.92 Mach at 17,600 feet).
I am interested to know whether speeds up to 0.92 Mach and vertical accelerations up to 1.462g are sufficient to cause aerodynamic flutter and separation of control surfaces in flight.
Don Thompson says:
August 21, 2018 at 10:51 am
@Air Life,
I don’t doubt you hold the opinion that a right turn may have occurred, if any officials have shared anything back to corroborate your opinion that’d be very interesting indeed.
The traffic count through the KL FIR was not high on the night 7/8th March. A combined picture of SSR identified plots vs PSR unidentified plots would have readily differentiated 9M-MRO’s post diversion track. The Malaysian reports of 8th March 2015 and 30th July 2018 describe that they made a conclusion for the post-diversion track, using radar playbacks, that 9M-MRO turned left.
To turn right, not be detected by any PSR, and make the line of position described by the 18:25 Log On, implies that the aircraft would have actively evaded PSRs. Those PSRs would include DCA TARs at WMKC, WMSA, WMKK, WMKJ, and WMKL; the RMAF operated TARs at WMKB and WMKD; and the long-range air defence radars at Western Hill-Penang, Kuantan-Phanang, Bukit Lunchu-Johor, and Bukit Gombak-Singapore.
Don, thank you for your reply. Yes, a right turn back over the SCS to KLIA, with technical difficulties, would be “atively” the safer/safest, manual manoeuvre to execute. I have had acknowledgement from both the ATSB, and finally from the Malaysian officials. I hope they are now actively looking into the right turn back to KLIA option.
@Paul Smithson
You stated “Anyone who now wishes to depend upon these (alternate) cases as evidence of BFO veracity also needs to provide a good reason why such an alternate configuration should be in place.”
Victor already answered you on that point in his linked comment:
http://mh370.radiantphysics.com/2018/08/19/end-of-flight-simulations-of-mh370/#comment-18069
“We suspect that the SATCOM was deliberately shutdown by flipping at least the bus tie and left IDG switches, and then restoring one or both switches at 18:24. (That can be accomplished without leaving the cockpit.) I don’t think it’s a stretch consider that after 18:24, the switch configuration was with the left IDG and backup generator remaining off, and the left bus powered off the right bus through the bus tie.”
@Paul Smithson: Calm down. I’ve seen some here claim that others are “disingenuous” when they present scenarios that are counter to their own. Once the accusations start to fly, it usually doesn’t end well for the accusers.
The end-of-flight dynamics for the alternate electrical configuration is being proposed as a possible explanation for the timing and values of the rapid descent derived from the final BFOs. The missing IFE log-on is consistent with an uncontrolled flight that ends soon after the log-on acknowledgement message that was sent to the GES.
What is your explanation for the final BFOs and the missing IFE log-on?
@DrB, @Andrew: The roll to the left about 40 s after shutdown of the right engine is not explained by a reduction in left thrust as it is not consistent with the subsequent altitude and groundspeed. Nor is the roll to the left explained by RAT drag, as the timing and direction don’t match. I believe what remains is a change in position of one or more control surfaces after the APU autostarts and comes online. Is there another possibility?
@Victor @Richard
“… I don’t think it’s a stretch consider that after 18:24, the switch configuration was with the left IDG and backup generator remaining off, and the left bus powered off the right bus through the bus tie.”
Could one reason for that config be to save fuel by running less generators? Or to get more power for who knows what manuevers we might gave missed after radar coverage ceased at 18:22.
@TBill: With only one main generator running (the right IDG), ELMS will shed loads, so there will be an additional increment of fuel available for propulsion, and also an increment of additional thrust capability. I doubt that would be a consideration, but I am guessing.
@Hank
DennisW dismissed suicide but raised the issue of political motive in an earlier post. There were some indications that Zaharie had concerns with the government. Assume that Zaharie wanted to grossly embarrass the government of Malaysia by making MH370 disappear and never be found. He will die as a consequence but this is a different objective than a pure suicide. If this was his objective he was extremely successful. Best ever!!!
I do think the diversion was politically motivated, but with the intention of getting something from the Malaysian government. Not simply to create an embarrassment. The early response of the government to the diversion strongly suggests that they were aware of what was going on at high levels. The decision to search the South China was made before Inmarsat surprised everyone with their data, and the government had to scramble to create the radar narrative.
A quick thought about the turn-back.
Given that there was an initial turn to the right segueing into a (longer) left hand turn, is it possible that this was intended to ‘feel’ like an offset manoeuver to the PAX and/or crew?
Thus they may think that they were still broadly heading in the right direction (for a short while anyway). Over the sea they would have few points of reference.
This might just be possible, acceleration can apparently play strange tricks. I recall reading about a fatal crash involving a light plane piloted by one of the Kennedys where this was reported as a (much more extreme) factor.
I do not fully understand how a turn can feel like a dive let alone whether this can be reverse engineered by a pilot.
If this seems like some over elaborate stage-magic trick then, assuming it actually occured, that is exactly what it was.
@DennisW
The search of South China Sea was a logical response to the loss of the transponder at IGARI. Everyone expected an immediate crash near IGARI.
I assume you meant the search off the west coast which was a surprise because the turnback observed on radar was not discussed. It was just odd that they were searching so far from IGARI.
I learned that Inmarsat checked the MH370 data within hours of the reported disappearance and noted that the aircraft was in the air for 7 hours and this was provided to the government before any BTO/BFO work was done by them.They did privately report the time in the air to Malaysia. This was not announced by Malaysia, but this could have been the wake up call that led to looking at the radar data and searching to the west,
It may be that there were other communications based on a hijack, but the private Inmarsat alert on flight duration may have been the driver.
@Hank
…or, there were people in Malaysia who knew the plane was airborne for some time after IGARI, but did not make that info public until the Inmarsat data forced their hand.
@DennisW
If you are correct, Z must have had a deadline for the government to comply, with a threatened consequence for non compliance, or a “plan B” if you like, otherwise the government would simply “wait it out”, thinking that Z would have to land somewhere, so he would loose. He didn’t land. The government lost the gamble instead. Then the scrambling commenced.
@Ventus
My thoughts, exactly.
BTW, my book “Occam to Metaphysics” is largely complementary to our community here. To me metaphysics tries to define what it is possible to know. The cognoscenti here have gotten very careful about that, and it is a good thing, IMO,
@Victor Iannello,
@Andrew,
Andrew, thanks for the RAT pictures. I recall a long time ago a drawing was shown with the RAT on the left side at the rear of the fuselage. Obviously that was incorrect for this aircraft.
Victor said: “@DrB, @Andrew: The roll to the left about 40 s after shutdown of the right engine is not explained by a reduction in left thrust as it is not consistent with the subsequent altitude and groundspeed. Nor is the roll to the left explained by RAT drag, as the timing and direction don’t match. I believe what remains is a change in position of one or more control surfaces after the APU autostarts and comes online. Is there another possibility?”
I can’t think of any. It seems like most/all of the hydraulics would still be operational with the R engine windmilling and the L engine running normally, but with no AC power until the APU started. A R wing control surface would need to be driven downward or a L wing control surface upward to causes a roll to the left when the APU restored AC power. Maybe someone here can identify the control surface which would move after AC power was restored.
The timing of the left roll is a bit off from the ATSB estimate. I think the ATSB said it took 1 minute from engine fuel exhaustion for the APU to restore power. In the Boeing simulations the left roll starts about 40 seconds after R engine fuel exhaustion. That is about 20 seconds sooner than the ATSB estimate. I don’t know if this timing difference is significant or not.
@Victor Iannello,
@Andrew,
Victor asked Paul: “What is your explanation for . . . the missing IFE log-on?
Is it possible that the IFE is load shed in the alternate electrical configuration, so that it is powered when the R IDG is operating but not after the R engine fails and the APU starts? I would have thought that IFE would be a good candidate for load shedding.
@DennisW
Unlike the ATSB decisions which are very well documented, the Malaysia activity is very opaque, even now. Could be a cover for blowing a ransom deal as you suggest. I have never seen any MOT description of why they started the expanded search – a detailed discussion of who decided what, when, based on what information would be nice to see documented unless they can’t.
If it was a hi-jacking, the Malaysian Authorities would have known about it and demands would have been received and that they would have made public. Why wouldn’t they? It’s not their fault! With COMMS down starting with ACARS at 17,07 UTC or 1.07 local time, then RADIO after 17.19 UTC or 1.19 Local time AND unanswered calls made to the aircraft later on, I really don’t see how the Malaysian Authorities could have received ANY demands. Hi-jacking is another red herring thrown into Cooking Pot of Causes to confuse and distract us up a blind alley. There are too many nonsense causes. Find the wreckage first, and then let the wreckage do the talking.
@Air Life: I don’t think Najib’s government would have been very eager to publicize demands that exposed the corrupt Najib government and demanded elections. The negotiations could have been with collaborators on the ground, and not directly with the pilot. In that scenario, something must have gone wrong, leading to a crash rather than a landing, in which case, both sides would have an interest in keeping the details undisclosed.
@Victor
@DrB
RE: “I believe what remains is a change in position of one or more control surfaces after the APU autostarts and comes online. Is there another possibility?””
Nothing that springs to mind. That said, I can’t find a reason for a control surface to change position after the APU autostart either.
@DrB wrote “ I would have thought that IFE would be a good candidate for load shedding.
Yes, the cabin located components of the IFE: the seat electronics boxes, the media players/storage, distribution boxes, etc. But not the few avionics units in the MEC that provide integration to the Cabin Services System, and so on.
@DrB, @Andrew: I asked the ATSB about the roll to the left, and they said they would discuss this with Boeing. Hopefully, Boeing will respond and the ATSB will share the response.
@Don Thompson, @DrB: In addition to Don’s belief that the IFE head in the MEC is not load shed, we know that there WAS an IFE log-on as part of the log-on sequence commencing at 18:25, so unless the electrical configuration changed after that IFE log-on, there should have been an IFE log-on during the log-on sequence commencing at 00:19.
@Victor
Reviewing past, I now see it took the IG about 6-months until at least Nov_2014 just to start to realize the pilot could turn off Left Bus to turn off SDU from the cockpit. Prior to that it was apparently thought a hijacker had to enter the MEC bay.
That has some implications:
(1) For 6 months nobody in the public/media had the understanding that the pilot could have done this act sitting in the cockpit…so that’s a lot of time for conspiracy and pilot-did-not-do-it theories to develop. Championing old out-dated theories is probably one reason some people do not accept pilot blame.
(2) Not sure we can solely blame MY for hiding the fact that the pilot could have done this from the cockpit. Boeing/JIT/CIA probably knew this from the get-go but nobody was super-anxious to let the public know that in part due to fear of public demand for corrective actions
(3) If it took IG up to 6 months to figure this out, the implication is ZS probably was the only one smart enough to do this.
@TBill: The IG was learning a lot about aircraft systems in that timeframe, and made many mistakes. Even now there is resistance to believing that the captain would know that isolating the left bus would bring down the SATCOM, even though there are EICAS messages advising this very fact, and he could have tested this on his simulator without poring over electrical schematics.
In the ATSB’s June 2014 update, they told us that the left bus could have taken down the SATCOM, although it took some of us some time to understand what they were saying:
The SDU was powered by 115 V AC from the left AC bus which was normally supplied by the left IDG. If power from the left IDG was lost, then a bus tie breaker would close and power would be automatically transferred from the right AC bus. Similarly, if power was lost from the right AC bus, power would be automatically transferred from the left AC bus. This power switching is brief and the SDU was designed to ‘hold-up’ during such power interruptions. To experience a power interruption sufficiently long to generate a log on request, it was considered that a loss of both AC buses or, a disabling of the automatic switching, would be required.
I’ll add we were thinking about how to disable the SATCOM without bringing down the entire left bus, which seemed extreme.
@TBill
The IG is not a self appointed global police force investigating aircraft accidents.
We are a voluntary self-funded organisation that spontaneously came together with a common interest to help solve the mystery of MH370.
We have no contracts, commitments or deadlines. We cannot be early or late in delivering any particular tasks.
You have apparently absorbed too much garbage from Jeff Wise.
For the record, the IG are not authorised, accredited, ordered, contracted, paid or recompensed by the ATSB, NTSB, AIB, CIA, FBI or any other body.
@Victor
OK thank you, so ATSB did say that in June and it took a while. Looks like a lot of early energy on MEC Bay access concerns.
@Richard
I was not criticizing IG…only trying to track the history of what we undertood and when we understood it, especially since I was not a discussion participant for some period after CNN coverage stopped.
@TBill: Yes, access to the MEC at one point seemed to answer a lot of questions. Later, I think we realized it was an unnecessary complication.
Victor Iannello says:
August 23, 2018 at 7:12 am
@Air Life: I don’t think Najib’s government would have been very eager to publicize demands that exposed the corrupt Najib government and demanded elections. The negotiations could have been with collaborators on the ground, and not directly with the pilot. In that scenario, something must have gone wrong, leading to a crash rather than a landing, in which case, both sides would have an interest in keeping the details undisclosed.
Victor, thanks, but I am not going down the Japanese Red Army incident route and I think the hi-jacking route is too far-fetched, as is the Suicide/Murder route. I think I’ll stick with the right turn back and Ghost Flight scenario as the most likely cause, even if it doesn’t turn out to be that, for my own sanity sake.
@Air Life: Many people persist in believing a particular scenario because of pre-conceived biases. That’s your right. Just don’t fault others that don’t share your bias (not that you have).
@TBill
In stating “it took IG up to 6 months to figure this out”, I find that a disparaging comment against the IG.
We never committed to solving the mystery of MH370, or any part of it, in a particular timeframe, let alone 6 months!
When you say “I was not criticizing IG…only trying to track the history”.
I am not willing to be tracked by you or anyone else.
Victor Iannello says:
August 23, 2018 at 11:22 am
@Air Life: Many people persist in believing a particular scenario because of pre-conceived biases. That’s your right. Just don’t fault others that don’t share your bias (not that you have).
Victor, my likely scenario isn’t based on pre-conceived biases, but on Tough Training, Airmanship and Experience. Thanks!
@Victor Iannello,
You said: “@Don Thompson, @DrB: In addition to Don’s belief that the IFE head in the MEC is not load shed, we know that there WAS an IFE log-on as part of the log-on sequence commencing at 18:25, so unless the electrical configuration changed after that IFE log-on, there should have been an IFE log-on during the log-on sequence commencing at 00:19.”
There definitely were major changes in the electrical power configuration after 18:28, in fact, circa 00:17-00:18. The right engine flamed out, AC power was lost from both main and transfer buses, the APU auto-started, and the RAT deployed. I understand Don believes the part of the IFE needed for log-on at 00:21 was not load shed. He may be right, but it would be useful to have this confirmed by ATSB or Boeing, since it is a vital clue.
@DennisW
Dennis, sorry for my late reaction regarding “Doppler integration”. I have taken a long summer break this year.
Henrik used something like Doppler integration, see eq. 9 in:
http://bitmath.org/mh370/mh370-path.pdf
Also, long time ago, I have been looking at an “BFO-only” approach, see:
http://nebula.wsimg.com/25f47f9d48f3cce925b3d9609e980c18?AccessKeyId=5D94BAB6DF613B220FB7&disposition=0&alloworigin=1
I’m not sure, is this in line with your ideas? Estimates of 00:11 or 00:19 latitudes based on Doppler integration will be rather inaccurate, even if you limit BFO errors to +- 7 Hz.
@DrB: By a change in electrical configuration, I mean a change in the bus switch states in the overhead panel. If one generator supplies power, there is load shedding. If two, there is no load shedding. If the left generator was isolated before 18:25, then at the time of the log-on at 18:25, only one generator (the right IDG) was supplying power. If no bus switches were flipped after that time, then at the time of the log-on at 00:19, again only one generator (the APU) was supplying power. The load shedding would be similar in both cases. If the IFE was not load shed for the first case, it would not be shed in the second.
@DennisW
Regarding your comment:
http://mh370.radiantphysics.com/2018/08/19/end-of-flight-simulations-of-mh370/#comment-18106
I’m not sure when exactly the existence of the Inmarsat data was known to the Malaysian government/MAS/SAR coordinators.
However:
A link to an early report (apparently dated Monday March 10th) about radar data showing a turnback was recently tweeted by Julie (@Nihonmama):
https://web.archive.org/web/20180813193345/https://peraktoday.com.my/2014/03/ada-isyarat-pesawat-berpatah-balik//
Google translate:
“Kuala Lumpur: The disappearance of the MH370 plane was enveloped in mystery when there was a signal on the radar showing it turned back to Kota Bharu at 2.40am yesterday.
According to sources, it detected the Royal Malaysian Air Force (RMAF) control tower, Butterworth and control rooms of several Royal Malaysian Navy (TLDM) vessels overseeing the security of the South China Sea and the Singapore Air Traffic Control and Surveillance Unit.
RMAF Chief, General Tan Sri Rodzali Daud, said it did not rule out the possibility of a plane returning to Kuala Lumpur Airport (KLIA) before being reported missing from the radar.
Impact of movement
“The military radar detects the movement of the aircraft back and causes search and rescue operations to expand into the waters of Penang,” he said at a press conference at the Operations Coordinator Center, in Sepang yesterday.
MAS Group Chief Executive Officer Ahmad Jauhari Yahya, however, said it did not receive any emergency calls on damages that forced the plane to return.
-Daily news”
@AirLife: “If it was a hi-jacking, the Malaysian Authorities would have known about it and demands would have been received and that they would have made public…”
I’d agree in part: I’m not at all sure that the Malaysian govt would be open, but it would surely be in the interest of the hijackers to broadcast the situation — make sure third parties — and the world at large — see what’s going on in real time, as a way of maximizing pressure on the govt, and inviting blame if negotiations go south.
Against VI, Dennis, others: There are just so many problems with this being a hijack with negotiations: it requires a pi-jacker to fly for hours with live passengers and crew on the other side of the bulkhead, and certainlyt in that scenario with more than one party trying to phone as they passed over the peninsula and Penang. It surely would have triggered a large aerial chase, and probably requests for assistance from other countries. The latter, even if confidential, would almost certainly have leaked by now..
It’s of course possible that ZS could have “negotiated” in bad faith, with the passengers/crew already comatose or dead. But that doesn’t explain why every interceptor in the region was not in the air that night.
@lkr: I could come up with reasons the negotiators would not want to be identified during negotiations, but I would prefer that @TimR weigh-in, since he is the theory’s staunchest proponent.
@TimR: Why is it that you believe negotiators did not try to leverage public opinion to strengthen their demands while the plane was still flying?
@Victor Iannello,
You said: ” If the IFE was not load shed for the first case, it would not be shed in the second.” I assume that is for the case of the APU powering up circa 00:18:30 What if it didn’t? Would that mean the SDU would not have logged on, because the RAT won’t power the SDU?
@Niels
So how would you explain the initial search in the South China Sea?
@DrB: In either the “normal” or “alternate” electrical configuration, there would be no power to reboot the SATCOM if the APU does not autostart. The RAT does not supply power to the left bus, which feeds the SATCOM.
@DennisW
Perhaps it is not surprising they started searching in SCS on Saturday. On Sunday 9/3 they expanded the SAR operations to the west, according to:
https://m.youtube.com/watch?v=soY45F3tgTk
Of course it puts in question the later statements that they followed the turn back in real time and knew it was MH370 (with no need to react)
Niels says: A link to an early report (apparently dated Monday March 10th) about radar data showing a turnback was recently tweeted by Julie (@Nihonmama):
https://web.archive.org/web/20180813193345/https://peraktoday.com.my/2014/03/ada-isyarat-pesawat-berpatah-balik//
There was a post by “Iceman29” on a French forum on the Sunday 9th at 09:06 EU time (08:06Z) about the turnback, with a link to:
https://quebec.huffingtonpost.ca/2014/03/09/enqute-pour-terrorisme-a_n_4928216.html?utm_hp_ref=canada-quebec&ir=Canada+Quebec
“Il existe une possibilité réelle que l’avion ait fait demi-tour”, a déclaré le chef de l’armée de l’air malaisienne, le général Rodzali Daud, sur la foi d’analyses radars.
[Google translate: “There is a real possibility that the plane has turned around,” said the head of the Malay Air Force, General Rodzali Daud, on the basis of radar analysis.]
Regards,
PL
***
@Victor
ELT: the Safety Info/Flaperon Report suggests a ditch which makes the lack of ELT signal upon crashing more suspicious yet again. Is there a breaker for that main ELT? I bumped into discussion along those lines.
DrB:
2014-03-07 at 12:50: the aircraft powered up on the APU. At that time, after SATCOM Log On, the IFE hosted applications opened their connections over SATCOM. We know that the APU was running because ACMS sent a message just after pushback that the APU had been running for more than 4hrs.
Sure ATSB, or Malaysia (lol), might confirm that, but the ACARS data tells us that. Their reference would also be the ACARS correspondence.
@TBill
RE: ”Is there a breaker for that main ELT?”
No. The ELT has an internal battery pack that is not connected to the aircraft. The ELT control panel on the overhead panel in the cockpit is powered by the master test and test circuit for the cockpit panel lights. Removal of power from the control panel would not affect the automatic operation of the ELT in a crash.
Dennis: “So how would you explain the initial search in the South China Sea?”
How about bull-headed sunk-cost stupidity in those in charge, and cowardice for those reporting to them. Combination of “I can understand” for plane crashing right after transponder goes out, to too-much-information-I-don’t-want-to-understand for Inmarsat’s informing the plane was still in the air.
In your book, I hope you address how often we find that incompetence beats conspiracy in explanatory power. ?Ockham’s razor-cuts?
@TimR
If MH370 followed a flight path out over the Andaman Islands, entering a holding pattern pending negotiations, followed by a flight around the south of Sumatra heading toward Cocos Islands (YPCC) or Christmas Island (YPXM), as you have proposed, then the simplest way to navigate would be to enter a couple of waypoints. Waypoint PILEK is nicely out of range of the Indonesian radar at Sibolga and waypoint EPGUP places you in a nice position between YPCC and YPXM, leaving both airport options open. The flight path shows a good general fit to the satellite data.
If the MH370 flight was at FL250, Mach 0.642 (which is the corresponding MRC speed), the TAS would be 383 knots and the range from the 1st Arc given in the Boeing data in Appendix 1.6E of the Malaysian Safety Report is 2,363 NM.
Making some assumptions about the holding pattern over the Andaman Islands, and following the flight path from PILEK to EPGUP, I calculate that at 22:41:22 UTC, MH370 is 27 NM short of waypoint EPGUP at 10.2463°S 99.3171°E (5th Arc) on a track of 150.2528°T. This fits the BTO and BFO satellite data exactly. At this point YPCC is 187 NM distant to the South-West with a tailwind and YPXM is 377 NM distant to the North-East with a headwind. Both airports are within range. I calculate the fuel remaining is good for at least another 583 NM.
Obviously MH370 did not land at either airport. If MH370 carried on following a track of 150.2528°T, the aircraft reaches the 6th Arc at fuel exhaustion. The end point on the 7th Arc is at 19.4130°S 104.7428°E.
I have recently completed a revision of my drift analysis using the data from 300 undrogued GDP buoys and a corrected windage calculation based on the CSIRO data from a genuine Flaperon, which points to a MH370 End Point around 20°S near the 7th Arc.
Here is a link to a GE map of the waypoints and airports in question. The points M5, M6 and M7 are where MH370 crosses the respective Arcs:
https://www.dropbox.com/s/yt6jdxdfhy3rh9q/MH370%20Flight%20Path%20Model%20V17.0%20EPGUP.png?dl=0
@DennisW
I should add that I do find it strange that the SCS continued for a week or so, see for example:
http://www.asiaone.com/malaysia/missing-mh370-3-countries-end-south-china-sea-search
One would expect that it would not take more than the whole weekend to “understand” the essence of the radar play-back.
@PaxLambda
It is possible that the info refered to on the French forum was based on the 9th March press conference, see the youtube link in my 4:52 pm posting.
@Victor Iannello,
In your August 21st additions to the header article, you said: “
If we consider that the end-of-flight Boeing simulations were representative of the actual conditions of MH370, we have three possibilities:
1. If the flight was uncontrolled after fuel exhaustion, and if we ignore the final BFO values, the plane impacted no farther than about 15 NM from the 7th arc.
2. If the flight was uncontrolled after fuel exhaustion, and if the descent rates suggested by the final BFO values are valid, then the aircraft impacted no more than about 8 NM from the 7th arc.
3. If the flight was controlled after fuel exhaustion, an efficient, stable glide starting at about 40,000 ft could have achieved a distance of 120+ NM from the 7th arc.
Possibilities (1) and (2) suggest the next search should occur along the 7th arc north of 25S latitude at a width of about +/-25 NM. The +/-25 NM is an estimate that includes uncertainty in the simulation models and uncertainty in the final BTO values.
Because possibility (3) requires pilot inputs after fuel exhaustion, if this occurred, we have to assume that pilot inputs also occurred during the powered part of the flight after 19:41. As such, there would be no reason to believe that paths reconstructed with the assumption of automated flight with no pilot inputs are representative of the path actually flown. The large width (+/- 120 NM) and the difficulty in objectively constraining the crossing latitude of the 7th arc would make it difficult to define a manageable search area without additional information.
Until additional information or insights become available, it will be difficult to establish a new search area.”
My Comments :
I disagree with several of those points.
First, I don’t think it is difficult to establish a prioritized list of new search areas. At present there is actually not much disagreement in the interpretation of the meager information we have to work with. However, there is no proposed location that is best indicated by all the methods of latitude discrimination. Therefore, one must consider how to weight the findings of the various methods so that multiple search zones can be prioritized. You took a binary approach (N or S of 25S). I think this is too crude to accurately represent what I think we know.
Second, you said: “Because possibility (3) requires pilot inputs after fuel exhaustion, if this occurred, we have to assume that pilot inputs also occurred during the powered part of the flight after 19:41. As such, there would be no reason to believe that paths reconstructed with the assumption of automated flight with no pilot inputs are representative of the path actually flown.”
I don’t agree with this statement, either. You don’t have to make that assumption. You can certainly have a pilot flying for a long period on autopilot and then manually flying after fuel exhaustion (alternatively, there could be other reasons why our inference of the 00:19 BFOs is flawed).
What I Think is Known :
1. The aircraft is very unlikely to be located within +/- 22 NM of the 7th Arc for latitudes south of 25S, and within +/- 40 NM for latitudes south of 33S.
2. The 00:19 BFOs appear to indicate a crash prior to 00:21, and therefore within about 22 NM of the 7th Arc. However, this interpretation is subject to several assumptions and could be wrong. In particular, it assumes no piloted glide after 00:19:37.
3. The great circle paths proposed by Godfrey and by Iannello (2016-2017) can match the satellite data, albeit with BFO errors larger than used by Inmarsat, although they are consistent with DSTG’s analysis. The latitude range is from 39.0S to 26.6S for 7 Hz RMS BFO error, with BFO errors increasing at the northern end of the latitude range.
4. The Cocos Island Flyby route proposed by Iannello (2018) ends at 22S and has similar BFO errors as the Great Circle routes.
5. The CMT route proposed by Ulich (2018) ends at 31.6S+/-0.4. It has the smallest BTO/BFO errors of any route flyable with a single, unchanged autopilot lateral navigation method or waypoint at 19:41. The BFO errors are consistent with Inmarsat’s more stringent criterion.
6. At least 8 drift studies have been done by many groups to infer 7th Arc latitude range:
Pattaiaratchi (2014, 2016): 28.3S-33.2S
Meteo France (2016): 25S-35S
CMCC (2016): 28S-35S
GEOMAR (2016): 14S-33S
NOAA (2016): 26S-38S
CSIRO (2016, 2017): 26S-38S
Godfrey (2017, 2018): <28S-37S (2017), 20.5S-34.5S (2018)
Nesterov (2017): 25.5S-34.5S
I would characterize these results by saying at least 6 out of 8 (all except GEOMAR and Godfrey 2018) are remarkably in agreement at both ends, with average latitude limits of 26.5S-35.6S. One of the studies, namely GEOMAR, is an outlier for the northern latitude limit (14S). Godfrey’s 2018 paper indicates a slightly more northern limit of 20.5S than the majority of predictions. Starting a new search north of 25S seems to be inconsistent with the predictions of almost all of the drift studies.
7. A new method, for which at present there is only one quantified result, is matching the temperature profile inferred by measurements of the barnacle growth patterns with time. Godfrey (2018) finds 25S-20S (from memory, Richard, so apologies if this is incorrect) as the best matching latitude range.
Potential Search Zones :
Consider three possible search zones, and give each area 3 points if it is the best match, 2 points if second best, and 1 point if poorest match. Then sum up the results for these three zones: A is 25S-39S, B is 31.2S-32.0.2S, and C is 20S-25S.
For each latitude discriminator, next rank the three zones from high to low in terms of best match:
Current negative search results: C, B, A (i.e., C is most likely, and A is least likely)
00:19 BFOs: C, B, A (but C is very uncertain)
If Great Circle Path: A, B, C
If Cocos Flyby: C, B, A
If CMT Route: B, A, C
Drift Studies: B, A, C
Barnacle Growth Temperature Profile: C, B, A
An unweighted average shows that Areas B (31.2S-32.0S) & C (20S-25S) are roughly equal and way ahead of Area A (25S-39S). Depending on how you weight each discriminator, either B or C is at the top. This is where we are right now.
I don’t think a binary choice between 20S-25S out to 22 NM versus 25S-39S out to 100+NM makes much sense. The probability density of MH370 being between 25S and 39S is certainly not uniform. There is a strong peak at 31.6S. It deserves to be given due consideration. It has almost all the drift studies and the sat data going for it, except the inferred discrepancy between 00:19 BFOs and an impact < 22 NM from the 7th Arc. North of 25S is OK (but not great) for the sat data, it is poor for the majority of the drift studies, and it is unknown how it will stack up against the 00:19 BFOs. It does have the appeal of being virgin territory for searching and possibly matching our current interpretation of the 00:19 BFOs (which could turn out to be wrong), but I am concerned about the disagreement with almost all the drift studies.
Extending the search out to 100 NM from 31.2S-32.0S is just 50% of the area from 20S-25S out to 25 NM, so we are not necessarily talking about just doing one or the other. My recommendation at this time is to focus only on those two areas. I don’t think searching out to 100+NM over the entire range from 25S to 39S is worthy of even a partial effort until the two other options are first completely carried out. Furthermore, pending some new insight, I don’t see a way to clearly prioritize between 20S-25S/22NM (Area C) and 31.2S-32.0S/100NM (Area B). Perhaps weather or logistical considerations might favor starting one before the other.
@Don Thompson,
Thanks for addressing the question of the IFE sending messages when on APU power. The pre-flight occurrence seems like a solid demonstration that this will occur if the APU continued running at 00:21 (which it certainly should for the alternate electrical configuration), and if the aircraft orientation allowed a line of sight to the satellite.
Air life wrote “my likely scenario is based on Tough Training, Airmanship and Experience”
Reads like bias confirmation.
Tough training: you’re choice
Airmanship: the crew of MH370 didn’t communicate, the aircraft ceased cooperative surveillance, 9M-MRO crossed active airways. Any posited right turn makes confliction more likely with traffic departing & approaching WMKK & WMSA.
Experience: contradicting data reported to investigators by the ANSP, and military surveillance.
@Richard Godfrey
“If MH370 followed a flight path out over the Andaman Islands, entering a holding pattern pending negotiations, followed by a flight around the south of Sumatra heading toward Cocos Islands (YPCC) or Christmas Island (YPXM), as you have proposed, then the simplest way to navigate would be to enter a couple of waypoints………….”
This is above my pay grade. I am really just a messenger.
I am pleased a destination to the North on the 7th arc is being given oxygen.
Using basic navigation (from the drift rate era) on the back of an envelope my thoughts were that around 21S or around 9S had merit as a possible ditching.
Your incredible work on all these in-depth calculations are to be applauded particularly if it helps find MH370.
Thanks.
@Andrew. The ATSB’s remarks as accredited representative. I have discussed this with the ATSB today.
Going back to the South African investigation the ATSB was requested to provide specific information about pilot’s licenses, so in that instance as an accredited representative its role was limited to what it was requested to do by the investigating authority. As to whether it will get to comment of the final report was left open.
However on the Safety Investigation Report the ATSB said the Malaysians had made clear that this was not a final report, though the ATSB did provide comment on it. I asked whether the ICAO would treat it as a Final Report in the Annex 13 sense and was told that no they would not since the Malaysians had made clear it wasn’t.
I asked whether it would be fair to describe the Safety Investigation Report as an interim final: agreed.
As to whether the ATSB was satisfied with the response by Malaysia to its remarks I could get no comment on that. My speculation is that, if dissatisfied, the ATSB like other accredited reps, would not be able to insist that its dissent be included in that Report.
My suspicions remain unallayed.
@Niels
@DennisW
Possible reasons for SCS Search (why Malaysia hid the IGARI turnback)
(1) Trying to hide something-
(a) military/radar capabilites; or
(b) known plot to divert MH370 (definite rumors of this); or
(2) Expecting to hear MH370 landed somewhere
(3) Incompetence/Pride/going with the flow of events without taking initiative
(4) Internal politics/not wanting to inflame politcal opposition by suggesting pilot deliberate hijacking, which the opposition was angry when Razak disclosed probable deliberate act on 15-March
(5) Various SCS “eyewitness” sitings (oil rig worker/fishermen/etc)
@ TBill
“lack of ELT signal upon crashing more suspicious”
Nope, absolutely not. As I have explained many times, I would expect the ELT to trigger in a ditching, any ditching. However the first transmission on 406Mhz, by design, occurs 50 seconds after triggering. By that time the antenna could be under water and hence the trasmission would be ineffective. There may be no satellite visible at the precise time of the first transmission and in that case there is a further delay until the second transmission.
@David
RE: “The ATSB’s remarks as accredited representative. I have discussed this with the ATSB today.”
Thanks for that. So we can look forward to Malaysia releasing further interim reports until they get around to publishing a ‘final’ report, as required by Annex 13? I suspect not!
The entitlement of an accredited representative to have its views heard (and published) is very clear under Annex 13: the State conducting the investigation ‘shall either amend the draft Final Report to include the substance of the comments received, or…append them to the Final Report’. In practice, that obviously doesn’t always happen, but I doubt many governments would have the political will to make representations to ICAO if their views were ignored. Even if they did, there’s nothing ICAO can do to compel the investigating authority to comply. Frankly, I think it makes a mockery of the UN conventions, but I guess that’s nothing new.
@Victor Iannello,
@Richard Godfrey,
Here’s another way to look at the choices for a new search area.
If you go 22S/22 NM, you are betting on three things that all must be correct: (1) a pilot made multiple maneuvers after 19:41, (2) a pilot did not glide after fuel exhaustion, and (3) our current interpretation of the events and BFOs circa 00:19 is roughly correct. You are also betting that almost all the drift studies are wrong.
If you go to 31.6S/100 NM, you are betting on two things that must be correct: (1) a pilot made no maneuvers from 19:41 to fuel exhaustion, and (2) either (a) our current interpretation of the EOF and the 00:19 BFOs is seriously wrong or (b) a pilot glided the aircraft after 00:19.
So 22S/22NM requires an active pilot long after 19:41, but not active at 00:19. On the other hand, 31.6S/100NM requires no active pilot from 19:41 until 00:19, but possibly one after 00:19.
@DrB: I don’t believe there is a strong peak at 31.6S latitude. In the Great Circles post, I included a file summarizing the BTO and BFO errors for constant Mach and LRC speeds. The minimum RMS BFO error of 3.4 Hz occurred at 34.3S latitude (due south), yet the RMS BFO error only increased to 4.6 Hz at 37.5S, and to 4.8 Hz at 30.6S. The BTO error over that same range was between 27 μs and 32 μs. These BFO errors meet the Inmarsat criteria. In fact, the due south path to 34.3S is very similar to the path that Inmarsat proposed.
You believe you have a methodology to constrain the crossing of the 7th arc to a small range of latitudes. You also believe there is a reasonable chance that the pilot provided no inputs until fuel exhaustion and then performed a long glide (after what was probably a steep descent). I place a much lower probability on both these claims.
Ultimately, it’s not my opinion that matters, but the opinions of those that might conduct and finance a new search.
@DrB said: So 22S/22NM requires an active pilot long after 19:41, but not active at 00:19.
That’s simply not true. The path ending at 22S latitude that included an alignment and flyover of Cocos Airport had no pilot inputs after 19:41.
@Brian Anderson
Re: ‘… the first transmission on 406Mhz, by design, occurs 50 seconds after triggering …’
The ELT transmission delay when triggered automatically is 30 seconds. The 50 second delay is when the device is triggered manually; the extra 20 seconds is because there is a 10 second self-test sequence followed by a 10 second period when the self-test result is displayed.
If you take Ethiopian Airlines flight 961 as indicative of a wide body ditching there’s about 10 seconds between the first substantial contact with the water (the left engine) and when the aircraft (or parts thereof) come to a stop. It is up for debate whether the tail section of the aircraft would have sank in the next 20 seconds.
@Victor Iannello,
Perhaps those maneuvers could be programmed before 19:41, but they occurred after 19:41. That was my point, that maneuvers occurred after 19:41.
@Victor Iannello,
You said: “I don’t believe there is a strong peak at 31.6S latitude.”
That is because you only considered a small subset of all lateral navigation methods and speed modes (Great Circle routes at constant Mach and LRC according to your counter-example).
I looked at all the possible nav and speed methods, and clearly the best single-setting route is 181.2 degree CMT at MRC, as shown in Figure 11 in
https://drive.google.com/file/d/1NM7c9Vs6dJgOczqnU0682Jqw5Mjp84-J/view?usp=sharing
CMT has a single sharp, very narrow optimum, as shown in Figure 11c. The great circle case has two broad optimum regions, as shown in Figure 11b. This comparison was done using the same fitting program, with only nav and speed methods being changed from run to run. I have high confidence in the results, especially for relative route comparisons.
You also misrepresented what I said about my case at 00:19. I don’t know whether or not a pilot glided after 00:19, nor do I know that the current interpretation of the EOF and BFOs then is correct, but for 31.6S to be correct, at least one of those must be true.
@Andrew. “….there’s nothing ICAO can do to compel the investigating authority to comply.”
As you say, no compelling. In this case though it could advise Malaysia that in the circumstances a Final Report is called for now. If that should be unsatisfactory or is withheld it could seek rectification publicly, applying embarrassment and possibly deterring others from what gives the appearance of a ploy.
Doing nothing implicitly is endorsement.
DennisW
You write, “…or, there were people in Malaysia who knew the plane was airborne for some time after IGARI, but did not make that info public until the Inmarsat data forced their hand.”
The first public reference to a possible air turnback by Malaysian authorities (based on military radar data, although that was not made clear at the time) was made by HH (and possibly Daud) at a press conference on Sunday, March 9, 2014 [Day 2], at 11:30 am Malaysia time. At that time, no one at Inmarsat had even yet begun to delve into the satellite data.
All of the above assertions are based on news articles available on the internet at the time and still retrievable today.
@DennisW @TBill @ikr @Neils
When MH370 went dark at IKARI and Malaysia and Vietnam ATC could not establish communication, it would be assumed it crashed into the SCS. Vietnam, Malaysia, and other nations responded as expected. Inmarsat did not check their data logs until well after the flight was reported missing. The turn back and radar story was not public or maybe not even understood. If you assume the aircraft just crashed or exploded, the world would expect a SCS primary search.
What surprised me was that Malaysia started a search at Straits of Malacca without any notice of how a crashed plane could get there!
Inmarsat privately advised Malaysia shortly after the plane was declared missing that it was in the air for 7 house after it went silent. This was not disclosed to public, but it clearly indicated that IKARI was not likely. 7 hours is a huge radius, so it may not have been sufficient to cancel search.
I don’t know when 7.hour info and radar information came together. It took Inmarsat a few days to define the arcs – it the 7 hours was a major early finding.
It would be great if the Malaysian government had a timeline for when they confirmed a turn back toward Penang and that it was in the air for 7 hours.
But there was really no reason to not start an Immediate search in SCS. This follows standard protocol. If they knew this was a hijack and didn’t want to advertise it, you would still let the expected search to begin.
It does seem stupid to have not stopped the SCS search earlier, although China didn’t help by claiming satellite detected debris in SCS.
@sk999
Explain the search in the South China Sea.
An air turnback would not be the way to describe a long diversion West over Kota Baru.
@sk999
Also, I have made the point several times that the radar data is really not necessary to determine the path from Igari to 18:25 range ring. The aircraft had to fly almost straight West at maximum speed to satisfy that BTO. The more intricate BFO data interpretation for an FMT to the South came later, but it did not take any analytical prowess to figure out that the path after IGARI had to be almost straight West
DennisW,
“Explain the search in the South China Sea.”
The LKP (last know position) to DCA was at IGARI, over the South China Sea. DCA was responsible for launching the search. What’s to explain?
“An air turnback would not be the way to describe a long diversion West over Kota Baru.”
Irrelevant. You claim that the radar narrative was forced by the Inmarsat data, whereas publically available news articles demonstrate the opposite. Why are you trying to rewrite history?
@Victor, @Don, @Andrew. Do any of you know answers to the following, I wonder?
1. What conditions, specifically, trigger an APU auto-start
2. What are the possible sources of power to pitot heat and are there two pitot tubes?
3. Could RAT deliver power for pitot heat?
Thanks in advance for any insights.
@sk999
Inmarsat knew or should have known that the aircraft flew West over the Malaysian peninsula in less than a few hours after the diversion. If the Malaysian claims relative to the radar data are valid they also would have known the direction the aircraft went from IGARI. There was no plausible reason to spend any time searching in the South China Sea.
The reality is the LKP was clearly West of Penang. The path after that took some time to unravel (four plus years and counting).
@Paul Smithson
“1. What conditions, specifically, trigger an APU auto-start?
Loss of power to both AC transfer buses while the aircraft is in the air.
“2. What are the possible sources of power to pitot heat and are there two pitot tubes?”
L & R AC transfer buses. There are three pitot tubes, L, C & R.
“3. Could RAT deliver power for pitot heat?”
No.
As regards to your earlier comment August 23rd at 02:13pm
http://thehuntformh370.info/content/malaysia-airlines-flight-370
Paragraph 6
Malaysia Airlines (MAS) issued a media statement at 07:24, one hour after the scheduled arrival of the flight at Beijing, stating that contact with the flight had been lost by Malaysian ATC at 02:40. MAS stated that the government had initiated search and rescue operations. It later emerged that Subang Air Traffic Control had lost contact with the aircraft at 01:22 and notified Malaysia Airlines at 02:40. Neither the crew nor the aircraft’s on-board communication systems relayed a distress signal, indications of bad weather, or technical problems before the aircraft vanished from radar screens.
Don Thompson says:
August 23, 2018 at 6:51 pm
Air life wrote “my likely scenario is based on Tough Training, Airmanship and Experience”
Reads like bias confirmation.
Tough training: you’re choice
Airmanship: the crew of MH370 didn’t communicate, the aircraft ceased cooperative surveillance, 9M-MRO crossed active airways. Any posited right turn makes confliction more likely with traffic departing & approaching WMKK & WMSA.
Experience: contradicting data reported to investigators by the ANSP, and military surveillance.
Don, it’s based on MY Tough Training, Airmanship, Experience, not others. I have looked both left AND right over 4.5 years, taking a balanced approach, whereas others have not, that’s confirmation bias. I’d say there has been too much BIAS directed against the pilots without a shred of evidence. ‘Innocent until proven Guilty’ is the cry. Finding the aircraft should be first priority and thereafter the post mortem, but it’s happening in the reverse order. Also, not everything that has been put in the reports can be believed. Cutting through all chaff isn’t an easy task, but it has to be done, AND so too narrowing down the options. Thanks.
@Joseph Coleman
Your quote states that “Subang Air Traffic Control had lost contact with the aircraft at 01:22 and notified Malaysia Airlines at 02:40.”
Local Malaysian time, obviously.
However, the Factual Information/SIR ATC voice transcript records a telephone call at 02:25MYT between ATC & MAS Ops where ATC notes an earlier conversation with MAS, “But earlier we checked with MAS I think your side somebody said that the aircraft still flying and you already send signal to the aircraft.“. At that point both ATC & MAS appeared unclear about the status of the aircraft.
@DrB
Please be careful how you represent the various drift analyses. There are multiple solutions that fit a drift analysis.
Pattiaratchi et al. of UWA only considered the area from 22.8°S to 39.3°S and therefore would not find a hotspot at 20°S.
Daniel of Meteo France considered a large number of areas from 7°S to 49°S but only 21°S and 23°S were in the vicinity of the 7th Arc.
Durgaddo et al. of GEOMAR had two favoured areas at 15°S and 25°S but neither were in the vicinity of the 7th Arc.
Griffin et al. of CSIRO only considered 27°S to 41°S and therefore would not find a hotspot at 20°S.
Trianes et al. of NOAA considered 15°S to 40°S and concluded only a wide area from 26°S to 38°S.
Jansen et al. of CMCC considered 15°S to 40°S and concluded only a wide area from 28°S to 35°S.
In terms of discounting a MH370 end point at 20°S, Pattiaratchi and Griffin can be ignored, because they didn’t even look.
Daniel of Meteo France and Durgadoo of GEOMAR both agree with my finding that 20°S is also possible.
Trianes of NOAA and Jansen of CMCC both disagree that 20°S is also possible.
A review of the drift analysis literature is not a clear cut as you make out.
Air Life says:
August 24, 2018 at 4:29 am
Don Thompson says:
August 23, 2018 at 6:51 pm
Air life wrote “my likely scenario is based on Tough Training, Airmanship and Experience”
Reads like bias confirmation.
Tough training: you’re choice
Airmanship: the crew of MH370 didn’t communicate, the aircraft ceased cooperative surveillance, 9M-MRO crossed active airways. Any posited right turn makes confliction more likely with traffic departing & approaching WMKK & WMSA.
Experience: contradicting data reported to investigators by the ANSP, and military surveillance.
Don, it’s based on MY Tough Training, Airmanship, Experience, not others. I have looked both left AND right over 4.5 years, taking a balanced approach, whereas others have not, that’s confirmation bias. I’d say there has been too much BIAS directed against the pilots without a shred of evidence. ‘Innocent until proven Guilty’ is the cry. Finding the aircraft should be first priority and thereafter the post mortem, but it’s happening in the reverse order. Also, not everything that has been put in the reports can be believed. Cutting through all chaff isn’t an easy task, but it has to be done, AND so too narrowing down the options. Thanks.
I may add, and that is exactly what IO has been doing too, narrowing down the location options or eliminating the location options. Is that confirmation bias? Absolutely not. By eliminating, or narrowing down the options, you get closer to the truth. By endlessly expanding the options/theories/causes, the further away you get from the truth. So, I asked myself, why would they turn back to KLIA? Answer: Turn back is about technical difficulties, it always is. Thanks.
@Andrew. Thanks for your earlier response.
I am still trying to think of a situation where AP might not drop out.
As I understand it, the generators on the engines are of a “variable speed” type so that the amount of power generated depends on the revs. If that’s the case, will a windmilling engine not generate some power, even if it’s not a lot? Last time this was discussed I think the consensus was that there would continue to be power generation for a limited period as the engine spooled down but that below a critical threshold it would stop generating power. I wonder what that time period and threshold are.
If the ELMS is doing its job properly, it should also be shedding successive systems in reverse order of priority as the generated power declines. That might conceivably also reduce the “resistance” on the windmill, enabling it to continue rotating and generating a little power.
Now, since this windmilling IDG is not generating enough power for critical systems, would APU not autostart – this being triggered by “too little power available” rather than zero power available?
If something like the above could be feasible, you would have a scenario whereby:
– AP remains engaged because pitot heat retained as a high-priority item on the little power being generated by windmilling engine. Also remain in “normal” rather than reverting to secondary.
– SDU is shed by ELMS when power severely limited under windmilling engine. Re-boots once the APU is up and running.
Alternative to windmilling is an engine that maintains revs above threshold due to repeated re-lights. You would still, however, need to have an APU re-start triggered by power upset rather than zero power. Enough re-lights to bridge a 2-minute gap seems to me a bit of a stretch, though, depending on how long a spooling-down engine keeps generating electrical power.
Is either scenario conceivable?
Perhaps it’s less than a 2 minute gap if we are only waiting for APU power to become available (rather than that plus SDU boot/logon sequence).
When I google RAT I find “the Ram Air Turbine provides hydraulic power to powered actuation of flight control surface and electric power for flight control computers, essential instruments, radio, cockpit lighting, windshield- and pitot tube de-icing”. Seems strange that pitot heat is not included on a 777? I’d have thought this would be pretty high on the list of vital systems.
@Air Life
If we were to add up, training years, flying hours and professional experience years for all contributors to Victor’s blog, we would all be astounded at the result.
We still have to argue our case and supply supporting evidence, so that others can transparently review our findings.
This reference
http://www.ieee.rackoneup.net/rrvs/06/Emergency%20RATs%20Presentation.pdf a
lso specifically says:
Emergency Ram Air Turbine
Purpose: provide emergency power to essential systems when all engines fail
• Hydraulic power for powered actuation of flight control
surfaces
• Electric Power for flight control computers, essential instruments, radio, cockpit lighting, windshield and pitot tube de-icing
@Paul Smithson
RE: “As I understand it, the generators on the engines are of a “variable speed” type so that the amount of power generated depends on the revs. If that’s the case, will a windmilling engine not generate some power, even if it’s not a lot? Last time this was discussed I think the consensus was that there would continue to be power generation for a limited period as the engine spooled down but that below a critical threshold it would stop generating power. I wonder what that time period and threshold are.”
No, the engine speed obviously varies, but the generators are driven by a constant speed drive to ensure constant frequency power; the combination of drive and generator is known as an integrated drive generator, or IDG. When the engine speed reduces below a certain point, the drive can no longer maintain the constant speed required by the generator, so the generator drops off line. In the simulator, it takes about 7-9 seconds for the generator to drop out after an engine failure.
RE: “Now, since this windmilling IDG is not generating enough power for critical systems, would APU not autostart – this being triggered by “too little power available” rather than zero power available?”
The autostart sequence is driven by voltage on the transfer buses. I’m not sure of the figure, but when the voltage drops below a certain threshold on BOTH transfer buses, ELMS automatically sends a start signal to the APU controller to start the APU. The voltage threshold is well below the normal 115V AC required by the equipment that is powered by the buses.
RE: “Is either scenario conceivable?”
I don’t have time to explain in detail, but the short answer is NO. I’ll try to explain in more detail later, if someone else doesn’t beat me to it.
RE: “When I google RAT I find “the Ram Air Turbine provides hydraulic power to powered actuation of flight control surface and electric power for flight control computers, essential instruments, radio, cockpit lighting, windshield- and pitot tube de-icing”. Seems strange that pitot heat is not included on a 777? I’d have thought this would be pretty high on the list of vital systems.”
I think the problem is that the really essential systems on aircraft such as the B777 consume a lot of power. The pitot heat, although important, is not as important as some other items. I guess the designers had to draw the line somewhere to make sure the loads do not exceed the RAT’s generating capacity.
Richard Godfrey says:
August 24, 2018 at 5:57 am
@Air Life
If we were to add up, training years, flying hours and professional experience years for all contributors to Victor’s blog, we would all be astounded at the result.
We still have to argue our case and supply supporting evidence, so that others can transparently review our findings.
Richard,
So far the supporting evidence and cases supplied by the official search teams have come up empty after 4.5 years of searching I’ve sent in my case and supporting evidence to the officials. Time to think outside the 7th arc box and look into it? I can’t do more than that. Thanks
@Air Life: I won’t engage you because you already said you have a preferred answer. (If you think technical failures haven’t been thoroughly studied, you are mistaken.) Also, like it or not, this is not a website where we dismiss scenarios because of “innocent until proven guilty”. This is not a court of law. It’s an investigation.
However, I can’t not respond to this statement:
So, I asked myself, why would they turn back to KLIA? Answer: Turn back is about technical difficulties, it always is.
Have you ever heard of 9/11?
@Air Life
Please see my post @TimR yesterday and then tell me I am not thinking outside the box.
@DrB also tried to prove yesterday that my Drift Analysis was an outlier (or outside the box).
Denial of evidence is not thinking out of the box.
@DrB: I have no doubt that with a particular set of assumptions, you get a narrow peak in your cost function. That doesn’t mean that other acceptable solutions with different assumptions that don’t exhibit a narrow peak are wrong. It comes down to the correctness of the assumptions. I am much less certain about the uniqueness of your solution than you are.
@Richard Godfrey, DrB
I encourage you to carefully check the update by David Griffin (August 10th 2018):
http://www.marine.csiro.au/~griffin/MH370/
David has extended the probability graphs and simulation/animations up to S8 degrees; see the links in his update.
His (extended) results do not support a hotspot around S20 degrees.
@Niels
Griffin makes so many mistakes and false assumptions in his various analyses:
1. Differences to Pengam 2016 not explained.
2. No ocean trials with genuine flaperon.
3. Windage results based on only a small selective data set.
4. Conclusions based on sporadic aerial search, claimed as being comprehensive.
5. Conclusions based on satellite imagery.
6. Conclusions based on one debris item out of 27.
7. …
ICYMI, MH370 is not where Griffin predicted, in study 1, 2, 3 , 4 or subsequent updates.
Please do not try and prohibit a search further north of 25S, because Griffin does not agree.
Victor Iannello says:
August 24, 2018 at 7:21 am
@Air Life: I won’t engage you because you already said you have a preferred answer. (If you think technical failures haven’t been thoroughly studied, you are mistaken.) Also, like it or not, this is not a website where we dismiss scenarios because of “innocent until proven guilty”. This is not a court of law. It’s an investigation.
However, I can’t not respond to this statement:
So, I asked myself, why would they turn back to KLIA? Answer: Turn back is about technical difficulties, it always is.
Have you ever heard of 9/11?
Victor,
So why do you engage me? 9/11 wasn’t turn back for technical reasons, but hi-jackings and diversion of aircraft to perpetrate terrorist acts on the Twin Towers in New York. The Petronas Towers weren’t affected! I stand by: “Innocent until proven Guilty” as far as the pilots are concerned, court or no court. It isn’t fair blaming them without concrete evidence. A right Turn back to KLIA with technical difficulties, imo, is the most likely scenario. If you have studied tehnical failures, then good for you, there are many as I have already pointed out. Furthermore, too few blips detected by Military Radars to be sure it was MH370 diverting across Malaysia and up the Malacca Strait. The 7/8/9th March 2014 was also the annual raptor migration event in Malaysia which complicates matters even further.
Richard Godfrey says:
August 24, 2018 at 7:26 am
@Air Life
Please see my post @TimR yesterday and then tell me I am not thinking outside the box.
@DrB also tried to prove yesterday that my Drift Analysis was an outlier (or outside the box).
Denial of evidence is not thinking out of the box.
Richard,
Drop your defensiveness. I wasn’t accusing you. Time to think outside of the 7th arc boX? It’s a rhetorical question. What evidence?
@Air Life
Stop playing psycho games!
When you write to me with accusations, then you are making accusations against me.
ICYMI, your comment was not addressed to anyone else.
@Air Life
Apparently, you have no regard for the ZS simulator data. Or do you believe ZS was particularly meticulous to practice a technical difficulty on the way to IGARI?
Furthermore, too few blips detected by Military Radars to be sure it was MH370 diverting across Malaysia and up the Malacca Strait.
You don’t need the radar data to be sure the aircraft flew West past Penang. The Inmarsat data is unambiguous in that regard. It is the only direction the aircraft could have flown based on very easy to interpret BTO data.
There is absolutely nothing in collection of information associated with the diversion that suggests an aircraft failure.
Richard Godfrey says:
August 24, 2018 at 9:36 am
@Air Life
Stop playing psycho games!
When you write to me with accusations, then you are making accusations against me.
ICYMI, your comment was not addressed to anyone else.
Richard,
What accusations? What psycho games? You’re making accusations, or reading too much into a rhetorical, challenging, question. I’m straight forward. That silly “psycho finger” of ours is point back at you!
@Richard Godfrey
My posting was triggered by your remark:
“In terms of discounting a MH370 end point at 20°S, Pattiaratchi and Griffin can be ignored, because they didn’t even look”
This is simply not true.
I agree with some of your critical points regarding the CSIRO work, especially where it concerns the “mixing” of aerial search results and satellite imagery with drift results. However, if you remove these “restrictions” I think it is valuable work which should not be ignored.
I have no interest to “prohibit a search further north of 25S”. I just feel that a decision to start a new search should be based on complete and state-of-the-art information.
I wouldn’t know where to search best based on the current state of available information. For example, I think it would be important to have a detailed update on the extent of “data holidays” and LPD areas in previously searched areas before taking a position where to search next.
@Air Life
Maybe you should read, what you write before pressing the send button.
@Niels
Whatever!
You are always right.
@TimR
@Niels
@all
I think @Niels may have orginally posted this sg.News.Yahoo article. Focus on the 527 comments, apparently many from Malaysians. It gives a feeling for public sentiment at the time (distrust of the government). Also more “chatter” about MH370 being a planned diversion.
https://sg.news.yahoo.com/hishammuddin-now-says-military-told-keep-eye-mh370-025118812.html
For exmaple-
Azura4 years ago
Razmatazz, my husband had told me from the beginning that this is a planned operation but I told him he is being ridiculous.But now after reading Hishamuddin’s statement and also what you have written, I think my husband is right after all.
But if they are safe, where can they be ? Where can you hide 239 passengers with so many countries searching for them ?
razmatazz4 years ago
A driver asked his boss who is a Minister about what has happened to the MH370 about 3 days after the incident..the Minister replied “Everything is according to plan” The driver then asked again…The Minister was speechless because he has said something which he shouldnt have said..He gave a stern look at the driver…and the topic was left hanging until now …A true story from the driver himself but he refused to reveal the name of the Minister..
>>So I do not know if FBI/CIA would say this above and @TimR’s account qualify as “chatter” but seems to me it is “chatter”. Additionally there was the reported email claim to one of the lesser-respected news outlets.
DennisW says:
August 24, 2018 at 9:52 am
@Air Life
Apparently, you have no regard for the ZS simulator data. Or do you believe ZS was particularly meticulous to practice a technical difficulty on the way to IGARI?
Furthermore, too few blips detected by Military Radars to be sure it was MH370 diverting across Malaysia and up the Malacca Strait.
You don’t need the radar data to be sure the aircraft flew West past Penang. The Inmarsat data is unambiguous in that regard. It is the only direction the aircraft could have flown based on very easy to interpret BTO data.
There is absolutely nothing in collection of information associated with the diversion that suggests an aircraft failure.
Dennis,
Simulator data? It’s worth JS if it can’t find the wreckage. So, if you’re so sure about the direction, military data and Inmarsat’s BTO data then why haven’t you found the wreckage in 4.5 years or to date? Also, I have never mentioned “aircraft failure”, but electrical failure/outages. BTW, a flight from KL to Perth is 5h30. A 7h00 flight after turn back would take it beyond Perth. You do the math. I don’t subscribe to 3h00 diversions that don’t make sense. Also, there were more than one geo-stationary satellite in play at the time of MH370’s disappearance from radar. So, until they sort out the mess or come clean, I’m firmly sticking to the timeline, it’s more precise.
@all
please don’t feed the troll
@TBill
I don’t remember having posted this article (Which is not a guarantee that I haven’t: nowadays it feels I need a memory extension)
@Air Life
Your interpretation of “turn back” mirrors sk999’s. To me there is a significant difference between a “turn back” and a “turn left”. The data can only be interpetated as a “turn left” (West). There was no indication of an attempt to return to KL. A search of the South China Sea after 7 hours of flight time after IGARI seems to assume the plane was flying in circles over that region.
Intelligent people would not make that mistake. It was a sham plain and simple.
The ZS simulator data clearly shows planning and intent. It has nothing to do with locating the aircraft.
Richard Godfrey says:
August 24, 2018 at 11:29 am
@Air Life
Maybe you should read, what you write before pressing the send button.
Richard,
You’ve been too long in the tetchy 7th arc box, get a life, get a new perspective.
@Air Life
Being the “gate keeper” for the whacko category I inform you that you are in.
@Niels
Below is your original post.
Important notes on the sg.News.Yahoo article “Keep and eye on” quote:
(1) I do not think Hish said (on 4 Corners video) that someone said to “keep an eye on MH370″…so I do not exactly understand why this media article indicates Hish said that….seems to be a bad press report again. So I am not able to verify that point.
(2) Note last sentence claims Razak started some Malacca Straits serach right away, which is also my understanding might be true.
So here you go below. Obvously I latched onto it in my memory, and later Victor challenged me on it because we really do *not* have any proof someone said to “keep eye on” MH370.
“Niels says:
February 11, 2018 at 12:30 pm
@Julia
Here is one link to the interview:
https://m.youtube.com/watch?v=2XQPCwY5SHk
In my observation Hishammuddin escapes the question about why there was no interception by going “ad absurdum”:
“If you’re not going to shoot it down, what’s the point of sending it (a fighter) up?”
See also:
https://sg.news.yahoo.com/hishammuddin-now-says-military-told-keep-eye-mh370-025118812.html “
@Air Life
Please do not try to give me advice.
You don’t even know your left from your right.
@Andrew thanks for your responses.
@DennisW
There was no coordinated information about any flying from IGARI toward Penang until well after the aircraft did not arrive at Bieging. While radar observation was being made people responsible for search only knew that all communication stopped near IGARI.
My understanding was the aircraft made a 270 turn to the right on radar and not a 90 left. But result is the same of a quick turn toward Penang.
If you believe (I don’t) that someone on the AC was in direct communication with the the person directly responsible for the search, then searching the SCS makes no sense except as a ruse. But without any confirmed information that the AC deviated from the approved route and did not arrive one would start a search along the route starting from the last transponder fix.
Malaysia has been terrible at laying out the timeline of what was known by whom when and how decisions were th3n made and by who. Inept or hiding something?
Inmarsat had no reason to check data until the AC was declared missing which was not un5il not arriving. This was when they checked the data records and noticed the 7 hours. There was no real time information provided by Inmarsat that night. There was no reason to not start a search at IGARI at first light.
When did Malaysia primary decision maker learn of flight toward Penang and 7 hours of fligh5 after IGARI? Which was discovered first by the person in charge. I don’t remember who was the actual person making the decisions and not just a talker.
Richard Godfrey says:
August 24, 2018 at 2:57 pm
@Air Life
Please do not try to give me advice.
You don’t even know your left from your right.
Richard,
I’m not trying to give you advice, your shortcomings are blatantly obvious.
I most certainly do know my left from right, unfortunately you do not, nor your light (UTC) from darkness (Local Time).
Good Bye!
@TBill
Thanks for digging that up; I remember it now in the context. Regarding your two points:
(1) The minister admitted that DCA contacted the military that morning (he avoids to mention a time), see around 9:30 in the youtube link. I agree that is not necessarily the same as what the article suggests.
(2) Difficult to verify when the Malacca Straits search was exactly started. If we may believe the Sunday press conference mentioned before, some searching at the west coast was ongoing on Sunday 9/3.
@AL: I shouldn’t pile on since Dennis has made his call. But “The 7/8/9th March 2014 was also the annual raptor migration event in Malaysia which complicates matters even further” at least adds to our collective store of facts..
Not Falcon and Eagle jets, but Falcons and Eagles!
@Niels
The military radar data showing a diversion toward Penang would have led to expanding the search to the Straits. The 7 hour information from Inmarsat would have made all of the local searches questionable. They asked Inmarsat to work the arcs after the new flight duration was recognized. They did not disclose the 7 hours for a while. Hijack may have been a concern at the time.
The decisions after Inmarsat initial air search work and then DSTG are well documented by a ATSB but MOT decision timeline is opaque in report. They merge the information in a timeline that was only incrementally known over days at the time. After the fact the searches in SCS and Straits were a waste, but real time decisions with conflicting information may have been reasonable.
When the news first broke I expected wreckage to be found near IGARI in shallow seas. I was mystified by search in Straits and blown away by the seven hour revelation.
@sk999
http://www.abc.net.au/news/2014-03-12/malaysian-military-denies-detecting-missing-plane/5314212
I estimate it would take me a 20 minutes using the Inmarsat data to know that MH370 flew West past Penang – 15 minutes to do the calculation, and 5 minutes to check it after being very surprised. This denial is dated 13 March. Inmarsat had to have known the plane went West of Penang by then. The search in the South China Sea ended March 18.
The question of info timing between Inmarsat and Malaysia is murky, but I am sure there are clues out there. I stand by my claim that the ISAT data forced Malaysia to come up with the correct (or fabricated to sound correct) narrative.
Why the South China Sea search went on to March 18 is an important question, that begs for an answer.
In the meantime wreckage was being dispersed at the impact location in the SIO.
@DennisW
You are correct in that there would be no reason for any SCS search after Inmarsat disclosed the 7 hours on March 8 or March 9. My point was only that it made sense to start in SCS at daylight on March 8. But not for long.
I don’t remember the date that Inmarsat provided the diagram of the north and south arc segments, but that should have ended any searches near Malaysia. They disclosed the 7 hours much earlier than the arc model by several days.
China did report debris by satellite in SCS which confused things because it was not consistent with SIO focus at the time.
@Hank
Inferring anything from the BFO is much more complicated. The approximate location of the 18:25 arc is easily derived from the last known position, BTO and time near Igari, and the BTO and time at 18:25. No need to get fancy and calculate BTO bias etc. My sense is the location West of Penang should have been realized before the end of the first day of searching in SCS.
DennisW,
The article you linked says the following: “Air force chief General Rodzali Daud was this morning quoted by Malaysian media …”
Have you searched back to the original in “Malaysian media”? I highly doubt it – the headline of the article, which is based on said “Malaysian media”, is sensationalist rubbish, as anyone who tried to track down its origins would realize. The original article appeared in Berita Harian (in the Malaysian language) and has long since disappeared. I spent quite a bit of time trying to track it down. If someone else here has a copy and can put a copy online, please do.
The following is my best reference to the earliest report of an “air turnback”:
http://english.astroawani.com/malaysia-news/
plane-may-have-made-air-turn-back-counter-terrorism-units-activated-31527
It is dated March 9, four days before your purported “his denial is dated 13 March.”
You also state:
“I estimate it would take me a 20 minutes using the Inmarsat data to know that MH370 flew West past Penang – 15 minutes to do the calculation, and 5 minutes to check it after being very surprised.”
Congratulations. Were you working for Inmarsat on March 8, 2014, with full benefit of hindsight?
@sk999
Congratulations. Were you working for Inmarsat on March 8, 2014, with full benefit of hindsight?
Oh please, you could do the same thing. I never worked for Inmarsat, and I don’t think you did.
The fact remains that something is very strange about the early search efforts, and it bears looking into. (And I am not trying to rewrite history.)
@DennisW
@sk999
At some point before 15-March (Razak press conf), it seems to me that Malaysian authorities developed a talking-point position to say the military observed an unidentified radar blip, not known to be MH370 until (a week) later, but known to be commercial flight (no threat perceived).
The recent SIR report, to my ears, surprisingly says, they knew all along it was MH370 turning back. Total contradiction.
@TBill,
In the link I posted the phrase “reciprocal route” was used. It is more definitve than “turn back”. Maybe sk999 will not think that is “irrelevant”. I don’t know what his problem is.
P.S.- I assume SIR report is more accurate, that they surmised within a few hours they had radar tapes of MH370’s turnback.
@Tbill
The reality is that the first thing any intellegent person would do, and I don’t doubt there are intelligent people at Inmarsat, is to compute the location of the 18:25 range ring. It is a trivial thing to do. At that moment it is “game over” for searching in the SCS.
@DennisW
Arc6 is near SCS so flight could have ended near there? Is that a stretch explanation for why Inmarsat did not over-rule SCS search at first?
@DennisW
@TBill
Dennis, as Bill has alluded to, wouldn’t the first thing that Inmarsat might do with the satellite data be calculate the BTO for the last transmission rather than the first. If they did that, lo and behold, that arc carves through the South China Sea a couple of hundred miles from the last known position of the aircraft based on the SSR data. You’d be a brave person to cancel the search in the SCS when, despite now learning that the airplane stayed aloft for 7 hours, you’ve got a piece of independent data that supports a crash in the SCS.
As to the availability of radar data immediately post-LOC I am firmly of the view that there is no Malaysian military radar data (to the extent there’s a ‘cover up’, that’s what’s being covered up). I think that the only radar data that the Malaysians had at their disposal until around 18-19 March was their own civilian radar data – the KL and KB SSR data showing the flight proceeding normally right up until it disappeared near IGARI after commencing its right turn toward BITOD, the KB PSR showing a target approaching from the NE and then heading away to the SW towards Penang and the Penang (Butterworth) PSR data showing a target approaching from the NE and then turning NW as it neared the southern tip of Penang.
In terms of accessing their own civilian PSR data I suspect that it would have been the KB data first and once they understood that then they would have called for the Penang PSR data. I think that process explains the delay in adding a search of the Malacca Strait to the SAR effort.
And if you consider the largely forgotten ‘other’ Lido slide – the one showing the fanciful 270-odd degree right turn at IGARI that hooks back to pick up the transit past KB towards Penang – I’d suggest that that is the product of trying to stitch all the then available radar data together into an ostensibly cogent picture. If they had radar data showing a turnback to the left when that presentation was put together, why show such a fanciful representation of the turnback? One possible answer is that they didn’t have any military radar data at that time (and I’d argue that’s because the Thais hadn’t yet given it to them).
We’re all captives of the tyranny of knowledge, we can’t unknow what we now know. Back then as things were unfolding off a somewhat shoddy start, the notion that the airplane might have crashed somewhere other than near where it dropped off radar would have been met with some scepticism, and rightfully so. We now place considerable trust in Inmarsat’s data and the various interpretations of it. Back then, however, if you’re running the SAR effort and someone says that we’ve got some ‘satellite guys’ from London on the phone who say the aircraft was still in the air at 7am, you’re realistically not just going to drop everything to factor in their input exclusively. Let’s not forget that you’ve previously had the airline’s ops people telling your ATC guys that the airplane was flying off the east coast of Vietnam per its flight plan at 1.30am. New and challenging data is going to be processed slowly and with caution and to the extent that any of it supports currently held views (eg the seventh arc carving through the SCS) you’re going to run with that first.
@Mick and TBill
You both make good points, and I suppose I am guilty of a personal bias here. If I had the ISAT data from the start my strong inclination would have been to ask where did the plane go after it dissappeared near IGARI. The 18:25 arc is a bit unique given that it can only be reached directly from IGARI at near maximum aircraft speed. It falls in the category of a certitude that the plane had to fly almost directly West. The BFO data was not needed. It is also unique in the sense that it represented a reboot – the ping occurred much later than it should have.
The path relative to subsequent arcs is not so deterministic as we all know given the struggles of the last four years, and BFO plays a much stronger role. Still, I acknowledge your arguments make a lot of sense.
@Hank
@DennisW
@Et Al
“You are correct in that there would be no reason for any SCS search after Inmarsat disclosed the 7 hours on March 8 or March 9. My point was only that it made sense to start in SCS at daylight on March 8. But not for long.”
With the knowledge available at the time it made sense to continue the SCS search. There could have been an explosion/collision that left debris near the turn-back area.
Mick Gilbert was spot-on when he suggested “We’re all captives of the tyranny of knowledge”.
@sk999 said: The following is my best reference to the earliest report of an “air turnback”:
http://english.astroawani.com/malaysia-news/
plane-may-have-made-air-turn-back-counter-terrorism-units-activated-31527
It is dated March 9, four days before your purported “his denial is dated 13 March.”
This BBC story is dated March 9, and includes the following: Air force chief Rodzali Daud said the investigation was now focusing on a recording of radar signals that showed there was a “possibility” the aircraft had turned back from its flight path.
About half-way down the page, Rodzali’s full statement can be heard in the video.
Here is a more complete version of the press conference, including more questions from reporters about the turnback.
https://www.youtube.com/watch?v=HKRkm11RyD0
@Air Life
You stated “I’m not trying to give you advice, your shortcomings are blatantly obvious.”
I agree with you, I have many shortcomings, failings and weaknesses. In fact, quite human and almost normal.
@Mick
Also one of the articles above is talking about a lot of radar assets near IGARI that we nener talk about, including Malaysian radar ships monitoring the shipping lanes, Singapore etc.
There is a very interesting exchange starting at 22:11 of this news conference on March 12. The Chief of Army, General Zulkifeli Mohd Zin, explains his version of the radar saga. (He is clearly distressed.) He is emphatic that MH370 was not tracked in real time. A reporter asks a question about whether Malaysia will release all radar data. Hishammuddin states that “If FAA and NTSB can confirm that this flight from the raw data from the military is the flight that we are concerned, tomorrow I will release it [all the radar data]”. Then at 28:40, a reporter asks a question about why there was no intercept of the unidentified radar target. The reply, this time from Air Force Chief Rodzali Daud, is that the radar operator classified the target as a non-hostile civil aircraft, so there was no response.
So, in the same news conference we have from two different generals that MH370 was NOT tracked in real time, and later that it WAS classified in real time as non-hostile and therefore was not intercepted.
https://www.youtube.com/watch?v=pWh2MGgq6Ak
Victor Iannello said : There is a very interesting exchange starting at 22:11 of this news conference on March 12.
From France, the link you gave (youtube HKRkm11RyD0) connects to a 14min57sec long video. Are you sure of the link?
Regards,
PL
@Victor
Well done for researching out this evidence.
The Malaysians could not keep their fabrications aligned, even at the highest level!
@Pax Lambda: Thanks. The link should be correct now.
@Victor
In combination with the SIR report, it seems to me they are saying that MY *human eyes* did not see MH370 in real time.
However, their radar system did see and record MH370 and, if I understand the SIR, their radar system somehow knew the aircraft was coded as MH370 (commercial) even after the turnback.
That would seem to imply either (1) their radar computer system is smart enough to fill the 10-minute KB radar data gap at IGARI and still know it was MH370, or (2) if @Mick is wrong about no military radar, maybe there was MY military radar data that their computer system used to connect the dots and keep the radar blip ID’ed as MH370 (commercial flight – no concern).
The implication to me is MH370 probably got totally away before they knew it. I had thought maybe by 18:22 they were looking for MH370 in real time around VAMPI.
PS- Another implication is the pilot perhaps knew how to outsmart the MY radar logic system and keep the flight ID’ed as commercial
@TBill: If the military radar picked up the turnback after IGARI as claimed, there would be no reason to doubt the target crossing the Malay peninsula was MH370. I suspect what they call the military data is the primary civilian data supplied to the military. The confusion about whether the unidentified targets were MH370 might stem from the uncertainty in associating the final SSR target near IGARI with the first PSR target as it approached Kota Bharu.
@Victor
…but the unanswered question is- why was the incoming target at KB classified as commercial no threat in real time?
SIR seems to say their radar system knew the attributes of MH370 and was able to know it was MH370 coming back at KB. So then you could argue, upon replay, the people had human doubts, even though the system got it right.
@TBill: Why do you think it was classified as non-hostile? Because the military says it was? It is a very obvious lie.
@VictorI, TBill
To add to the confusion:
In the 4 corners interview the minister of defence states that an aircraft was detected in real time, but it was not identified as MH370, see around 6:00 in
https://m.youtube.com/watch?v=2XQPCwY5SHk
So, what to think about the statement in the SIR p.19/20? I have serious doubts. After seeing several press conference / interview videos I get the impression it took time afterwards to realize and confirm they had been looking at MH370.
From the SIR:
“On the day of the disappearance of MH370, the Military radar system recognised the ‘blip’ that appeared west after the left turn over IGARI was that of MH370. Even with the loss of SSR data, the Military long range air defence radar with Primary Surveillance Radar (PSR) capabilities affirmed that it was MH370 based on its track behaviour, characteristics and constant continuous track pattern/trend. Therefore, the Military did not pursue to intercept the aircraft since it was ‘friendly’ and did not pose any threat to national airspace security, integrity and sovereignty.”
@Victor
I agree with you. I am convinced that there was no real time military radar available. The military looked at recordings of the civilian primary surveillance radar.
Civilian radar operators looking at the real time civilian primary surveillance radar concluded the target was MH370, that was flying back over Malaysia and out over the Malacca Straits. As they were convinced that is was MH370, they did not consider the aircraft hostile. This was the wrong conclusion in light of 9/11. A commercial airliner that was flying back toward Malaysia, without any distress signals, should raise an alarm. There is always a possibility of a hijack taking place and that a 9/11 style terrorist act could happen.
As it turns out, there was no 9/11 style terrorist act, but it would still have been helpful if jets were scrambled to intercept MH370, as we would know more about the direction, speed, altitude, final turn south, pilots in the cockpit, etc.
Perhaps there were no jets available to be scrambled, just as there was no military radar.
Western Hill (military) radar at 5.424622°N 100.250972°E is 243.521 nm from the last radar point at 18:22:12 UTC at 6.5776°N 96.3408°E and still within the nominal range of 250 nm (the range will be further due to the altitude of the radar station at 814 m above sea level).
Butterworth (civilian PSR co-mounted with civilian SSR) radar at 5.472106°N 100.394621°E is 250.972 nm from the last radar point and at a nominal range of 250 nm, just disappearing out of range.
I conclude that it was the civilian PSR at Butterworth that captured MH370 at 18:22:12 UTC and not the Western Hill military radar.
I also conclude that there were no military aircraft on standby at 02:22:12 MYT to be scrambled if required.
The Malaysian authorities are just smoke and mirrors.
@Niels: Even if the military identified MH370 in real time (which General Zulkifeli was adamant that they did not), the claim that they did not send an intercept because they “knew” it was friendly is just plain silly, to put it mildly. At the minimum, they are hiding the inadequacy of their defense systems.
@Niels
The 4 Corners is very unlcear, except I like when Hish says you’ll get the same talking points from PM Razak (of course we will).
So you two are saying SIR above is false info. I am saying that explains why the system kept track of MH370. I would suggest the pilot perhaps knew how to accomplish that behavior.
To me the SIR is possibly correct, and that means I think the statement about not knowing the blip was MH370 was a little white lie that was agreed to in advance. The purpose of that white lie could be to cover-up something or possibly to defray criticism for SCS search and account for why they kept the turn-back secret.
Or alternately, for some reason, upon playback, they did not agree with the software that told them it was MH370 coming back in. Seems unlikely though obviously it was MH370.
@Niels
From the SIR (SLIGHTLY MODIFIED):
“On the day of the disappearance of MH370, the KOTA BHARU CIVILIAN PSR radar system recognised the ‘blip’ that appeared west after the left turn over IGARI was that of MH370. Even with the loss of SSR data, the KOTA BHARU CIVILIAN PSR long range air defence radar with Primary Surveillance Radar (PSR) capabilities affirmed that it was MH370 based on its track behaviour, characteristics and constant continuous track pattern/trend. Therefore, the Military did not pursue to intercept the aircraft since it was ‘friendly’ and did not pose any threat to national airspace security, integrity and sovereignty AND HAVING IGNORED EVENTS LIKE 9/11 AND AS NO ASSETS WERE AVAILABLE ON STANDBY”.
@Richard Godfrey: I don’t think the Butterworth PSR has range to reach past MEKAR, as it is approach radar for the airport.
I also note that the Malaysia is reluctant to release the military radar data before “calibrating” it with all the other radar sources. What the hell does that mean?
@Victor
If the radar capture 10 nm beyond waypoint MEKAR is not from Butterworth, where do you think the radar source was?
General Rodzali Daud’s famous denial that he never said that it was Butterworth that captured MH370 flight back over Malaysia and out into the Malacca Strait, makes it all the more believable that it was Butterworth.
Butterworth is an Air Force Base and not a civilian airport. Could it not be that the radar is not a typical approach radar.
@Richard Godfrey: According to this DCA document: A 50 NM Terminal Approach Radar with co-mounted 250 NM conventional SSR located one each at Butterworth, Labuan and Kuantan Air Forces Bases
I don’t know about the 18:22 data point. It could be from Western Hill, as most believe. It could be from Thai radar. It could be a “mobile radar asset”, be it airborne or on the sea surface. Or it could be fabricated. I’m open to all possibilities.
@VictorI, Richard Godfrey
“the claim that they did not send an intercept because they “knew” it was friendly is just plain silly, to put it mildly”
It is clear this is a sensitive point. When pressed on this point in the four corners interview, the conversation takes strange turns with the minister starting to talk about “shoot down”, see 7:40 – 9:30 in:
https://m.youtube.com/watch?v=2XQPCwY5SHk
It is possible that there were no assets available on stand-by and they don’t like to admit. Remember also that RMAF Butterworth is “home to the Headquarters Integrated Area Defence System (HQIADS), part of the Five Power Defence Arrangements for Malaysia and Singapore, which is commanded by an Australian Air Vice Marshal.” (Wikipedia). So the embarrassment could be significant.
@TBill
I tried to convey what is my impression. There is not much we can prove here, except that there are clear inconsistencies in the different statements from Malaysian side.
@Niels: Yes, most of us remember that interview quite well. It demonstrated how defensive and evasive Hishammuddin was. It might be because the ineptness of Malaysia’s military was on display for the world. It might be more.
To further belabour radar.
The Malaysian’s had 2, and most likely 3, ‘pictures’ available from air surveillance assets on the night of 8th March (local time): A) the DCA civil ATC radar; B) the RMAF operated ATC radar (at WMKB); and the RMAF air defence surveillance network.
Assets included in A) include the radar heads at WMKK, Genting Highlands, and WMKC: a mix of SSR only and combined SSR/PSR systems.
B) is the NEC combined SSR/PSR on the aerodrome at Butterworth serving both Butterworth and Penang International.
C) includes the RAT-31DL on Western Hill and the S-743 Martello at Bukit Puteri. I suspect the Martello did not contribute anything.
Assuming readers are satisfied with the SSR reports until 17:20, I’ll go to the Factual Information part of the 2015 Interim Report and the 2018 SIR that show tracks from plots captured by the terminal area PSRs at Kota Bharu/WMKC and Butterworth/WMKB, groups A and B, above. These PSRs have limited range, stated to be 80NM and 60NM, as the primary function is terminal area control under the transition altitude defined for the area. Both reports depict tracks determined to be 9M-MRO. The Kota Bharu PSR investigation was initiated by DCA at 05:34MYT on 8th during a call with AAT, DCA’s radar tech services provider.
Now C), the ADS network. The RAT-31DL also comprises dual, SSR and PSR, antennas on its head. A MSSR antenna sits above the 3D PSR planar antenna. It has the ability to interrogate using SSR in ‘all available modes’. The manufacturer’s description for the performance of this site states a range compatible with tracking 9M-MRO throughout its path from departure to IGARI (with SSR confirmed identity), the turnback, across the peninsula to Penang Island, and out over the Straits of Malacca. These type of systems include a plot combiner so that SSR and PSR targets can be correlated. I’ve described before how the Sentry C² system, that integrates the ADS network, can also integrate with the civil network to corroborate assigned SSR identity. If the RAT-31DL executes Mode-S all-call interrogations, it receives the 24-bit a/c address, 4 digit ‘squawk’ code, and baro altitude replies: with that information it classes a target as ‘friendly’. After loss of SSR returns the RAT-31DL may continue to associate the PSR plots with the previously identified target. This is the capability Selex and Raytheon describe for their systems. However, following the loss of SSR, the surveillance reverts to PSR and plots are subject PSR’s inherent precision issues.
It is scandalous that Malaysia has never adequately clarified the issue of ‘military radar’. Scandalous that Malaysia, after five days was still in headless chicken mode.
In the press conf clip referenced by VI, above, the Chief of the Armed Forces, General Zulkifeli Mohd Zin, does state the track was determined from a recording. The PC clip was recorded on Weds, March 12th. The previous day, 11th March, the press sectretary to Najib’s Prime Minister’s Office, Datuk Tengku Sariffuddin Tengku Ahmad, strenuously contradicted remarks made by the chief of air force, Rodzali Daud, that the ‘turn-back’ had then extended across the Malaysian Peninsula and over the Str of Malacca (NY Times, Guardian, Malaysian Insider). Sariffuddin’s statement appeared to allow only the turnback as certain, at that time.
@Mick
You mentioned an image depicting a 270º turn. If it was this image, that was created to support the NoK’s presentation back to Malaysia (my annotation translating the title), I understand that it was not part of any Malaysian material.
@Richard Godfrey,
You said: “@DrB also tried to prove yesterday that my Drift Analysis was an outlier (or outside the box). Denial of evidence is not thinking out of the box.”
It is unclear to me exactly what you meant. Yes, I said your result was different from most of the drift study results. If you want to call that an outlier, I won’t argue the point. But being an outlier is quite different from thinking out of the box. I don’t consider any of the drift studies, including yours, as thinking out of the box, except possibly Pattiaratchi, because he was the first to do it and he assisted Blaine Gibson in determining where to search for debris. They all use similar methods and the predictions vary. That is to be expected since they don’t all use the same assumptions or the same drifter data. You seem to be saying that I am denying evidence, if that last comment is directed to me instead of Niels. If you do think that of me, please give me an example. I included all the drift studies of which I am aware. That is certainly not a “denial of evidence.”
I am glad you listed the latitude limits for the drift studies. That is helpful in understanding the results. Still, the majority of results to date favor latitudes between 26S-36S. Only 3 of 8 studies indicate a “possibility” north of 25S, whereas 5 of 8 indicate it is more likely to be south of 26S.
You also criticized Grffin’s drift study by saying “2. No ocean trials with genuine flaperon.” That’s not a valid criticism of Griffin. The flaperon is not his to use. Besides, you could say the same thing about all the drift studies, including yours. Why didn’t you use the real flaperon? No French prosecutor in his right mind is going to let the flaperon be used by CSIRO in a long-term sea trial for fear of damaging the evidence.
You further criticized Griffin’s results by saying: “4. Conclusions based on sporadic aerial search, claimed as being comprehensive. 5. Conclusions based on satellite imagery.” On this point we agree. I have only used Griffin’s drift study predictions, unfiltered by the aerial search and satellite data assumptions he made. That gives an apples-to-apples comparison with the other drift studies, which is what I have done.
@Don Thompson
G’day Don, yes, that is the slide I was refering to. My understanding is that that slide was shown to the NOK on 21 March 2014 at the Lido Hotel in Beijing as part of the official briefing. I think that slide had to be part of the official Malaysian presentation.
@Don Thomspon
Therefore, do you feel the SIR statement below is probably correct?
And if SIR is correct, then that suggests the historic rhetoric from Malaysia that they did not know the blip was MH370 turning back, was not true?
Seems like we all agree Malaysia is talking about what the recordings showed.
I am going to stick my neck out and suggest MH370 made it to MEKAR undetected in real time.
From the SIR:
“On the day of the disappearance of MH370, the Military radar system recognised the ‘blip’ that appeared west after the left turn over IGARI was that of MH370. Even with the loss of SSR data, the Military long range air defence radar with Primary Surveillance Radar (PSR) capabilities affirmed that it was MH370 based on its track behaviour, characteristics and constant continuous track pattern/trend. Therefore, the Military did not pursue to intercept the aircraft since it was ‘friendly’ and did not pose any threat to national airspace security, integrity and sovereignty.”
@Victor Iannello,
You said: “@DrB: I have no doubt that with a particular set of assumptions, you get a narrow peak in your cost function. That doesn’t mean that other acceptable solutions with different assumptions that don’t exhibit a narrow peak are wrong. It comes down to the correctness of the assumptions. I am much less certain about the uniqueness of your solution than you are.”
I am not certain the 181.2 degree CMT route is the one flown by 9M-MRO, but here are some of the arguments why I think it is the most probable route:
1. The sharpness of the objective fitting function (what you call the cost function) is remarkable. It is only 0.1 degrees wide, whereas for great circles the width is about 5 degrees. Thus the CMT fit is 50X sharper. I spent three weeks adjusting the track angle in very small steps to make sure there was only one peak and that I did not miss it by making steps that were too coarse.
2. There is a distinct connection of this path prior to 19:41 to nearby waypoints and prior procedures. I first fit the route from 19:41 onward. Then I generated a “back track” to earlier times (as far back as 19:00). Then I found the magnetic deviation for that area. It was -1.2 degrees, meaning that the true track approaching 19:41 was 180.0 degrees. That means that the aircraft was flying due true south when the CMT commenced (perhaps using TRK HLD via the MCP circa 19:28). Next, I identified a pair of nearby waypoints that lay on a 180.0 degree true N-S line (AGEGA and BULVA). Now the “back track” path prior to 19:41 did not pass directly over AGEGA-BULVA. It lay to the west by 14.1 NM. In other words, prior to 19:41 the plane would have been flying ~15 NM to the right of AGEGA-BULVA. Sound familiar? That’s the same 15 NM right lateral offset called out by the Malaysia Airlines Contingency Procedure for Lost ATC Clearance and the same 15 NM right offset that fits the 18:25-18:28 satellite data very well. Perhaps the pilot was mimicking the contingency procedures so as to appear like a non-threatening civil flight with technical issues, thereby reducing the chances of a military intercept.
3. The best-fit speed is MRC, and the best-fit altitude is FL336, which are both quite reasonable. Another step in the Contingency Procedure is to change the flight level by 500 feet, and this is also consistent with the best-fit flight level being essentially FL335 (I.e., not an increment of 1,000 feet, but in between the normal flight levels in this area).
Regarding the correctness of my assumptions, I will point out that this route has the minimum number of assumptions among all proposed solutions. That doesn’t prove it is correct, but this has the fewest chances to get an assumption wrong.
@TBill
RE: “I am going to stick my neck out and suggest MH370 made it to MEKAR undetected in real time.”
I think that’s most likely true and that all the contradictory statements made by Malaysian officials are simply an attempt to cover up the fact they were caught with their pants down that night. How can they expect anyone to believe the military identified the aircraft in real time and “did not pursue to intercept…since it was ‘friendly’ and did not pose any threat to national airspace security, integrity and sovereignty”? If that were true, why did Malaysia persist with searching the South China Sea for days after the event? Did the military ‘forget’ to tell anyone? I think the real story is that the turnback was captured on radar and recorded, but not analysed until some time later.
@Niels
RE: “It is possible that there were no assets available on stand-by and they don’t like to admit. Remember also that RMAF Butterworth is “home to the Headquarters Integrated Area Defence System (HQIADS)…”
Absolutely. I think it’s extremely unlikely the RMAF had aircraft on alert, especially in the early hours of a Saturday morning. The threat level in the region is low and there would have been no reason for them to have aircraft and crews on 24-hour alert. IADS was set up to coordinate military action in the event the Malay Peninsula ever came under attack, but it is not NATO. They regularly exercise the military forces that make up the FPDA, but they don’t normally have forces on instant readiness.
@Paul Smithson
RE: ”If something like the above could be feasible, you would have a scenario whereby:
– AP remains engaged because pitot heat retained as a high-priority item on the little power being generated by windmilling engine. Also remain in “normal” rather than reverting to secondary.
– SDU is shed by ELMS when power severely limited under windmilling engine. Re-boots once the APU is up and running.
Alternative to windmilling is an engine that maintains revs above threshold due to repeated re-lights. You would still, however, need to have an APU re-start triggered by power upset rather than zero power. Enough re-lights to bridge a 2-minute gap seems to me a bit of a stretch, though, depending on how long a spooling-down engine keeps generating electrical power.
Is either scenario conceivable? .”
As I said previously, the short answer is NO. The following is a more detailed explanation.
The engine and APU generators are rated at 120 kVA and produce power at a constant 115 V AC, 400 Hz. To that end, the IDGs are driven at a constant RPM by a constant speed drive (CSD) over a wide range of engine RPM. However, the CSD can’t drive the generator at constant RPM once the engine RPM drops below about 50% N3, so the generator control unit (GCU) trips the generator off-line. Essentially, the generator output is either 115V AC 400 Hz, or nothing; there is no gradual decrease in output as the engine RPM decreases. The APU generator is similar, except that it is driven at a constant speed by the constant RPM of the running APU.
Two operating generators provide more than enough power to operate all the aircraft’s electrical loads; however, a single generator cannot do so. If the total load exceeds the available power, ELMS automatically sheds specified loads in sequence until the total load is within the capacity of the operating generator. The following loads are shed in sequence:
– Galley loads
– Utility buses
– Equipment cooling vent fan
– Galley chillers
– Recirculation fans
– Lavatory/galley fans
– Electronic seat equipment (ie IFE)
– Hydraulic pumps.
In the alternative electrical configuration scenario, only the right IDG was operating before the right engine flamed out. Consequently, ELMS would have shed some or all of the above loads until the total load was within the capacity of the operating generator. Upon engine flameout, that generator would have continued to produce power for about 7-9 seconds. The GCU would then have tripped the generator off-line as the engine RPM continued to decrease. At that point, the voltage on the AC buses, including the transfer buses (assuming the left backup generator was selected OFF), would have dropped to zero. ELMS would have automatically sent a start signal to the APU and, after 15 seconds, automatically sent a deploy signal to the RAT.
The loss of power to both AC transfer buses would have caused the loss of a number of systems and sensors that are required for the operation of the PFCS in normal mode. Consequently, the PFCS would have automatically reverted to secondary mode, resulting in the loss of the AP and TAC, etc. In the past it was thought that the mode reversion was caused by the loss of pitot heat. However, recent advice indicates the reversion is caused by the loss of other sensors required by the PFCS. The loss of pitot heat would only cause a mode reversion if the system detected invalid air data.
@DrB: You are obviously placing a lot of weight on the sharpness of the objective function as a criterion for testing assumptions and prioritizing routes. In the end, the validity of a path is only as good as the underlying assumptions, even if the objective function is sharp, and even if the number of assumptions is less than for other scenarios.
Independent of the sharpness of the objective function, we have no reason to doubt the final BFOs, and we have no reason to doubt the end-of-flight dynamics from the Boeing simulations, although we cannot be absolutely certain that either is correct. Assuming both are correct (which I think is our best estimate), your scenario requires a pilot to select track hold for a long distance trajectory, not provide inputs for hours, run out of fuel, enter in a steep descent, then glide a long distance. I question the realism of those assumptions.
I won’t try to persuade you that your assumptions are incorrect, but I am much less persuaded than you are that your solution is more likely than others.
@DrB
In your last post the term “best fit” is used frequently. I have long ago (from the get-go actually) rejected the notion of “best fit” being a valid route discriminator. The intrinsic uncertainty associated with the ISAT data (BFO) does not support the use of “best fit”.
I think attaching more weight to the softer issues of motive, intent, human factors,… is a better approach at this time. The pure analytics have simply run out of gas. Nobody is taking them seriously anymore.
At this moment I am inclined to support Victor’s path toward a Cocos landing.
@Andrew
So I think the real story may be that ZS knew he could turn off the transponder at IGARI and make a sharp U-turn, and by-pass the military alert system right under their eyes. Total inside job. Next morning business as usual in the air, but Malaysia is having a 9/11 crisis…what’s going to happen next? More AWOL planes? Negotiations? Razak keeps a lid on it, and it turns out working in his favor to have a great mystery instead of what really happened.
PS- The reason it is such a mystery, ZS was thinking everyone would know he was the only one who could do that deed, and why (Anwar trial outcome). Crazily, Razak did not blame ZS so everyone is mystified who done it and why.
The radar data has never held any interest for me. The path from IGARI to the 18:25 range ring is very deterministic. Who cares about details of the turn West? I did get involved with the Kota Bharu radar discussions just for fun (and it turned out not to be fun), but not because I thought they would produce anything of value relative to a terminus.
At this time, I think we all need to take a step back and contemplate the “why” of this event.
@DennisW
“The pure analytics have simply run out of gas. Nobody is taking them seriously anymore.”
Refreshing thinking Dennis, and long overdue, in my view.
May I now respectfully suggest, that we discard the Malacca Strait straitjacket / blinkers, and instead, take the deterministic performance limited 18:25 arc segment, from it’s most northern possible limit, to it’s most southern possible limit, and then apply the softer issues of motive, intent, human factors, and plausible “why’s” ?
I won’t push my route / reasons again, I will just sit quietly on the sidelines, and wait and see what transpires.
@Victor – Thank you for all the detailed analysis on this blog. Just wanted to point out that the blog is showing as being “Not secure”. I don’t recall this being the case before. Maybe some server permissions or patches are causing this issue. Wanted to make sure you are aware.
@DennisW said: “At this time, I think we all need to take a step back and contemplate the “why” of this event.”
Yes. ’bout time! And frankly I don’t see that Shaw had much reason to do this. I do see that the Uyghurs had plenty of reasons to do this and the Chinese had plenty of reasons to stop them.
@ST
Concerning ‘not secure’. That’s a warning generated by your browser that a website is hosted by an HTTP server that does not use SSL encryption.
@TBill wrote “I am going to stick my neck out and suggest MH370 made it to MEKAR undetected in real time.”
I won’t argue with that. I’ve attempted to explain how the air defence surveillance system identifies a target, how it can combine PSR & SSR to track the target. It’s quite possible that 9M-MRO never reverted from a ‘friendly’ categorisation to ‘suspect’ in the ADS system after -MRO’s transponder ceased. There should be no doubt that an aircraft departing WMKK, under ATC supervision, with an ATC assigned ident code will be categorised as ‘friendly’ in the ADS system.
DCA’s ATCC staff didn’t track MRO after 17:21 when -MRO’s transponder ceased, at that time no DCA PSRs had coverage. ATCC called upon the services of AAT at 0535MYT on Saturday 8th to recover their systems’ recordings. Analysis of PSR data from WMKC produced the two tracks either side of WMKC.
It remains astounding that, apparently, the military continued to struggle with the task of discriminating -MROs path even on Wednesday, day 5, as shown in Awani clip referred above. It’s a straightforward task: remove the ‘dots’ for all the targets for which an ID was determined by SSR interrogation.
Both functions in the military, ATC (operating at WMKB) and ADS, tracked -MRO & eventually produced data. Our request still stands, for that data to be put on the table.
Mick: I’m quite certain that image was not part of the Malaysian material presented on 21st Mar. I have a trail (somewhere!) with a record of the date.
@All
It would be nice to see the whole radar clip or images of all PSR detections/hits for the SOM track especially from Vampi onwards to determine how they picked out 02:2212, for the public to be fully convinced it wasn’t fabricated, misjudged or genially mistaken for something else in that recording period.
@Paul Smithson
@Andrew
@Victor
As someone who hasn’t gone down the nefarious pilot route, I favour an alternative power setup where the L BU gen hasn’t been switched off but is powering the R transfer bus. Therefore, the autopilot, RAT and primary flight controls only drop out on the left eng failure.
I thought, the general consensus was before this thread started that loss of control was after the second(left) engine failure. Why has it changed to loss of control after the first eng fail?
@Don Thompson
G’day Don, here’s an article dated 22 March 2014 from the UK Mirror featuring that slide;
https://www.mirror.co.uk/news/world-news/missing-malaysia-airlines-flight–3272635
@Mick Gilbert, @Don Thompson: On March 16, 2014, Chinese media published an interview with a Malaysian air force officer in which he disclosed that MH370 passed IGARI and turned back, flying low to avoid radar detector. the turn after IGARI was a clockwise loop. A video was produced by the Chinese to illustrate this path. (The link I have for the video no longer works.)
Interview: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=http%3A%2F%2Fwww.soulu365.com%2Fa%2Fnews%2F32582.html
Video: http://www.zhuatieba.com/video/XNjg1ODM0ODUy
On March 22, 2014, Malaysian officials brief the next-of-kin (NOK) at the Lido Hotel in Beijing, but leave the room without taking any questions. After the Malaysian officials leave, the NOK continue with questions, and project a slide showing the path of MH370 with the clockwise loop shown in the Chinese video. This enrages the NOK as there are discrepancies in time and path from the radar data shown to the NOK the previous day (March 21). I documented this on Twitter.
https://twitter.com/radiantphysics/status/648202471425376256
@Victor Iannello
Thank you for the links and provenance of that slide, Victor.
Tim asked: Why has it changed to loss of control after the first eng fail?
The Boeing simulations with the alternate electrical configuration are more consistent with the descent rates as derived from the final BFO values.
@Victor, thank you for the detail.
@Mick, further citation via an India Today article which provided the following caption to the “question of time” image, “A screen shows the questions from family members of passengers onboard Malaysia Airlines Flight MH370 after a routine briefing given by Malaysia’s government and military representatives at Lido Hotel in Beijing March 22, 2014.”
Other reports for 22nd March described that the family group became aggresive towards the briefing team & the delegation left with the protection of the security guards. The Malaysian team, normally sat at the ‘top table’, under the screen, is absent in both the Mail and India Today images. The Mail’s image shows the audience on their feet, turned to the press at the rear of the room (it’s a large room).
@Mick Gilbert, @Don Thompson: It has never been explained why the turn near IGARI was described by the “Malaysian Air Force officer” as clockwise. The article also says eyewitness reports peg the flying at low altitude near Kota Bharu. In fact, much of what was reportedly said by the Malaysian Air Force offer is not consistent with our current understanding of the facts.
Try as we may, we will never sort this out with the raw military radar data.
@ST: Yes, I am aware of the “unsecure” label. More browsers are showing this for sites that are “http” instead of “https”. I have not yet migrated the site to an https server. Perhaps at some point I will.
@ventus45 said: May I now respectfully suggest, that we discard the Malacca Strait straitjacket / blinkers, and instead, take the deterministic performance limited 18:25 arc segment, from it’s most northern possible limit, to it’s most southern possible limit, and then apply the softer issues of motive, intent, human factors, and plausible “why’s” ?
Do you also suggest we ignore the BFOs at 18:28?
@DennisW said: At this time, I think we all need to take a step back and contemplate the “why” of this event.
@TimR has been me writing to me since 2014 with a rumor out of Malaysia that we have not been able to substantiate. In summary, he believes that:
– In the days after the disappearance, a politically-connected individual relayed to him what was planned for MH370.
– The objective was to embarrass the Prime Minister into agreeing to call new elections.
– There was to only be communication from the ground to the aircraft, and no communication from the Captain to the ground.
– Once landed and passengers released communication, was to resume.
– As it turned out, there was apparently no communication during the flight due to the break down of the negotiations in Malaysia.
– The ground team was unaware of the final outcome of the flight.
– The plan was for the Captain to fly MH370 over the Malacca Strait, out of Malaysian radar contact into the Andaman Sea, then listen out for confirmation the demands had been agreed to.
– The plane would then proceed to Banda Aceh and land to allow the passengers to be repatriated.
– In the event confirmation was not received early on the intention was to continue at a steady loiter around off the coast of Indonesia to suitable island landing sites while the negotiations continued.
– Possible landing sites included Cocos Islands and Christmas Island.
– There was no suggestion of suicide in the plan which was put in place weeks before the disappearance.
As @TimR continues to contribute here, perhaps we can engage him in more discussions on this scenario.
@Victor
However we slice it- low altitude to hide from radar, or high altitude to purposely paint a contiguous domestic flight path to avoid human eyes detection, the implication is an expert flown path that apparently managed a great escape over Malaysia without ever alerting the military.
The question becomes what were Malaysia’s criteria in their radar computers to signal an intruder alert? Whatever it was, it appears the pilot knew how to avoid those “intruder” flight characteristics.
The FIR gives us clues about some of Malaysia’s criteria, and they are weak:
(1) A flight is allowed to turn off transponder and still be considered “non-hostile”; (2) A flight is allowed to deviate from planned flight path, as long as there is a contiguous flight path inside Malaysia it is considered “safe”. One wonders if there are some areas like Kuala Lumpur Twin Towers where there could be more rigorous radar rules to detect intruders.
@Victor
Re: @TimR hypothesis
As you know, further above I posted on-line “chatter” from “Razmatazz” confirming a somewhat similar story to the @TimR story.
I accept the @TimR hypothesis as one possibility. Razak handled it so well, ignoring ZS-connection in the follow-up days backfired on ZS the attempted scheme. The world was much happier inventing a million-and-one alterbate theories.
@TBill: It’s hard to put any weight on the “chatter” from Razmatazz.
@Victor,
What I’m trying to say is; it is possible to have an alternate electrical configuration (but with L BU GEN still powering R Transfer) that allows flight onto second eng failure. This still should produce the high rate of descent observed.
@Tim: If the left backup generator is on, there is no loss of A/P and no autostart of the APU until the left engine flames out. The Boeing simulations say that does not produce high descent rates anywhere close to two minutes after the autostart of the APU.
@Don Thompson
I am sure you can correct my understanding of the ACARS Logs.
The ACARS message requesting urgent contact sent at 18:03:23 UTC to the 9M-MRO Cockpit Printer resulted in a response that it had failed because the “up intercept aircraft not logged on”.
In the Inmarsat Log we see a transmission of user data starting at 18:03:41.405 UTC and ending at 18:05:59.905 UTC with five requests for acknowledge of the user data. There was no record of any response.
We are told that the ACARS message was then retransmitted at 18:08:09 UTC and every 2 minutes until 18:43:33 UTC. Each of these 20 attempts resulted in a response that the message had failed.
In the ACARS Log in the Malaysian Factual Information and the Safety Report only the first 6 messages are recorded.
Why were the other 14 messages not included?
In the Inmarsat Log we see no further transmission of user data that aligns with the second message attempt reported in the ACARS Log at 18:08:09 UTC or any other attempt up until 18:43:33 UTC.
Was the ACARS messaging switched to VHF after 18:03:23 UTC and prior to 18:08:09 UTC?
The ACARS Log in the Malaysian Factual Information was filtered on:
1. Aircraft Registration = MRO.
2. Source/destination = QXSXMXS.
3. Timeframe = 7th March 2014 12:48:00 UTC to 7th March 2014 20:00:00 UTC
Based on this filter, 37 ACARS messages are shown.
We do not see any messages prior to 15:54:41 UTC, although there was a successful Logon to POR at 12:50:19.735 UTC, whilst 9M-MRO was at MAS Engineering. We do not see any message after 18:15:25 UTC, although the urgent contact message was retransmitted until 18:43:33 UTC.
The ACARS Log in the Malaysian Factual Information has been truncated and is incomplete.
The ACARS Log in the Malaysian Safety Report Page 1 was filtered on:
1. Aircraft Registration = 9M-MRO.
2. Timeframe = 7th March 2014 15:54:00 UTC to 7th March 2014 18:15:59 UTC
Based on this filter, 5 ACARS messages are shown.
We see 3 ACARS messages via VHF between 15:54:31 UTC and 15:54:36 UTC via BKKXCXA. We see 2 ACARS messages via SATCOM between 15:54:41 UTC and 15:54:53 UTC via QXSXMXS. We see no more messages after 15:54:53 UTC until 18:15:59 UTC despite the filter setting.
The ACARS Log in the Malaysian Safety Report Page 1 has been truncated and is incomplete.
The ACARS Log in the Malaysian Safety Report Page 2 onwards was filtered on:
1. Aircraft Registration = MRO.
2. Source/destination = QXSXMXS.
3. Timeframe = 7th March 2014 15:54:00 UTC to 7th March 2014 18:15:59 UTC
Based on this filter, 37 ACARS messages are shown.
We see 2 ACARS messages via SATCOM between 15:54:41 UTC and 15:54:53 UTC via QXSXMXS repeated from Page 1. We do not see any message after 18:15:25 UTC, although the urgent contact message was retransmitted until 18:43:33 UTC.
The ACARS Log in the Malaysian Safety Report Page 2 onwards has been truncated and is incomplete.
What are the Malaysians trying to hide by presenting an incomplete record of the ACARS Log in both official reports (Factual Information and Safety Report)?
Was the end time of the timeframe filter changed from 20:00:00 UTC in the Malaysian Factual Information down to 18:15:59 UTC in the Malaysian Safety Report as an attempted cover up?
@Victor
Upshot of that, if true, is that MH370 is not over yet. Probably everyone in power on both sides would know about this and there is a day of reckoning ahead.
Although this is entirely speculative, we have some interesting developments to follow and to better understand:
1) Claims from @TimR that the aircraft was to be contacted from the ground with the status of ongoing negotiations.
2) Truncated ACARS logs that might be missing ACARS messages submitted between 18:15 and 18:43 by MAS operations for the VHF link.
@Victor
I am happy to accept @TimR hypothesis, but there are some weaknesses, like how could it possibly be kept secret all this time? And some other weak points.
We should be inclusive of possibility that Z was on a secret protest mission. In either case Z made a great escape without detection, snookering the air defenses, and Malaysia hid the facts of what they thought happened and is still hiding the facts and their thoughts.
@Richard & Victor,
I will respond presently, too much today already.
@TBill: If the intention of the planned diversion was to draw attention to a corrupt administration without hurting anybody, and then that plan caused the disappearance of 239 people, neither side would want to put the details in the public domain. The (then) opposition responsible for the diversion might want to quietly make it known what was planned so that perhaps that knowledge might be used by the search effort to find the aircraft. Perhaps the leak to @TimR was part of that effort, but the leak was ignored because it was considered to be an unsubstantiated rumor, of which there have been many that have proven to be false.
@DennisW,
In term of finding a direct solution from the satellite data, I would agree that we have already extracted as much useful information as we are going to get. That process has produced several solutions, each of which fails in some way or the other to provide a complete match to our our current understanding of the data (satellite data and Boeing end-of-flight simulations).
Recently, there have been several attempts to add maneuvers after 19:41 in order to generate impact points north of the currently searched zone (while also matching the satellite data). They are also assumed to match the Boeing descent predictions. However, that is a big unknown at present. One of these late-maneuver routes could be correct, but the only way to find out is to search there. The cause of the failure of the current search is currently unknown – did the aircraft impact farther from the arc or is it farther north?
The “analytics”, as you call it, still serve a very useful purpose in eliminating newly proposed routes that are inconsistent with the satellite data.
Interestingly, the drift studies have not allowed the latitude range on the 7th arc to be restricted compared with the predictions based only on the satellite data. In my opinion, we would be in basically the same situation today if there were no drift studies at all.
@DrB
Yes.
I did not intend to sound or imply a lack of appreciation for your work. It is consistently very good. I am simply burned out analytically.
@Victor @TBill
There were reports that Z had political concerns but did not show signs of depression or other indicators of suicide. This would align with a theory of well executed diversion to beyond MEKAR and holding to await some resolution, where the fallback was to dispose of the AC. It could equally have been a Z plan to just lose the AC.
In the recently discussed hijack and negotiate plan, this was said to require some circling about in vicinity of Christmas Island (for example) and this would directly impact arc to arc crossing times. This is not dissimilar to Z performing a number of wide 360 turns for the purpose of decoupling max range and max endurance calculations for a pure disposal plan.
In either scenario the effect of holding or clearing turns is a huge change from the ATSB/DTSG particle filter models. This clearly moves arc 7 crossing to the north – maybe above 25S.
It would have been interesting if DTSG performed some PF simulations that inserted some random 360 turns at a reasonable holding latitude. Clearly the longer Z would hold waiting a decision, this would significantly impact all arc crossings. But they only assume Ghost flights.
Thanks Victor for the clarification.
So we are now assuming the autopilot/Primary flight controls drop out on the first engine(right) failure. But if this is the case, then surely, if the left engine is still running then there will be an overwhelming roll to the right, not to the left.
@DrB
I should have mentioned that my threshold for “consistency with the satellite data” is less stringent than yours, but we (you, Mike, me,…) have a long history of discussions on that topic. Victor’s Cocos path is an LNAV path with a runway track adjustment at the end. In my view it is acceptably consistent with the satellite data.
@Hank said: But they only assume Ghost flights.
Actually, in the DSTG study, they allowed maneuvers, with prior distributions for the number of maneuvers and the time spacing between maneuvers. In my discussions with the DSTG, I suggested that they expand the types of maneuvers (which were restricted to changes in altitude, direction, and speed) to include a structured “hold” maneuver which would essentially hold the plane in place for a given amount of time (again determined by a prior distribution). They seemed receptive, but I don’t think that was ever carried out. One of the problems is the posterior distribution becomes wide with no discernable peak. That means a broad range of solutions (including ones terminating further north) become acceptable. So this kind of analysis may provide justification for a search further north, but little else.
@DrB
All the searches were based on the particle filter simulations conducted by DTSG as documented in their book. While this uses the BTO and BFO data, the model is highly dependent on the random flight profiles generated to feed the filter. The designed the simulation to favor creation of long straight airline flights and validated it against other successful flights. The results are only as good as the assumptions. For the specific assumptions – which are for a ghost flight – the likely flameout at arc 7 is at 38S latititude with decreasing probability to north and south along the arc.
If DTSG had required every flight to include one or more large 360 degree turns, this would have moved arc 7 crossing to the north. But the scenario of a live pilot performing intentional holding turns was never considered.
While a live pilot that knew about BTO/BFO and the concept of the arcs could enter a best range glide after flameout to increase the uncertainty, there is no indication Z knew about the arc capability. So after flameout he would plan the final decent to minimize debris on the surface. I don’t know how that is best done in rough sea state – but a rapid descent to just above the water, hold off for a bit, and then do a 360 back loop to enter vertically at stall speed might be best????
The Boeing simulations are unpiloted spiral dives – classic out of control demise. But a piloted ending could use the same spiral to get down quick and the pilot art would be setting up the airspeed and attitude for water entry.
The SEA Q400 ramp worker did some very impressive maneuvers. Z could have practiced many terminal maneuvers on his simulator.
Victor suggested a ditching and sinking. Ditching without power in the ocean usually is very difficult. If he could enter vertical at 100 knots, maybe the aircraft would stay together until the tail cleared the surface. Usually a pilot is trying to land on the water and not dive deep under it.
@Victor
I reviewed their book in detail and particularly the modelling for the flight profiles. They were using a very simple model for turn magnitude and turn frequency which was heavily biased by the filter toward small turns. This model is OK for simulating direct airline flights and that is how they validated it. If they wanted to include a single 30 minute 360 turn they would have had to allow for it (and could have) and set probable location. Doing this would have created a very different probability map. The Mach and altitude modeling did not concern me because Z would most likely have left these alone.
They also assumed infinite fuel with the idea of post simulation correction of the results. Bad. They do not seem to recognize that the duration and the route are decoupled. They ignored the separate problem of the time for arc 7 fuel exhaustion is a different problem than the location for crossing arc 7. Maybe the PF could not handle two different objectives so they ignored fuel burn. The book discussion of maneuvers and dynamics showed a very amateur discussion of aircraft flight dynamics and little understanding of the difference between range and duration models.
While the stochastic modeling by DSTG in the PF may have been world class, the aircraft dynamics were simplistic. The result is for the best straight airline flights to the SIO.
If they inserted piloted intentional holding patterns at random times and duration the heat maps would be different. They could not mix these because these would represent different scenarios.
So everything that has been done has been for long straight airline flights to the SIO.
For the scenario that you raised o holding while negotiations were conducted was never considered. If DSTG used their existing model and just required one 30 minute duration 360 turn at the latitude of Christmas Island, a very different heat map would result.
They could run for different numbers and duration of 360.
The heat map would be different for each assumption of location, size, and number of 360 turns.
I’d love to see a PF simulation with just one planned 360.
@Victor
Your suggestion to DSTG was good. You wrote – “One of the problems is the posterior distribution becomes wide with no discernable peak. That means a broad range of solutions (including ones terminating further north) become acceptable.”
By ignoring a live pilot – maybe under pressure from Malaysia – DSTG only analyzed routine airline flights. You are exactly right that each holding loop moves the arc 7 crossing north.
They could have looked at other scenarios for what a live pilot might do and they had no reason to not consider that a pilot was active until water entry.
And they spend 200 million on a search. The sole justification was the the wonderful particle filter made a nice heat map. You would get a nice heat map somewhere else along arc 7 if you planned one 30 minute hold.
It seems stupid to spend the search time on a simplistic assumption and to not have run more assumed scenarios. Someone really didn’t want this airplane found. I don’t know if this is nefarious or just inept.
@Hank said: Victor suggested a ditching and sinking.
No, that wasn’t me. I believe the debris shows evidence of a high speed impact.
By ignoring a live pilot – maybe under pressure from Malaysia – DSTG only analyzed routine airline flights.
Yes, this was a fundamental error. The flight of MH370 might have looked nothing like the airline flights they used to calibrate their model.
Another assumption was that the aircraft was level at 18:40, which constrained the track to be southerly at this time.
@Victor
May you could get DSTG to modify the model to include a planned holding pattern as the AC crosses the latitude of the Cocos islands. Use a 30 minute 360 turn to the left. Then have them run another simulation for two loops at the same location. Each of these Monte Carlo runs would generate its own heat map. This would be a rough analysis of a hijack and negotiate scenario – a 30 minute orbit and a 60 minute orbit. Each planned orbit would move the heat map more north. It would be interesting to see how these maps with some holding would correlate with the drift models.
@Andrew. “In the past it was thought that the mode reversion was caused by the loss of pitot heat. However, recent advice indicates the reversion is caused by the loss of other sensors required by the PFCS.”
That’s new and sounds more sensible. Are the transfer buses the source of power for the “other sensors” or are the main buses? Any more on which they are?
@ Victor – Thank you for the inputs. I tested some other sites as well and this seems to be like you mention how browsers are reporting different sites these days.
@Hank: I already proposed to the DSTG that they include a holding pattern as a maneuver. I don’t control their activities, nor set their priorities.
On the other hand, if you feel that a particular case is worthy of further examination, you should examine it.
@Hank
Like Metron, DSTG is coin operated. You don’t get unfunded freebies. Likewise the absence of university or corporate people providing inputs is largely throttled by their legal departments. Only folks who are truly independent investigators can contribute.
@DenisW. @Victor
Far too complex to create the particle filter from scratch for me. It uses the BTO and BFO data to assign conditional probabilities to each of the randomly generated flight profiles. It would be relatively easy for DTSG to add the ability to the path generator to perform a hold. The hold has a much larger impact on the arc 7 then just the time delay between two adjacent arcs. It could swing the probable downstream routes.
But I expect that there is no incentive for DSTG to do free work to show the entire basis for their search area maps used an overly simplistic maneuver model on which over $200 million was spent.
I’ll have to think about whether I can pick a location on arc 7 (such as 23S) and use a constant Mach and altitude and only adjust a time delay for one or more holding circles to match all of the crossing times. It may be that a single time delay between two arcs could allow many crossings of arc 7 above 25S?
I’ll think about a very crude proof of concept simulation.
@ Dennis
This flight only has a mathematical resolution if the crew and passengers were unconscious (Newtonian). If the plane was being actively piloted we need to begin with motive. A suicide motive leads us to a race to the southern Indian Ocean as one possibility. What levels of confidence do we have in each data set? We know with a high level of certainty that the plane made it to the transponder shut down. I assume the military and commercial radar back each other up to the fmt. The Inmarsat data has no other back up. It came back on with no flight identifier and it is the easiest data to manipulate. If we ignore the satellite and looked more at range and motive what does that leave?
We have to look to the actors on board. First is the crew. The pilot and co-pilot didn’t appear to know one another so one of them would have to be acting alone. This complicates matters for the other pilot. The 3 Ukrainians are serious guys and capable of handling oxygen deprivation (experienced diver). They would not go down without a fight. I have taken flights from China to Malaysia several times and there is no quiet time with lights out. The stewardesses would have been in the aisles hawking duty free goods. Everyone would have been trying to get on their internet. Many of the passengers would not have turned off their calling/texting. Chinese do not follow rules well and the passengers would have been up opening overhead bins. I have seen passengers get up from their seat on final approach and open bins. The sudden turn would have included a warning from the flight deck saying they have encountered mechanical electrical issues and will be trying to resolve. This message would come from the flight deck from whoever was flying the plane. The only way to hijack the flight for 6 hours would be to kill all the passengers. Otherwise you would definitely have an angry mob outside the cockpit. Inflight real time falsifying Inmarsat data seems reasonable. The Ukrainians would have to overcome the flight crew. I think they’re much more capable of that than the flight crew would be.
Which brings us to motive
For the flight crew we have suicide although I can’t believe that the other crew members couldn’t come up with something to regain control.
Flight crew also has political motives.
Ukrainians would be trying to neutralize an asset. What a great way for them to call attention to US military than to attack Diego Garcia knowing that the US would not shoot down a commercial jet liner. And doing it before the start of a war game would leave a strong message.
In that scenario the US shoots down the plane and then falsifies the Inmarsat data. The US wants this to go away as much as the Malaysians do. The US radar on Diego Garcia would have been tracking an incoming plane. When there was no response they had no choice but to shoot it down.
I think the Ukrainians would definitely be able to kill all on board. As a matter of act the US may have shot it down after it became apparent that all of the passengers had been killed.
I think Diego Garcia is a reasonable target. Excuse the wildly unfounded speculation but what we’ve tried so far hasn’t really worked.
@David
RE: ‘Are the transfer buses the source of power for the “other sensors” or are the main buses? Any more on which they are?’
I don’t have a comprehensive list, but my understanding is that they are mostly position sensors associated with actuators controlled by the L2 ACE. The L2 ACE is only powered by the L 28V DC bus, which is normally supplied by the L AC transfer bus (via the L TRU).
@Andrew. Thank you.
@Nederland, Victor. Here is the new appraisal of the French DGA’s right flaperon separation hypothesis, of the evidence in its report and of alternative theories. It replaces that of 5th August.
It is more comprehensive and incorporates various changes plus rethinks. It attributes the most likely cause of the separation to overstress causing a wing break in an unmanned high speed spiral.
While hardly seminal it does bolster the case that there was no ditching. Also, it may prove a useful reference for others later, so I will correct and amend it as necessary.
The Boeing simulations might include descents within and outside the flight envelope and simulator data base which overstress the aircraft though I have supposed they will have looked at that possibility.
https://www.dropbox.com/s/kqgi5tjcnjzc25e/Flaperon%20Separation%2C%20MH370.docx?dl=0
@Trip said: Excuse the wildly unfounded speculation but what we’ve tried so far hasn’t really worked.
Now is not the time to throw out all the data and latch on to theories that are not supported by any facts. Rather, now is the time to build upon what we know based on the available evidence, including the most recently failed search, and possibly supplement that with plausible theories that include human factors.
Fascinated observer here (hence the name), I normally avoid commenting.
However, TimR’s postings about a rumoured plan to land on an island reminded me of this from 22 June 2014:
https://www.telegraph.co.uk/news/worldnews/asia/malaysia/10917868/MH370-captain-plotted-route-to-southern-Indian-Ocean-on-home-simulator.html
“Sources close to the investigation confirmed to The Telegraph on Sunday that a deleted flight path had been recovered from Capt Zaharie’s simulator which had been used to practice landing an aircraft on a small runway on an unnamed island in the far southern Indian Ocean.”
Was it ever substantiated that the simulation involved landing on an island in the SIO? Was this simulated island ever identified? Does this tally with what TimR had heard, or would they have to be different islands?
Regarding the rumour reported by TimR, we know from Zaharie’s public life he supported an opposition party, and we known he was very angry about his government’s behaviour. We know the flight was just after the (supposedly politically-motivated) conviction of the opposition party’s leader. Zaharie also had an apparently stable character, in contrast to the Germanwings pilot who crashed his own plane in 2015.
Putting these together, it really is not that much of a stretch to think Zaharie might want to use his flying skills in some kind of peaceful protest that involved returning the passengers safely afterwards.
But if the plan was peaceful, how could it go wrong in a way that let the plane fly until fuel exhaustion, and which kept the final location of the plane a mystery? This seems to be the biggest barrier to a “peaceful protest” theory?
Pings 1,2,3,4,5 don’t align with 6 & 7. Ping 1 @ 18.25 is too far north west. Once shifted south east, or in the opposite direction, the timeline aligns and shifts further south as it should in a 7h00 endurance after a right ‘or’ left turn-back at IGARI (Malacca Strait diversion excluded) The Timeline is CORRECT, but the geo-graphical depiction or illustration of Inmarsat’s 7 ping rings on the google map is WRONG.
@Air Life: You’ve presented too many falsehoods as facts. This is not Facebook or Twitter, where that behavior is common. You’re banned from here.
@Layperson: The deleted simulator data sets (the “flight files”) that we have do not show a landing on an island in the SIO. On the other hand, we are missing significant portions of the data files, including the portions that would show the route programmed in the flight computers, so we don’t know for sure.
I am in agreement that if the diversion was meant to cause no harm, something went wrong.
@Victor
@Layperson
I sense the recent SIR report has caused some misundertsanding of the Microflight flight simulator program. As FS9 or FSX users, we are constantly hitting “SAVE” but all we are saving is the current location and conditions. We are also saving the future flight plan, but as Victor has discussed, each and every truncated file from FBI was missing a ton of case information including the forward flight plan is missing (corrupted file fragments is all we have).
The SIR report suggests a weakness that no complete flight history was saved for the sim flight. But nobody saves complete flight history (not sure it is possible). Normally we are saving a sub-case at one spot along the flight path. So of course all we have is saved “waypoints” case fragments…but that’s all we usually ever have.
In his prior paper, Victor reconstructed/estimated the sim flight path to the best of his ability. And Victor’s ability is pretty good…no one really has improved on Victor’s analysis (as hard as I try).
As far as island landings, I believe that is in reference to unrelated flight cases to Diego Garcia.
@Victor: Do you know if these simulations are the ones indicated in the 2015/12/10 update of the ATSB report of 2014/12/03?
ATSB update 3th december 2015 wrote (copied from file “ae2014054_mh370-definition_of_underwater_search_areas_3dec2015_update.pdf”)
Simulator data
The aircraft behaviour after the engine flame-out(s) was tested in the Boeing engineering simulator. In each test case, the aircraft began turning to the left and remained in a banked turn. In many cases, but not all, a phugoid oscillation in pitch was observed. The final position of the aircraft was within a region defined by 10 NM forward and 10 NM left of the position where the flame-out occurred.
…/…
Basic turn analysis
The ATSB performed a basic trajectory analysis of an uncontrolled, but stable aircraft. The trajectories commenced from estimated flame-out locations and used starting conditions consistent with the DST Group analysis results and ATSB assumptions outlined in previous sections. The analysis included constant and increasing bank angles, but did not include variations in speed or pitch angle. There were many resulting trajectories consistent with an intersection with the 7th arc at the appropriate time. Some of these solutions ended between 20 and 40 nautical miles from the 7th arc.
The majority of the solutions from the turn analysis involved a left turn. A left turn is consistent with the simulation results. A larger proportion of the trajectories ended on the north-western side of the 7th arc.
…/…
Glide area
A simulation was performed to determine the glide distance of the aircraft under active control to maintain wings-level attitude. The simulation (from FL330) resulted in the aircraft gliding for a total distance of approximately 125 NM from the point of the second engine flame-out.
In order to make this distance the aircraft would travel approximately 15 NM in the first 2 minutes of the descent (approximate time required to start the APU and initiate the log-on sequence). Therefore, from the 7th arc, the aircraft has the potential to glide around 110 NM.
…/…
Bold by me.
Before discardings the few data we have, it seems that some areas remain to search…
Regards,
PL
@Pax Lambda: You are citing results of the “Basic Turn Analysis”, which were not Level D simulations, and included some unrealistic assumptions such as constant bank angle (for some simulations) and constant speed and pitch angle. I don’t see how we can apply these results to realistic flight dynamics.
@David: “Here is the new appraisal of the French DGA’s right flaperon separation hypothesis..”
Very interesting, and if validated, certainly a big deal for EOF and search radius — is this your own [anonymous] work, or did I miss the signature?
@lkr
I too wondered.
If you look at document properties, author is David.
We must insist David take credit for his good work.
Victor said: You are citing results of the “Basic Turn Analysis”, which were not Level D simulations…/…
OK, agreed.
But the main question was (which was perhaps not very clear): Are the End-of-flight simulations of this thread the ones mentioned in the December 2015 ATSB report (“Simulator data” section)?
Regards,
PL
@Ikr, TBill. Thank you, yes my work.
As you say Ikr, ‘if validated’ it could help, though that is an ask.
@Hank, I can not think of any sensible reason to loiter with 360 degree turns late in the flight, (at the latitude of Christmas or Cocos Islands), much earlier, near Ache perhaps, but only if the intent was to land in Ache. If negotiations were still ongoing “that late in the game”, they had effectively failed.
@Victor, re the 18:25 BFO’s on my route, there were significant thunderstorm cells west of Sibolga at the time, so some maneuvering due weather avoidance could be expected.
@David: Very interesting paper. What probability would you assign to separation of the right flaperon due to breakup of the wing in a high speed, tight spiral descent?
@David
Do you think the suggested wing break location is consistent with the YouTube video of the stress test break?
@Victor. It is hard to express high confidence in that alone since it depends on assumptions, estimates and judgements; and there could well be questions about it which I cannot answer. It is just a best (to my knowledge) fit amongst the alternatives.
Also had any of the simulations raised this as a possibility, including those beyond the flight envelope and data base, I would have expected Boeing to have been alert to that. Because nothing has been nothing heard that puts something of a dampener on it.
To help assess confidence in it I would forward it to the ATSB, for Boeing would not respond, but then I do not think the ATSB could these days either and there could be no confidence in Malaysia looking into it. Suggestions are welcome though.
To me there can be more confidence in the negative, that it wasn’t a ditching. That does not fit at all though I gave it a fair go. My feel is 4 chances in 5 it wasn’t that.
Given a few weeks if there is no rebuttal about that or the wing break, confidence in both should rise.
One other comment on probabilities is that were the (lesser likelihood) flutter included as an either/or, that would increase somewhat the prospects of the crash close to the 7th arc.
@PaxLambda
RE: “Are the End-of-flight simulations of this thread the ones mentioned in the December 2015 ATSB report (“Simulator data” section)?”
Apologies for butting in on your question – the EOF simulations discussed above are additional scenarios that were conducted by Boeing in April 2016, at the behest of the ATSB. They were reported in the ATSB report MH370 – Search and debris examination update, dated 2 November 2016 (amended 18 August 2017), pp.7-8.
https://www.atsb.gov.au/media/5773389/ae-2014-054_mh370-search-and-debris-update_aug2017.pdf
@TBill. In case you have the short video this one is the clearer, starting at 3 mins and also gives a description (about 5:40) as to break characteristics, including these being at about the same place on each wing.
https://www.youtube.com/watch?v=x4vEfZFx4hk
To me, judging by the tank access holes at 5:04, the break looks to be at about the place I have drawn it though it was diagonal whereas mine, indicative, was fore and aft.
There are refinements such as that diagonal, squarer with the leading edge, reducing lift and increasing drag as the outer wing rises after fracture. Also outside the envelope the chordwise and spanwise lift distribution of the test could change as the stresses could with thrust vs no thrust.
The 5 part You Tube series relates a magnificent story, though missing were the recollections of the rubber ducky last seen on the wing tip at 3:03, just before the wing broke, apparently unbriefed.
@Victor. In the second last line of my two above please read, “…flutter included as an either/or….” as, “….flutter included as a secondary possibility….”.
@David
Many thanks for your excellent analysis of the likely flaperon separation as a result of a wing break in a high speed spiralling descent and the exclusion of a ditching as the cause.
@Richard Godfrey. Thank you.
Timing. I was about to send you this:
@Richard Godfrey. Recovered item 5 was the R1 door closet panel part recovered from Rodrigues Island on 30th March 2016. Its ATSB photograph was after barnacle removal and in the Malaysian debris summary I can see none.
I regret I cannot find your latest drift analysis but my memory has it that this was not amongst those you included as recovered shortly after beaching.
Looking for other things today I came across the below which shows barnacles best at 0:44 and 1:20 – 1:23. In case that should be of interest in any continuing drift studies:
https://www.youtube.com/watch?v=FJEvqExMHbE
@Andrew: Thank you for your clarification about simulations.
I hope I will not bother you too much, but I take the opportunity to ask some questions: Could you confirm the 6 statements below are true and, most important, what happens in 7?
1 – If both main AC buses are unpowered in flight (both IDGs and both backup switches OFF), the APU starts automatically, whatever position of its rotary knob.
2 – As the APU doesn’t power the buses quickly enough, the RAT comes out and goes on line before APU could power the buses.
3 – If the plane was flying by the autopilot, it goes OFF and the plane has to be flown manually.
4 – If the APU knob was on OFF position, turning it to ON then back to OFF (before the APU comes on line) shuts the APU out and prevents it to power the plane.
5 – From then, the plane flies with only the RAT, and can do that for a long time, say one hour or more, if somebody manually flies it.
6 – If, later, an IDG or both come back on line, the plane is again with “normal” electrical power (excepted the RAT remains out, and some loads are shed if there is only one IDG). The autopilot can be used again.
7 – Now, if both buses are unpowered again later in the same flight, does the APU automatically start again, or does it remain OFF because some system “remembers” it has been shut (line #4)?
Thanks by advance,
PL
@David,
You’ve set out a good analysis that a wing break, that could be expected during the end-of-flight descent, is much more consistent with the condition of the right flaperon, and the adjacent flap. A thorough structural analysis for the consequence of bending the flap along its span would be useful, as would an analysis for the flaperon hinge-actuator fittings.
@David: That you are 80% sure the flaperon damage is not consistent with a ditching is helpful. If the aircraft impacted south of 25S and outside of the search zone, then our best estimation is a steep descent followed by a recovery followed by a long glide followed by a steep descent. That sequence doesn’t appear to make much sense.
@David
“…not think the ATSB could these days either and there could be no confidence in Malaysia looking into it. Suggestions are welcome though.”
…also maybe consider France as we know they have active MH370 effort.
@Don. Pages 18 & 19 of the report note that corrosion prevented scanning electron microscope examination of the fracture faces of the flaperon actuator attachment lugs and that there was distortion of them in the outboard direction. The detail of that is in Annex 11 pages 219-226, though in French. That does seem to be all that would be available.
Yes, the flap in ATSB hands received little attention, at least that the ATSB has disclosed, other than that relevant to its finding about flaperon and flap positions at separation. It is now in Malaysian hands.
The chordwise break at the support assembly, the separation of the auxiliary support track carriage internally and the shape of the trailing edge missing and its skin stresses are of interest as are the distortion, twisting and bending you mention.
A reconsruction of what happened to these two items surely could have been expected. With the flap and with DGA report to hand and access to aircraft plus Boeing should have all led to much more but there is no sign of that.
@Victor
We now have 3 independent evidences that there was a steep descent at the end of flight:
1. Two BFO satellite data at 00:19:29 UTC and 00:19:37 UTC.
2. Four Boeing Simulations in the alternate electrical configuration.
3. The Debris Analysis of the Flaperon and Outboard Flap by David.
Therefore the likelihood that the MH370 end point is outside the 22 NM search width, is remote.
The main wreckage of AF447 was found on the seafloor at a depth of 3,900m over an area 600m x 200m. If the main wreckage of MH370 is similarly spread, then the probability that it was missed by Fugro and Ocean Infinity is also remote.
The case for continuing the search north of 25°S and ±22 NM is very strong.
@PaxLambda
1 – If both main AC buses are unpowered in flight (both IDGs and both backup switches OFF), the APU starts automatically, whatever position of its rotary knob.
Yes. The trigger for the autostart is the loss of power to both transfer buses, which are normally powered by the main AC buses. Note: The backup generators only power the transfer buses.
2 – As the APU doesn’t power the buses quickly enough, the RAT comes out and goes on line before APU could power the buses.
Yes, however, the RAT only powers the standby electrical system to ensure power to essential systems. The entire electrical system would be powered by the APU when it becomes available, subject to switching.
3 – If the plane was flying by the autopilot, it goes OFF and the plane has to be flown manually.
Yes.
4 – If the APU knob was on OFF position, turning it to ON then back to OFF (before the APU comes on line) shuts the APU out and prevents it to power the plane.
Yes.
5 – From then, the plane flies with only the RAT, and can do that for a long time, say one hour or more, if somebody manually flies it.
Yes, assuming at least one of the engines remains operating! Only the essential systems would be powered.
6 – If, later, an IDG or both come back on line, the plane is again with “normal” electrical power (excepted the RAT remains out, and some loads are shed if there is only one IDG). The autopilot can be used again.
Yes, provided the crew cycle the primary flight computers DISC switch to restore the PFCS normal mode.
7 – Now, if both buses are unpowered again later in the same flight, does the APU automatically start again, or does it remain OFF because some system “remembers” it has been shut (line #4)?
Good question! As I said above, the autostart trigger is the loss of power to both transfer buses. If that should occur again later in the flight, then ELMS should send the same start signal to the APU controller, regardless of the position of the APU start switch. As far as I’m aware, the system has no ‘memory’ of the previous autostart attempt.
@Richard Godfrey: Let’s all try to develop a list of evidence that suggest the impact might not have occurred further north than 25S. I’ll start.
1) Some drift model results predict the debris arrives to East Africa too early. (Some drift model results are consistent with a more northern impact.)
2) For the automated path over Cocos Island ending at 22S, the BFO at 00:11 is in error by about 12 Hz, which is 5 Hz greater than the Inmarsat has stated in the past is acceptable. (The DSTG suggests larger errors than 7 Hz are possible. The temperature of the SDU at 00:11 might have been colder than normal due to depressurization of the cabin. There might have been an ongoing descent with higher BFO errors.)
Others?
@Victor
I am not at all “in love” with forward drift models. They are just fundamentally bad science (no disrespect intended, Richard). At the end of the day they anwser the question “what is possible” instead of “what is probable”.
In the case of MH370 the forward models are “sort of” OK because we think we know the starting point had to be on the 7th arc, but using that constraint contaminates the purity of the analytics.
@DennisW: Even if we recognize the uncertainty in using the drift models, we still have a large discrepancy between drift model results that needs to be resolved.
Is there other evidence that contradicts a possible impact along the 7th arc north of 25S?
@Victor
3. Despite their best efforts Fugro or Ocean Infinity missed the plane in the underwater search.
@Victor
What I am suggesting is that there is less discrepancy in reverse models. If you drop enough debris at the South pole eventually a piece will land on ReUnion at about the right time.
@Victor.”That sequence doesn’t appear to make much sense.” To me either.
As to a ditching to me there is little possibility of one along the DGA hypothesis lines.
There might still be a possibility of a flaps-up ditching being consistent with the damage evident under the flaperon and flap lateral-displacement approach if the right wing digs in.
I will look into this again tomorrow and my page 16 conclusion which I think applied to a ditching flaps down.
@Dennis
I agree drift modelling is not a precise science.
However, if I come up with 10 possibilities, but I do not know how probable each possibility is, then I can still go check each possibility out.
Henrik Rydberg in his drift analysis did work out the probability of floating debris in the SIO beaching on a coast somewhere and reflected the result in his forward drift analysis. He concluded the most probable origin was 34.5S.
@Victor Iannello
Re evidence that suggest the impact might not have occurred further north than 25S.
In the limit, going further north risks encroaching on Indonesian radar coverage if a great circle path is used.
A more subtle objection is that ‘further north’ implies ‘slower flight’ and probably ‘lower flight’.
Personally I advocate ‘further north’ with a lower and slower flight and a fuel dump to restore the expected end-of-flight time however I do recognise that there are constraints on ‘how low’ and ‘how slow’ based on both BTO and BFO data.
@Victor
Reasons for points below 25 South:
This is where I am stuck at:
(1) Absoluetly perfect match to BFO’s until 22:41 for original IG 187T deg path, good match after 22:41.
If I had a nickel for future coin-op DTSG studies, I would study all possible intentional flight paths after going 187 degT until 22:41 Arc5. I think it could head southeast to Dordrtech Hole with the winds, but also could divert to north of 25 South, so it does not exclude north of 25 South.
per my prior essay: On the Straightness of the BFO Trend
https://docs.google.com/document/d/1d5_ZM6lfcgjVelP6k6RjPIvsZeOgDYtJXELYHDewg3k/edit?usp=sharing
@ventus45
You noted: @Hank, I can not think of any sensible reason to loiter with 360 degree turns late in the flight, (at the latitude of Christmas or Cocos Islands), much earlier, near Ache perhaps, but only if the intent was to land in Ache. If negotiations were still ongoing “that late in the game”, they had effectively failed.
I was primarily discussing the impact of holding on the latitude for crossing arc 7. I only referenced the Cocos latitude because people had discussed these islands as a possible landing site.
Just a note. Every hold circle moves the flight path to the east between arc segments because of the SW-NE orientation of the adjacent arcs. A 15 minute hold takes over 100 miles off the available arc to arc range and this may require a more easterly course. If a pilot performed one 360 every hour between arc 1-2, 2-3, 3-4, 4-5, and 5-6 this would result in a significant shift. The number or length of holding between arcs would be limited by the arc spacing and reasonable Mach number for the cruise portion.
Sorry for backtracking a bit, but I don’t quite get the comments on those large radius 360 turns.
Would such turns not have changed the location of the corresponding arcs?
I mean, if the pings happened to be during one of those loops, it would show an anomaly, wouldn’t it?
In other words, the pings should not be somewhat equally spaced apart or there should be something else obviously inconsistent if those 360 turns did happen.
And another comment, if I measured on the image above (with the inherent lack of accuracy) the distance between the 7th arc and the calculated/simulated point of contact for all 10 scenarios, is seems to me that it should all have been within the searched area.
Please be gentle, I don’t mean to push any buttons, just stating some things I observed, right or wrong. 🙂
Argh, sorry, I meant “the arcs should not be somewhat equally spaced apart”, not the pings.
@TBill
(1) Absoluetly perfect match to BFO’s until 22:41 for original IG 187T deg path, good match after 22:41.
A perfect match to a random walk is what got the early analytics to converge on 38S. BFO can tell us the plane went South at the FMT, and that it came down rapidly at the end. Beyond that it is not useful.
@DennisW
If we ever find the black box (doubtful?) it will be very interesting to see how a loiter path or Jean-Luc’s path conspired to appear like a straight flight to Arc5.
@flatpack
You may not have followed by periodic 360 degree turn. You drop ground speed to zero if you perform a 360 degree turn between the arc segments. It takes exactly one hour to go between arc 2 and arc 3, but if you perform a 360 degree circle for 15 minutes, you now have only 45 minutes to cover the arc to arc distance. The aircraft burns one hour of fuel but only uses 45 minutes for transit. Because of the SE-NE orientation of neighboring arc segments the most probably arc intercept moves to the east with each hold. If a pilot just did a clearing turn every hour, this burns the fuel but moves the course to the south east versus more southerly.
Since there was plenty of time it is easy to just hold every hour (to avoid a slick on crashing and avoids a vapor trail of a dump).
I also mentioned the periodic holds decouple the range calculation from the known duration. The time between arc 1 crossing and arc 7 is 5 hours and 54 minutes.
My view is that an experienced pilot would have engaged in periodic holding to mess up any Boeing calculations of range/duration. One 360 per hour would have significantly shifted the arc 7 crossing to the North. This has a more dramatic impact on arc to arc routes than just shortening the line toward 38S and shifting it over.
The total number and duration of each 360 between neighboring arcs is limited by the arc spacing because the aircraft has to cover the each distance.
@ArthurC
Think of a 360 as the aircraft just stopping – fuel is consumed but no progress is made. The BTO arcs reflect an actual possible position at the time.
So between arc 2 at 19:41 and arc 3 at 20:41 the aircraft was flying for 60 minutes. If it spent 15 minutes in a circle, the leaves on 45 minutes to transit between the two arcs. This eliminates many possible arc to arc connections.
The DSTG work never allowed for any intentional holds.
A pilot that wanted to burn all of the fuel, but not simultaneously achieve the maximum range could break this up by just doing a 360 every hour or so.
The airplane could be in a circuit at the arc time and I do not know what that does to a BFO.
An experienced pilot that wanted to mess with max range max endurance simulations would mess this up and plan for it by doing periodic turns or even s-turns. The idea is to burn fuel but not go to max range.
This is not unlike the recent holding done by a Gulfstream that blew tires at Teterboro and circled over Bridgeport CT to burn fuel prior to landing at Stewart.
@Hank
Here’s what I am struggling with:
If this was an intentional flight to hide the crash-
I’d just keep the SDU off to avoid relaying flight duration data…
unless, it was intentional and I wanted to let people to know it was still flying. In which case, your maneuver or more simply turns could confuse the crash location issue.
@TBill
Maybe Z did not know about the SDU pings. He may have thought that by switching ACARS off this stopped the reporting. But he clearly would know that flying to maximum range to burn fuel adds some predictability to a future Boeing simulation.
And if he did know about SDU pings what a mystery he created by allowing everyone to know the plane was inthe air to 0019 UTC.
If he did the s-turns or period hold (as I think Z could have done) it clearly worked to confuse DSTG.
Their maneuver model limited turns to between -180 and +180 degrees at exponential distribution between each turn. Their particle filter mostly tossed out everything but straight southbound routes with a center at 38S latitude. Add lazy s-turns or a 15 minute hold in each one hour arc to arc segment and the course would be more southeast than directly south.
Over 25S maybe???
@Andrew: a BIG thank you for your detailed answers.
PL
@Hank
I recognise that you are one of the few who advocate deliberate decoupling of range from the known (or assumed) endurance.
You noted that “Since there was plenty of time it is easy to just hold every hour (to avoid a slick on crashing and avoids a vapor trail of a dump).”
Sure it is possible but unless the ping ring times etc were ‘known’ then the pilot would have to be quite lucky to avoid a clash with the periodic holds.
To my thinking a vapour trail would not be a bad thing but rather another ‘prop’ to support a ‘wounded bird’ interpretation.
After all, ZS did seem rather exercised about the fidelity of the depiction of fuel dumping within his flight simulator.
Slightly further down the rabbit hole one can simply assume that the report by Kate Tee was accurate:-
https://foxtrotalpha.jalopnik.com/the-ten-best-f-111-pig-dump-and-burn-moments-of-all-t-1601915237
@TBill, @Richard Godfrey, @DennisW, others:
I probably was not clear when I asked the question. I didn’t mean to ask for the reasons why the plane might be found south of 25S. I was asking if there is any evidence that taken on its face value refutes the hypothesis that the plane impacted north of 25S. I know of two that have been proposed: some of the drift studies, and an interpretation of the BFO that permits little error at 00:11. There are reasons to doubt the validity of both of these. Are there any others?
@Victor
Sorry for being slow.
Apart from the 2 reasons you mention, I do not know of any other evidence taken at face value, that refutes the hypothesis that MH370 is north of 25S.
@Hank
Thank you very much for the details. I do understand the mechanics and the reason for the 360 loops, just not how that would not be obvious plotting the arcs.
If it’s an argument for searching farther north on the 7th arc, then it makes sense. IIRC, the southernmost possible point was calculated based on maximum range, but with such “holding” loops, it would certainly move the final point farther north.
@Victor
I can’t come up with anything else.
I will get back on my BFO soap box and say that the Inmarsat 7Hz criterion is nothing more than a comment to a graph, yet it is cited over and over again as some sort of authoritative reference.
@flatpack
Z would not need to know about any ping times. The ping could happen in the middle of a 360. Assume that ZS needed to fly to around 0019 UTC to consume all of the fuel. There is no rush to dump it because he has the time. S-turns or a complete hold have the effect of decreasing the available time to complete the arc to arc crossing. The effect is that a more southeast course would result than a more south course to meet the arc BTO constraints.
The entire ATSB/BSTG search analysis assumed that there was no active pilot that would have intentionally engaged in period s-turns or holds while enroute.
My other issue is the water entry problem. ZS would have thought about this. The objective is to enter that water with the least breakup.
An uncontrolled aircraft would enter a tight spiral dive and have a near vertical high speed entry – maybe slightly inverted. A flat ditch approach at the end could just breakup and scatter on waves. Is it possible to fly a maneuver to enter vertical at 60 knots. A low speed vertical entry or into a wave.
Could I pull out of a terminal dive to level to the water and then do a loop over to get to a low velocity vertical entry. I would think a 60 kt entry would be better than 600 kts?? What would ZS do. The SEA Q400 ramp worker did some impressive flying – what could ZS do?
The ZS problem was to burn up all of the fuel and enter the water with the least breakup in a well executed terminal maneuver. No rush to go far so do some s-turns or holds along the way. That’s a good plan because it worked.
4 years and not a clue.
@Victor
We of course also have the NZPG simulator cases.
But we also have the straight flight/high altitude appearance of the data (by NTSB/Inmarsat/IG/and so on). Not saying it is true, just saying.
@ArthurC
The arcs are what they are and have nothing to do with the route. These represent all possible locations that the AC could be located at the exact time. How the aircraft could go between any specific point on one arc and feasibly get to any specific point on the subsequent arc is the problem.
From any arbitrary point on arc 2 to any arbitrary point on arc 3 you could eliminate any pair that would require a ground speed of Mach 1. That reduces many possible points. If you also assume that the plane can only fly in straight lines between arc points you could delete any point pairs that were below some ground speed. If you assume that Mach was always greater that say 0.7 for expected winds aloft, you could drop the routes too close. Here is where an s-turn or 360 creates havoc because the time delay makes the close route acceptable. The Inmarsat and the DSTG models just eliminate all of these “too short” routes (which would be in the SE direction).
The plane did need to get between the arcs at the exact time, but it is the complete elimination of all “too short” transits from consideration that worked against the over 25S location.
By not ever considering any intentional delay behavior by s-turn or holds the investigators biased the search to the south. Maybe this is OK because ZS never performed any delaying tactic.
And all of the Boeing work just simulated an uncontrolled “splat.” Anybody that’s been to an airshow has seen great low altitude maneuvers following a death spiral or spin.
@TBill: The fact that the BTO/BFO error is satisfied by a straight flight at cruising altitudes and speeds does not contradict the paths further north than 25S. The data agreement with straight paths has been used by some (including me in the past) as justification for prioritizing paths ending further south than 25S. Without any other evidence, simpler is better. However, we now have ample evidence now that is contradictory to an impact further south.
I’m looking for other evidence that taken on face value contradicts an impact further north than 25S, not evidence that says impact points further south than 25S are higher probability.
Taking this to the extreme, then the final point might even be much farther north, just enough to satisfy the shortest possible distance that still matches the points of crossing each arc.
@Victor
I am of course favoring north of 25 South myself.
However, I also suggest simuator data shows about 30 South@Arc7 heading towards Dordrecht Hole. Conceivably the real MH370 flight headed due south and then picked up the NZPG operating line as per the actual ZS simulator run. Overflys reasonably close to DrB’s point.
I am going to work that idea a little.
Since the BFO/BTO was a new application for the Inmarsat data has anyone compared the pings and the trajectories of actual flights? How do the pings compare with radar? Flying over land the pings could be used to calculate a route and then compared to the radar traces. If you only had the pings to calculate a route how close would they match the radar route? How does it play out in real flights? And what do we know about frequency variations from real flights?
@flatpack said: Re evidence that suggest the impact might not have occurred further north than 25S.
In the limit, going further north risks encroaching on Indonesian radar coverage if a great circle path is used.
A more subtle objection is that ‘further north’ implies ‘slower flight’ and probably ‘lower flight’.
If you consider any reasonable range for the Indonesian radar station at Lhokseumawe, Sumatra, MH370 targets should have been captured before reaching Penang Island and up the Malacca Strait. The fact that Indonesia has supplied no radar data indicates that either the radar station was not operational that night, or targets were captured but not shared. Either way, I don’t see how you can realistically use Indonesian radar as a way to qualify potential paths.
I also don’t think that slower, curved paths can be dismissed. As I said previously, all other things being equal, a straight path at cruise speeds might be preferred, but at this point, all other things are not equal.
@Trip: I refer you to the DSTG Bayesian Analysis, which presents the uncertainty of the measured BTO and BFO values based on data from previous flights.
This is a “hypothetical thought exercise” only – for those considering some theoretical way of putting a 777 cleanly into the water, without much structural breakup.
Look at the video of the crash of Czar52, the B52 crash at Fairchild Air Force Base. It was a near vertical, low speed impact, restricting the debris field to a very confined area.
A deliberate maneuver, to achieve minimum velocity water entry, for a B777, could be similar.
Method.
Enter a deliberate, low level, low speed, nose up, tight wind up left turn, with left engine at max power, and a boot full of top rudder (right rudder).
This leads to high bank angle, “g” coming on, and speed washing off, leading to an accelerated stall break, with resultant free fall, from a low height.
Immediately cut left engine and apply a rapid boot full of bottom rudder (left rudder – to swing the nose down) at the break.
This results in rapid roll left, probably going inverted, with rapid pitch down, resulting in relatively clean, near vertical water entry, at virtually nil forward speed, and perhaps as little as 5,000 ft/min (50 knot) vertical speed.
In such an entry, wing trailing edge components could be liberated with little further damage.
@Victor, others. In my flaperon separation analysis, in the Introduction and Summary at page 1 I said of the DGA ditching hypothesis that it was not supported by the evidence. The hypothesis depicts the flaperon as deployed. At page 8, repeated on page 9 I added that the hypothesis was unlikely to support a ditching even if not deployed. That all stands.
In the conclusions at page 16 I said that a ditching was ruled out. Since by then I had introduced an alternative hypothesis, implicity that too did not support any ditching.
However because a flaps-up ditching can be supported by the alternative hypothesis if during it the right wing breaks upwards and backwards, I now have added a paragraph covering that to page 14.
So instead of a ditching being ruled out the conclusion should have said that a ditching under the DGA hypothesis was ruled out as was a flaps down ditching under either hypothesis. It has now been amended to that effect.
My apologies for the error.
@TBill. “…also maybe consider France as we know they have active MH370 effort.”
Thanks for the idea but I doubt they will renew activities in this field.
@David: On Twitter, @PatrickTissot, who has contacts with official French investigators, claims the military have never stopped investigating the disappearance, although they have chosen to remain quiet.
@Victor. Thanks. I rang the embasssy here this morning and they said someone from the military would call back. I did not imagine there would be interest but maybe there will be after all.
I have asked the ATSB if they retain any interest.
@All
Reading this somewhat amused that so many here seem to be giving Tim R’s ‘hypothesis’ serious consideration.
How do you reconcile a hypothetical failed negotiation that then leads to the disappearance of the aircraft AND to the deaths of all 239 souls on board?
I have heard from Dennis in the past that it was necessary for Z to be prepared to follow through on his ‘bluff? If so, this would paint Zaharie in an even worse light. A true monster who would have deliberately and calculatingly murdered 238 human beings ostensibly sans mental illness, as he had a cadre of conspirators diabolically plotting with him.
And if there were an intervention on board (breached cockpit etc.) that thwarted Z’s plan, certainly we would by now have located the POI and corresponding debris field, as a deliberate disappearance is no longer on the table with said scenario. And furthermore, surely the debris finds would be far more numerous than the relative paucity of finds to date?
This whole negotiation thing is bollocks imho and makes absolutely no sense whatsoever. And it flies in the face of what the evidence to date strongly suggests occurred…a cunning, brazen and callous final fit to secure infamy and forever a place in history.
Seems he did quite well for himself.
@Donald
And it flies in the face of what the evidence to date strongly suggests occurred…a cunning, brazen and callous final fit to secure infamy and forever a place in history.
What evidence are you referring to, Donald?
[References to “David” were changed to “Donald” to avoid confusion.]
@Donald writes that the demise of MH370 involved “…a cunning, brazen and callous final fit to secure infamy and forever a place in history.“.
I accept that such a notion has become widely held, I just don’t hold it.
The evidence to date may be interpreted that the execution of a plan failed, initiating a paniced and unforeseen third option: no return.
@Donald
I am casting a wide net on what happened, I do believe pilot was probably involved. Whether it was straight suicide or large plan, I am open.
@Don Thompson said: The evidence to date may be interpreted that the execution of a plan failed, initiating a paniced and unforeseen third option: no return.
The third option might have been foreseen but not preferred.
@Victor Iannello
“If you consider any reasonable range for the Indonesian radar station at Lhokseumawe, Sumatra, MH370 targets should have been captured before reaching Penang Island and up the Malacca Strait.”
Indeed, you are correct. Unfortunately I did not make it clear that I was thinking about radar stations covering the south coast of Sumatra and Java, plus the Sunda strait.
An example on Sumatra would be Sibolga, which has a reach of approx 600 Km to the southwest.
@flatpack: I am not aware of a radar station at Sibolga. Can you please provide any details you have?
But assuming a radar station is there, a more reasonable radar range is 250 NM, not 324 NM (600 km). At 250 NM, MH370 targets near Penang Island would have been captured, and targets along flight paths to the SIO would not. (The path I generated that overflies Cocos Island is about 270 NM from Sibolga.)
I think we will be hard-pressed to use Indonesian radar to draw any conclusions about possible paths of MH370. I really wish this wasn’t the case.
@TBill
I agree with you that a piloted flight to water entry is a very possible scenario and cannot be ruled out until the aircraft is found. The result is the same whether ZS did it for reasons of mental health or to embarrass Malaysian government. In either case the objective is to hide the aircraft forever and the mission is very carefully planned to water entry maneuvers.
It has been proposed that a mechanical failure occurred at IGARI forcing a diversion toward Penang and it became a ghost flight to the SIOat some point. This is the basis of all of the ATSB search effort where a piloted flight has not been considered past Penang or maybe MEKAR. It’s a big ocean so this can’t be ruled out.
The hold and negotiate and conditionally kill everyone, for me, is a stretch.
If ZS just did a few lazy circles on a SE route and executed a clean water entry after fuel exhaustion north of 25 S latitude he would have placed the wreckage on the sea floor where no one has searched. Maybe the drift analysis supports debris originating near the 7th arc above 25S.
@Victor Iannello
Re Sibolga, the information came from a comment (including coverage diagrams) made by Richard Godfrey on your previous article. All credit to him:-
http://mh370.radiantphysics.com/2018/07/30/mh370-safety-report-raises-many-questions/#comment-17756
I believe, but cannot demonstrate, that there is Indonesian military radar coverage of the Sunda strait.
Civilian radar coverage of the west of Java (Jakarta and Bandung) may not be as useful as one would hope because of the mountainous terrain to the south.
@Victor,
TNI-AU has 4 long-range ADS sites in Sumatra: Dumai, Lhoksuemawe, Pulau We/Sabang, and Sibolga (SATRAD 232, 231, 233, and 234 resp).
The Sibolga site is actually 25km NW of Sibolga City, near Sorkam. I located it a few months back and it was noted here. The location is N-1.9096° E-98.6179°.
@Don Thompson
Do you have any info re coverage of the greater Sunda strait region?
Given that Jakarta is the capital, one would think that there would be an extended air defence system, possibly utilising outlying islands.
@Hank
OK thank you.
To quote one far-sighted TV contributor, finding MH370 is like eating an elephant, one bite at a time. All we can do is help decide where the next bite should be.
@Don Thompson, @Richard Godfrey, @flatpack: Thank you for the information about the additional radar sites in Sumatra. Those are additional sites that didn’t see MH370 as it flew past Penang Island and up the Malacca Strait.
So either Indonesian radar was operational and detected MH370, and the military is quiet because there was no military response, or the Indonesian radar was not operational and did not detect MH370, and the military is quiet because there was no military response.
That’s about where we were four years ago.
@flatpack
You’ll be interested in TNI-AU Kosekhanudnas I.
@Donald: “And furthermore, surely the debris finds would be far more numerous than the relative paucity of finds to date?….”
I don’t think you quite spelled out your argument. But please note that it’s easy to demonstrate that very, very few people randomly encountering a “non-charismatic” piece of debris [eg, everything but the flap and flaperon bodies” took any notice. More than 50% of debris items were found by Blaine Gibson’s directed search. On a very tiny sample of IO coastline. Even with expert advice, it’s unlikely that as much as 5% [more likely closer to 1%] of debris that reached shoreline [and no more than 10% or aircraft exterior was capable of floating more than a few weeks].
I have to remember to harp on this every couple of months.
Part of that dropped out; I meant to conclude: “Even with expert advice, it’s unlikely that as much as 5% [more likely closer to 1%] of debris that reached shoreline HAS BEEN RECOVERED [and, FURTHER, no more than 10% or aircraft exterior was capable of floating more than a few weeks].
@lkr: Based on Blaine’s success at finding debris, my takeaway is that there was a lot of debris generated and few people searching and/or having the ability and/or having the desire to report potential parts.
A question about supposed End-of-Flight Inmarsat data (ref. document is “Update to Signalling Unit Logs December 2014” PDF). Sorry for the length of the post…
The last transmission of the “normal” period after take off was an “Acknowledge User Data” at 17:07:48.907. Normally, in the absence of a Log-Off from the AES, a Handshake Request from the GES was due around 18:07. But there was none. Why ? Probably because from 18:03:41.405 to 18:04:29.413 there were transmissions from the GES of the ACARS from MAS requiring an answer from MH-370, then an automatic retransmission from 18:05:11.405 to 18:05:59.905.
After the 5 Request to Acknowledge (by the GES) from 18:03:50.905 to 18:04:29.413 without answer from the AES, and another 5 (automatic retransmission) from 18:05:11.905 to 18:05:59.905 also without answer from the AES, the AES of 9M-MRO was probably written off the system***.
At 18:25:27.421 the AES of 9M-MRO began a Log-On request. The sequence ends at 18:28:14.904, after 26 lines of data on the Inmarsat log.
It is largely admitted that this Log-On request happened because the SATCOM has been unpowered and powered again.
This request around 18:25 was from an AES written out.
At 00:19:29.416, began a new Log-On request from the AES of 9M-MRO. It is also largely admitted this was because the SATCOM was unpowered and powered again. But this time for some minutes only, so the AES was not written out the Inmarsat system.
There are only 8 lines of data, finishing by a transmission of the GES P-Channel: TX OxlS – Log-on/Log-off Acknowledge [copied from the data log where “OxIS” is probably a misreading by the OCR for 0x15].
This Handshake as been described as “incomplete” because of the SATCOM was unpowered again, or because the antenna lose the LoS of the satellite.
The line 8 of the Log-On sequence at 18:25 is the same as the 8th and last line of the Log-On sequence of 00:19.
Could the difference was not because of an interruption of transmissions, but because the Log-On of 00:19 was different: the AES of 9M-MRO was not written off at 00:19 and, perhaps, in this case, the data exchanges between the aircraft and the ground would be shorter?
PL
*** BTW, it is probably why there is not other data from Inmarsat about these ACARS message from MAS in the report: 9M-MRO being written off the system, the “failed” transmission came directly from the GES to MAS, without trying to “ring” the plane.
@PaxLambda
The 00:19 Log On has been described as incomplete, but strictly that is incorrect.
The GES Log On did complete at 00:19, hence the two Log On Acknowledge SUs exchanged between the AES and the GES.
However, the GES Log On was not followed by any additional signalling to establish ACARS comms with the ground (neither did that happen at 18:25), nor was it followed by the two ISO-8208 SSN connections (which did occur after the previous GES Log Ons).
Regardless of whether, or not, the GES has the AES marked as active in its Log On Table, the AES executes the same sequence of actions: GES Log On; ACARS establish; opens the two ISO-8208 SSN connections for the IFE applicaitons. A Log On doesn’t assume any prior state is maintained.
@Don Thompson: Thank you for your answer.
What did you think:
– about the absence of Log-On request from the GES around 18:07: is it because 9M-MRO has been written out as I wrote?
– about the retransmitted ACARS from MAS which has been described above is this thread as “missing” in the report? Could the 18:43 a typo for 18:13? Or the AES was written off? But at 18:43, it was on line again. Could be because, then, the flight ID was no more there?
… Perhaps, too many questions: hope I don’t bother you!
PL
@PaxLambda
About “the absence of Log-On request from the GES around 18:07“. The GES does not make Log On Requests. At 18:03:41, the GES assumes the AES is still listening (there has been no implicit or explicit Log Off for 9M-MRO’s AES since 17:07:48) so it makes the transmission of the first ACARS block, 220 chars, of the ‘URGET REQUEST’ message.
About the “retransmitted ACARS from MAS“. The GES transmitted this message twice, without acknowledgement, as a consequence of the message submission into the system by MAS Ops. Failure of the second SATCOM transmission resulted in the GES clearing -MRO’s AES from its Log On table. Accordingly, the ACARS hub lost its known path to forward messages to -MRO. So, the subsequent events are message re-submissions into the system from the originating terminal.
About the “typo for 18:13“. The final message re-submission described in Appendix 1.9A is timed at 18:15:13, a typo for 18:43 isn’t consistent. And “the AES was written off?, the GES cleared the AES from its Log On table at 18:06. The AES made a successful GES Log On at 18:25:35.
About “at 18:43, it was on line again“. The AES had successfully completed a GES Log On at 18:25:35, however, the aircraft Data Communications System did not initiate ACARS correspondence. It is the aircraft that initiates ACARS correspondence when a datalink is established. The AES, and its SATCOM datalink, is merely a medium for ACARS correspondence. It is the Data Comms System (and its component functions) that initiates/responds to ACARS messages, and it has been explained how access to SATCOM can be disabled for ACARS.
The use of a Flight Identifier in the Log On Request is not involved with ACARS, it’s purpose relates to an Inmarsat network management function. It’s of no consequence to packet data operation (ACARS and ISO-8208 SSN).
Returning to Appendix 1.9A: the Safety Investigation Team has presented two versions of this information, one in 2015, and the other in 2018. Neither satisfactorily explains “messages are auto transmitted every two minutes […] 1843:33 UTC“.
@Don Thomson: Thank you again for these very detailed and informative answer.
About the “messages are auto transmitted every two minutes […] 1843:33 UTC“, I agree that a typo is inconsistent vs the times values.
Is it possible these transmissions were only between MAS and GES, or even a subsystem: the first went along all the system to the satellite, but once 9M-MRO AES was cleared from the GES AESs table, and perhaps after some delay***, the retransmissions were rejected (> failed) and then not recorded following the ones seen in the reports, but, perhaps somewhere else… or not recorded at all.
***If this has happened, some delay had to exist between queries times and their “rejections” and also for explaining why, if retransmissions persisted between 18:25 and 18:43, those ones are not recorded in the data logs.
Even if I believe Inmarsat has nothing to hide and their data are right, I have to admit there is something suspicious on Malysian side if some data have been put aside.
Regards,
PL
@Pax Lambda said: Even if I believe Inmarsat has nothing to hide and their data are right, I have to admit there is something suspicious on Malysian side if some data have been put aside.
In addition to the missing traffic between 18:15 and 18:43, there are some red flags with the ACARS Traffic Logs that need to be explained by Malaysia.
For the Traffic Log in the FI, the choice of filter parameters ensured that the VHF traffic was not included. On the other hand, for the Traffic Log in the SIR, the VHF traffic was included on the first page, but the filter parameters changed starting with the second page to exclude VHF traffic. It appears that this log is actually an assembly of two different reports. In addition to the difference in filter parameters at the top of the pages 1 and 2, you can see that the messages received by MAS at 15:54:41 and 15:54:53 appear on both the bottom of page 1 and at the top of page 2.
What we need is a complete log of SATCOM and VHF traffic for the time period 12:48:00 to 20:00:00. Until this occurs, we can’t be sure there wasn’t other traffic that was not included in the logs.
@Victor
“The third option might have been foreseen but not preferred.”
The fourth possibility could be that ZS actually used the @TimR story as a ploy to get pre-flight “assisitance” he needed. The assistance would be (1) not letting Razak claim the accident was a mechanical failure/crash (also turning on SDU at 18:25 accomplished this task); and (2) aside from potentially wiping CVR, one step better than SilkAir, he may have left behind “knowledge” that this accident was not an intentional suicide. Possibly other assistance.
So here is the list:
@TimR Story Outcome Matrix
1. Peaceful Diversion
2. Peaceful Diversion that crashed; something went wrong on aircraft
3. Peaceful Diversion with planned suicide option if demands not met
4. Peaceful Diversion was a false cover story for a suicide plan
5. Peaceful Diversion plan was true, but plan was later changed to suicide
P.S.-
6. Peaceful Diversion is an urban myth (untrue after-the-fact rumor)
Victor said: I didn’t mean to ask for the reasons why the plane might be found south of 25S. I was asking if there is any evidence that taken on its face value refutes the hypothesis that the plane impacted north of 25S.
More you go North, and more you go East: doesn’t there would be a change from IOR to POR as it happened for MH-371 (and even before take-off at KL for MH-370)?
Not a very strong argument, but…
PL
@Don Thompson
Thank you, that is indeed interesting!
@Victor Iannello
My take is that there are multiple radar sites on Sumatra (and from the link supplied by Don Thompson likely also in Java) that have significant coverage to the south and southwest. I consider that these constrain ‘northern’ paths.
Please check out the map in the link provided by Don. It is nicely divided up into ‘named’ rectangles. At the bottom of the map is Christmas Island. There is also what seems like a radar station on the southern coast of Java that looks interesting.
@PaxLambda
We know that MH370 was logged on to the Inmarsat satellite IOR until 00:19:37 UTC.
At all points along the 7th Arc in the Indian Ocean, even right up to the South coast of Java at 8°S 108.0145°E, MH370 would be much closer to the IOR satellite, than the POR satellite.
@Pax Lambda: Once logged onto a satellite, automatic handover to another satellite requires degradation of the P-channel signal or an elevation angle below the threshold. That never occurred for MH370, independent of overlap or relative proximity. MH370 would have to be east of 121E longitude to be closer to POR than IOR. That is well beyond the 7th arc, even as the arc crosses the equator.
@Victor and Richard: OK. However, I just look at MH-371 inbound flight data***.
For this flight, the last connection with POR was at 06:48, the plane was then around 105°E. The change from IOR to POR has been at 06:11. At this time the plane was around 107°E. So, it is possible there was a change from IOR to POR largely west of the 121°E longitude.
But transition from IOR to POR not happened west of 107°, so I agree my argument is not a good one… May be 9M-MRO is North of 25° on 7th arc after all…
***Data from unredacted Inmarsat Logs Excel xlsx file authored Greg West. Correspondances between time and longitude taken from decoded hexadecimal SSUs PDF published by Richard on June 12th 2017 (thanks a lot, Richard for all your studies!).
Regards,
PL
@PaxLambd asked “Is it possible these transmissions were only between MAS and GES”
No, the particular GES being used by an aircraft is entirely transparent to the MAS ground user who submits the message for relay to the aircraft.
ACARS relies on a ‘store-and-forward’ technique. After 18:06, the main ACARS hub in SITA’s network (addressed via QXSXMXS) had no knowledge of any path over which to forward ACARS messages to 9M-MRO.
So, from 18:06 to 18:25, 9M-MRO was “discarded” from the AESs table of the main ACARS hub in SITA’s network, and later, when 9M-MRO was again in line with the satellite (after the Log-On of 18:25), there was no more a way to forward the ACARS because the ACARS link has not been activated in the 18:25 Log-On of 9M-MRO. And for the Log-On didn’t include ACARS link activation, the only way was to uncheck the “option case” SATCOM of the ACARS screen?
From the fact that at 18:03 the ACARS was effectively transmitted to 9M-MRO (without success besause the SATCOM was unpowered), can we conclude that at this time, the SITA’s network didn’t “know” that ACARS transmission to 9M-MRO was “out” and therefore unpowering the SATCOM had been done before unchecking the SATCOM “case” for ACARS?
And, therefore, unchecking the SATCOM “option case” for ACARS has been done knowingly before 18:25 because powering again the left bus was (also) allowing ACARS transmission!
With the SIO simulator points and this, is it really possible to consider this was an “accident” or an “electric problem”?
PL
@Pax Lambda: Along similar lines to your last comment, when a pilot gets handed off from one controller to another, common practice is to dial in the new frequency as the standby, acknowledge the handoff and read back the new frequency, hit a button on the radio that exchanges the active and standby frequencies, and call ATC on the new frequency. The time between the two transmission would be less than 15 seconds. It was 13 seconds at MH370’s previous controller handoff.
At the handoff between Kuala Lumpur and Ho Chi Minh, the pilot acknowledged the handoff, but never read back the new frequency, and the call to Ho Chi Minh was not made. That leaves a window of only about 15 seconds for the abnormal event that caused the turn back.
@Victor: An interesting observation. I would rather say that during the handoff the sequence of abnormal events has already taken place. The situation initially was not too serious but it quickly evolved to the critical point when the transponder was switched off.
@Greg: I think you may be missing my point. There is a mountain of circumstantial evidence that points towards a deliberate diversion by the captain.
@Victor: Of course, you are right. But I remember that you were interested in the motivations and role of the human factor. Perhaps that is why I am suggesting possible (but not certain) explanation of known facts. By the way, switching off the transponder is circumstantial evidence. It is certain that the signal has ceased to be received.
@Greg: Yes, what I said is There is a mountain of circumstantial evidence that points towards a deliberate diversion by the captain.
@Victor. I’m presume that two of the items of circumstantial evidence referred above are:
1. Simulator data
2. Timing of diversion
But that wouldn’t amount to a “mountain”. Would you mind listing the items that you consider circumstantial evidence of deliberate diversion? For present purpose, “deliberate diversion” taken to refer to diversion with nefarious intent and not an accident/systems failure.
@VictorI
With apologies for my late contribution; I’m having a busy week..
“Let’s all try to develop a list of evidence that suggest the impact might not have occurred further north than 25S. I’ll start.”
In addition to your point (1) (“Some drift model results predict the debris arrives to East Africa too early. (Some drift model results are consistent with a more northern impact.)”):
The same can be said for the arrival time of the Flaperon at Reunion
A more general comment:
While imo there are strong indicators for an impact south of 25S, there is no watertight evidence “against” an impact north of 25S. For the Inmarsat data we know that for many years. For the drift studies, even if the methodology followed is in accordance with state-of-the-art, there are inherent inaccuracies and the results are expressed in probabilities.
Remember, however, that before the OI search started most of us here were optimistic and probably many contributors would estimate the probability that MH370 should be located somewhere between S36 and S25 “near” the 7th arc perhaps as high as 80 -90%.
So for the moment, I remain at the position that MH370 has not been found in the area where we expected it to be located. Hence my call for taking our time to collect as much as possible of the relevant information which would perhaps be within our reach.
Some ideas for additional analysis which would advance our understanding and which could help us to formulate a better founded search recommendation:
– A detailed estimate of the detection probabilities in previously searched areas
– An in-depth meta study of the different drift models to better understand the different results and get a better feel for sources of inaccuracies.
– A review of the BFO analysis performed by DSTG on the 20 earlier flights of 9M-MRO
– (Better understanding of the BFO calibration process (timing, calibration target etc.): perhaps less urgent)
– (Analysis of possible end-of-flight scenarios: already ongoing)
@Niels said: While imo there are strong indicators for an impact south of 25S, there is no watertight evidence “against” an impact north of 25S.
That seems to be the consensus here.
So for the moment, I remain at the position that MH370 has not been found in the area where we expected it to be located.
That’s true. The locations remaining were not long ago considered by most here to be low probability.
And that’s the reason that I have not made a recommendation about where next to search. I am holding out hope that new evidence or insights surface. Perhaps drift modelers will converge on a smaller area of the 7th arc. Perhaps we learn more about communications between the parties that helps to eliminate some scenarios. I will say that my patience is not infinite. It is possible that with the evidence we have in hand, we have taken this as far as we can go in bounding where the debris field is.
@Paul Smithson: Perhaps others are willing to compile a detailed list making the case for pilot hijack. At one point, @DrB was attempting to do this.
In a nutshell, putting motives and even the sim data aside, there is no evidence that cannot be explained by a pilot diversion just after the last radio transmission, and no accident sequence that comes close to being a likely explanation for the evidence.
@Niels
You stated “For the drift studies, even if the methodology followed is in accordance with state-of-the-art, there are inherent inaccuracies and the results are expressed in probabilities.” I agree.
I am currently running a re-calibration of my drift simulator software having made some changes to remove all data prior to the year 2000 from the database of undrogued drifters, which Griffin points out are out of alignment and too slow. I have found 10 transoceanic drifter buoys from the years 2009 to 2017 that passed the 7th Arc at various latitudes from 13°S to 33°S and various months including Jan, Mar, May, Jul, Sep, Nov and Dec. I will compare their tracks with the tracks predicted by my simulator software.
I have analysed the latest update and conclusion from Griffin, with results up to 8°S along the 7th Arc. The conclusion is still based on one item of debris arriving too early in Reunion, “According to our model, the flaperon’s arrival at Reunion is not consistent with crash sites north of 25°S.” By contrast, Griffin’s Fig. 1 still shows high windage Items arriving in Africa from 23S northwards are much more likely, than latitudes further south.
There are 27 items of debris to fit into an analysis, not just one. What if the windage of the flaperon does not fit the Fig. 1 scenario with 10 cm/sec at 20? Low windage items in Griffin’s view (see his Fig. 2) are more likely to reach Reunion from 17S to 20S and to reach Africa from 23S northwards. There were no ocean trials with the genuine flaperon, how can we be sure the windage was, as Griffin predicts. What if the flaperon was slightly submerged and had no or only low windage? The low windage results from Griffin show, in this case, 17S to 20S as more likely MH370 end points.
@Neils
@Niels said: While imo there are strong indicators for an impact south of 25S
Maybe your strong indicators are all based on drift analysis – I can’t comment on that.
But if you believe the ATSB/DSTG particle filter modelling, that model only works for a very special case with a live pilot at the controls. And the failure of the search based on the modelling suggests that assuming a ghost flight could be a bad assumption.
If you assume that ZS was at the controls to water entry, this makes a crash location above 25S much more likely.
[Comments here are closed. Please continue the discussion under the new article.]