If you were leading a high-profile international aircraft investigation, in command of the world’s most qualified technical experts and in possession of all the relevant data, would you bother listening to a rag-tag band of internet commenters, few of whom actually work in the space or aviation industry, and none of whom have access to all the data?
Most likely, you’d say: certainly not! But as time goes by, and the puzzle remains curiously impenetrable, you might find it worthwhile to pay a listen to what the amateurs were saying. You might even abandon some of your own conclusions and adopt theirs instead.
This appears to be the case in the search for Malaysia Airlines Flight 370, which disappeared en route from Kuala Lumpur to Beijing back in March. From the beginning, the authorities running the investigation — first, Malaysia’s Ministry of Transport, and later the Australian Transportation Safety Board (ATSB) — held their cards close to the chest, releasing very little information about the missing plane and maintaining a posture of absolute conviction. The investigators’ self-confidence reached its apex in April, when their methodology led them to an area of ocean where underwater accoustic signals seemed to be coming from pingers attached to the plane’s black boxes. Officials assured the press that the plane would be found in “days, if not hours.” But then it wasn’t. A scan of the seabed found nothing; the pingers were a red herring (perhaps literally!). Back to square one.
Meanwhile, on the internet, a group of amateur enthusiasts had come together from all around the world to trade ideas and information about the missing flight. The group, which came to call itself the Independent Group (IG), emerged from various online comment threads and eventually grew to about a dozen individuals. This was a truly spontaneous, self-assembling crowd: there was no vetting of credentials, no heirarchy of any kind. (Full disclosure: I count myself among this group.) Basically, if you seemed to know what you were talking about and could comport yourself in a collegial fashion, you were accepted into the crowd.
While the mainstream press was reporting the ATSB’s pronouncements as received wisdom, the IG was raising red flags. IG members were among the most vocal critics of the ATSB’s contention that the accoustic pings probably came from black-box pingers. And later, after a public outcry led Inmarsat to release a trove of data received from the aircraft, and the ATSB issued a report explaining how it had come to identify its current search ear, the IG dove into the new information with abandon, quickly identifying holes in the data and weaknesses in the official approach. In a pair of papers, the group recommended its own search area, hundreds of miles to the southwest of the ATSB’s officially designated zone.
Today, the ATSB has released an update to its earlier report, explaining why it has decided to reassess its conclusions and move its search zone to a new area — one that overlaps, as it turns out, with the IG’s recommended area. (In the graphic above, the white bracket shows the ATSB area; I’ve added a yellow dot to show the IG area.) Needless to say, this has caused elation within the ranks of the IG, who see the move as vindication of their methods, and indeed validation of their combined efforts over the last few months.
A few observations on the new report:
— One of the reasons the ATSB gives for the shifting of the search area is the recognition that Inmarsat data related to an unsuccessful ground-to-air telephone call attempted at 18:40 indicated that the plane had already turned south at that time. The IG had been basing its analyses on this data point for months.
— Since the June report, the ATSB has improved its BFO model by taking into account various factors — such as temperature shifts caused by the Inmarsat satellite passing through the Earth’s shadow and the mis-location of the Perth ground station in an important Inmarsat algorithm — that IG member Mike Exner has been working through in detail for months.
— The ATSB has fundamentally changed its approach in how it is assessing the plane’s likely path. In its June report, the focus was on what I call the “agnostic” approach: it generated a large number of flight paths based on as few initial assumptions as possible, then graded them based on how well they fit the timing and frequency data received by Inmarsat. This resulted in a population of potential flight paths that fit the data well, but did not make any sense in terms of how a plane might be flown. Some of the routes, for instance, involved multiple changes in heading and airspeed. Today’s report explicitly excludes such flight paths. The ATSB and the IG alike now assume that the last several hours of the flight were conducted without any human input — the crew were presumed to be incapacitated by hypoxia or other causes, so the plane flew on autopilot until it ran out of fuel and crashed. This has been the IG’s starting point for ages, and the fact that the ATSB has now adopted it is a major reason for why the two group’s search areas have now converged.
— You can see in the graphic above how an emphasis on matching the Inmarsat data will tend to lead you in one area (“Data error optimisation”) while an emphasis on routes that comport with real-world autopilot functioning will lead you to another (“Constrained autopilot dynamics”). To be sure, they overlap, but the peak area of one is far from the peak area of another. I think it’s important to realize this, because it helps us to understand why it has been so hard to get a handle on where MH370 went, why the official search area keeps moving, and why knowledgeable people have been furiously debating possible flight paths for months: the BFO and BTO data just do not match up that well. In order to arrive at its recommended area, the IG has been willing to accept much wider deviations from Inmarsat data points than the ATSB has been comfortable with.
— Finally, it’s worth nothing that the ATSB approach is superior to the IG’s in one important regard: it is at heart statistical, looking at families of potential routes rather than proposing and assessing one at a time. There is a tendency, as an individual–and I have fallen into this myself–to cook up a solution, run an analysis, and to be so impressed with the result that one wants to shout about it from the rooftops. (Ask me about RUNUT some time.) The IG has come up with a search area essentially by pooling together a bunch of individual solutions, each of which is generated by a different set of procedures and different set of assumptions. It’s a herd of cats. To really move the ball forward a more rigorous approach is needed, one that takes each procedure and sees how it would play out if the assumptions are methodically modified.
The upshot is that, since the early days of the investigation, the attitude of search officials has changed radically. Once dismissive of amateurs’ efforts to understand the incident, they have clearly begun to listen to the IG and to turn to it for insight and ideas. Indeed, you could say that since the release of Inmarsat data and the issuance of the ATSB report in June, the search for MH370 has become effectively crowdsourced: a de facto collaboration between the professionals and a spontaneous assemblage of knowledgeable experts.
UPDATE:
The overlap between the ATSB’s analysis and the IG’s is more evident in the image below, courtesy of Don Thompson. It shows the fan of values calculated by ATSB to match likely autopilot settings.
No guarantee that posts with multiply URLs will see the light of day in a timely fashion.
Rand – It’s ok, I’m no grumpier than usual and someone has to take up the slack with John’s departure. If it’s not in one of the high priority boxes then I reckon the game could be up and it will fizzle out. This why they have been widened – if I’m right on that? That is early evidence of a concentration of resources – to me.
Bobby – Thankyou thankyou thankyou….for asking those questions – systematic error. As a sat-comm layman I’ve tried to articulate that this(MH370-Inmarsat satellite) is an awful way to measure anything much. Systematic error I regard as a lot more likely than spoofing.
@Bobby: “Insofar as I know, there is no identified route that fits -3 knots per hour better than just a constant speed.”
There is a path that works well with one stipulation. There is a constant Ma number path with a due south (180 deg) track aligns longitudinally with the waypoint BEDAX that has intrigued me for months. Much as Inmarsat did in their recent paper, begin at 19:41 without regard to what happened before that time. Start around at 19:41(2.6N, 93.8E) and go due south at a constant 0.811 Ma number, adjusting air speed for temperature and ground speed for wind, and ending at (34.2S, 93.8E). (Inmarsat chose a constant ground speed, but not until 20:41. Their track was close to 180 deg but ranged between 186 and 179 deg.) The RMS range error is around 4.9 nm and the RMS BFO error is around 3.3 Hz, even though the path is not chosen on the basis of minimizing the BFO error. The speeds are also very close the LRC speeds.
If you assume the path between 19:41 and 00:11 was due south and constant Ma number, the longitude of 93.8E can be found by hitting the ping arcs without any other assumption about a waypoint, although it is consistent with the BEDAX longitude.
The one problem with this solution (let’s call it a feature) is that there are not a “simple” path that connect the position at 19:41 with the radar tracks that end at 18:22. So if this path was actually the one followed, it would require either a circling or a landing (a “loitering”) in the vicinity of the Aceh Province of Sumatra. Many people will reject this path because of this “complication”. I won’t argue and say this path is the “correct” one. Rather, I would say that this path is correct given the assumptions and constraints therein.
I prefer to not classify paths as “correct” or “incorrect” because each model is laden with assumptions. All we can do is subjectively rank our solutions based on the perceived likelihood of our assumptions. Using one set of assumption, I calculate the end point (34.2S, 93.8E). With another set of assumptions (a continuous path with one major turn)I get the end point I contributed as part of the IG. I stand behind both solutions.
Here are some questions I would pose about your path:
1. Do you have evidence that a plane that overflies a waypoint remains on a great circle path?
2. Your air speeds seem considerably higher than the LRC air speeds. Do you have evidence that you can achieve the required range at these higher speeds?
3. If I understand some of your statements and how you calculate the speed between ping times, it appears that you begin the speed descent due to fuel exhaustion well before 00:11 UTC. Do you believe that this occurred?
Again, I will not debate “correct” or “incorrect” because for all I know your assumptions could be more accurate than other solutions. No solution will be correct or incorrect until the plane is found.
Thanks,
Victor
Hi @Victor,
“The one problem with this solution (let’s call it a feature) is that there are not a “simple” path that connect the position at 19:41 with the radar tracks that end at 18:22.”
Would there be a “simple” path which would connect 19:41 with the radar track at the sharp corner or somewhere near, further along the incoming tangent?
Cheers,
Will
Hi @Jeff,
“No guarantee that posts with multiply URLs will see the light of day in a timely fashion.”
I remember a long delay for my post in response to JF’s “moniker” comment, with email, phone number, twitter handle, but no URL (I think).
Cheers,
Will
There are certain characteristics that make WordPress assume it’s spam, too many outlinks or contact information being prime among them, I assume.
@Bobby Ulich:
I agreed that using particular statistical tests can be questioned but I think it is a reasonable first approximation; I wasn’t proposing a threshold chi-squared statistic to discriminate if a track could be excluded, which is where the precision would be needed. The ATSB reports and the Paper are indeed short on statistical tests that might be applied (see my conclusion below).
Using the wide +/-7Hz (or 5Hz one sigma from the ATSB report) errors is a conservative approach but the BFO data then becomes unconstraining. There is information in the precise distribution of errors, though with few MH370 data points there is always the issue of demonstrating significance. There may be other sources of error apart from short term noise, but those should in principle appear in the Amsterdam flight data and be covered by an error measured from that flight. That data set indicates that the BFO is deterministic with remarkably low error, given it was not designed to be so accurate; possible MH370 tracks with low BFO r.m.s. error cannot be ruled out as anomalous.
In any event selection of tracks based on their statistical significance (however judged) may be a moot point as it appears the selection of the final search area by the Investigation does not use the detailed BFO data. The Southern limit is set by the Southern fuel exhaustion point. The Northern limit can be obtained by simple constant speed models and a maximum speed for the 18:29 to 18:40UT leg (so using the BTO data only). In both cases the relationship of the BFO data to its errors is not relevant – the statistical significance of the BFO measurements for any track is effectively not a criterion.
I suspect the investigation was not prepared to rule out a complete family of possible paths based on a few BFO errors, so have covered both bases. The search areas (that is length of the 7th arc) dictated by the two models are similar, not a factor of 10 different, so the cost of including both was judged manageable avoiding a difficult statistics-based decision.
Rand, Nihonmama, MuOne etc – Yes you are right, it is getting into beer on the veranda weather here in the Perth Hills. I’m on about acre with creek and a few different types of marsupials frequenting. In the event there is something to celebrate the chairs are ready.
Rand
Posted October 16, 2014 at 3:34 AM
“It’s interesting to note the logic of Sir Tim Clark, as revealed in the Spiegel interview. Despite the fact that he is an industry insider, he is apparently hung on the conclusion that the aircraft was under human control for the entire duration of the flight.”
Here are the (4) times that Sir Clark mentions “control” or “controlling” in the Spiegel interview:
(1) “My own view is that probably control was taken of that airplane.”
(2) “The transponders are under the control of the flight deck.”
(3) “We must find systems to allow ACARS to continue uninterrupted, irrespective of who is controlling the aircraft.”
(4) “MH370 was, in my opinion, under control, probably until the very end.”
Hi Rand ~
Would you please explain your understanding of Sir Clark’s intended meaning? Did he mean:
(A) under human “hand” control?
(B) under human “remote” control?
(C) a combination of (A) and (B)?
(D) to express intentional ambiguity?
Thank you!
~LG~
@MuOne: “Would there be a “simple” path which would connect 19:41 with the radar track at the sharp corner or somewhere near, further along the incoming tangent?”
I believe there is a path from IGARI towards Port Blair (Andamans) to BEDAX and south that would work. The plane would need to be descending around 18:40 to match the BFO at that time. Of course, that would mean rejecting the primary radar data. For this reason, I have not studied this path in detail.
@Victori
I would suggest trying some paths with a descent earlier, compatible with the eyewitness reports of the plane flying low and westward at Kota Bharu at 17:36 UTC (per Ulich’s chart, eyewitnesses merely said around 17:30), and with the celltower reconnect uncovered by the Malaysian police investigation and relayed to US authorities.
I note that there seems to be no BTO or BFO data for a considerable period around the Khota Baru airport-Penang airport traverse (17:36-17:52 in Ulich’s chart). The flanking data points come from 17:07 and 18:25. Looking at the “Example Reconstructed Flight Path Results” on page 20 of the Journal of Navigation article, there are only two abnormally high BFO error values attributed to the uncompensated vertical motion component due to the climb to cruising altitude: those are from 16:42 and 16:55, the effect disappearing in the 17:07 results. Note that those data points are all pre-diversion, so presumably informed by real, not modeled data. My inference is that there would be no clues in the available BFO data regarding a descent to low altitude during the overflight of the Malaysian peninsula.
@Luigi: The large errors during the initial climb come way down when the ascent is properly modeled in the BFO prediction. (It baffles as to me why Inmarsat does not correct this in their model as it is relatively easy to do.)
As for other paths that ignore the radar data, I will allow others to take a lead in those calculations. Until I have evidence to the contrary, I will accept the radar data as valid, as both Thailand and Malaysia reported to seeing the plane.
I do believe, however, that we do not have all the raw radar data that is available. The Thai and Indonesian radar installations should have tracked the plane well after 18:22. Having the radar tracks around Sumatra would help tremendously in narrowing the search area and helping to explain the disappearance.
@Victori
Of possible interest regarding the BFO values from the data associated with the satcom reboot, the first data to follow the Malaysia overflight:
“The spike in the measured data at 18:28 is not fully understood and was originally ascribed to a possible manoeuvre of the aircraft: although it could be related to frequency changes during the logon sequence described in Section 3.3.”
So, they discarded the BFO data from 3 minutes after the satcom reboot (18:25) for the purposes of model-fitting, citing the unreliability of post-Logon Request BFO values. However, given the spike is positive, it could also suggest that the plane was put into a climb concomitant with the satcom reboot, which also happens to be concomitant with the plane leaving military radar coverage (last primary radar fix was at 18:22).
@Victori
>> The large errors during the initial climb
>> come way down when the ascent is properly
>> modeled in the BFO prediction. (It baffles
>> as to me why Inmarsat does not correct
>> this in their model as it is relatively
>> easy to do.)
I also thought it was surprising they didn’t figure that into the example data set — it should have been an easy calculation and would have made the fit look better. The vertical motion is uncompensated here at two levels: the plane’s electronics don’t figure it in adjusting the signal frequency, and here the analysts did not compensate for the associated BFO shift in their modeling. Anyway, my point was that, because they didn’t, we have some idea of the magnitude and duration of the BFO changes that could be associated with a large change in altitude. And, we don’t have the data to speak to the question of whether the plane descended during the Malaysian overflight, as indicated by other lines of evidence. Unless, that is, the discarded BFO data from 18:28 (which prima facie suggests a climb following satcom reboot), is actually valid.
LG: THAT was a very well crafted, rhethorical multiple choice question.
While I do perceive what you are getting at, I actually percieve his remarks are rather straightforward. He is an airline executive, and in his world aircraft are systems that enable air travel, of which the pilot is but a part (and probably more of a line item than anything else during budget season). Sir Clark, I believe, is commenting from the perspective that the system that was MH370 had been compromised and that it did not functon as intended in delivering passengers and goods from VKL to PEK. He thererfore projects a rather analytical description of how the aircraft was ‘controlled’ in a way that was not intended, diverted from its programmed flight path and flown to a terminus the SIO. It seems to me that his meta-point is that someone hijacked the system and ended the flight in the sea and that this could have been prevented. My point is that he really has not thought through the loss of MH370 beyond how the system was compromised or otherwise engaged the mystery of the thing. I would also repeat that I believe that he has based his conclusion that the aircraft was flown into the sea at the hands of a human after having concluded that it was diverted from its intended flight path at the hands of a human. This is logical, but it is not necessarily valid. Logic can be tricky this way, it is comforting, yet it does not reveal all, no matter its symmetry and grace (a word just for you).
My view, rather, is that aircraft appears to have been diverted intentionally while the termination in the SIO was unintentional. Sir Clark, then, is missing an element in his analytical sequence: it appears that there was some catalyst that transformed the flight post-diversion from one exhibiting human input to one devoid of human input. And it is this catalyst, whatever it was, that I perceive to have been overlooked in most expressed views of the incident.
I think it is Sir Tim, not Sir Clark.
@airlandseaman: just re-sending my requests to you, in case you missed them last week:
1) does a detailed “decompression scenario” model exist? If so, can it be published?
2) how did the IG rationalize this scenario’s slow, trigonometric curl?
3) does the IG now have performance models to test the ATSB’s current claim (per its Oct.8 report) that MH370 lacked the endurance to reach the 7th arc at speeds below 350ktas? If so, can this performance model be made public?
4) do you agree acoustic pings didn’t MOVE the search NE (just KEPT it there)?
Thanks,
Brock
@Bobby: thanks for your response.
Thanks for clarifying that you a) had a logical reason to test a range of speed deltas, and b) (effectively) tested this range under a wide variety of possible paths. I missed both points on first read.
And thanks for the work you did on those two “s21” scenarios. While I didn’t mean for either of them to be modeled as “great circle” routes (both passed through the Inmarsat arcs on cue, and were thus severely curving paths), I won’t trouble you for any further research, as a) I wouldn’t expect the signal data fit to improve materially, and b) my primary goal is to nail down feasibility of those scenarios with respect to ENDURANCE – if you lack the data at speeds that low, there’s nothing you can do.
I appreciate the reference to Lauren’s posted fuel stats; but I’m pretty sure I’ve already traced those values back to an extremely sparse (i.e. 2 minutes’ flying time) set of trials, so I’m not sure I can rely on those, either.
I think the IG has a working performance model, now (they seem to have used it to invalidate your recommendation, after all) I’ll just try to get a copy (or at least the essence) of that.
Brock
@Victor,
Let me address your points one by one:
180 Degree True Track
I fit the 180 degree True Track route you described, and my results appear consistent with yours. I get a constant speed of 484 knots between the end of the turn and 22:41 if you allow the radial arc errors to be 4.9 NM. I get the end point at 93.71 degrees, but you have 25 minutes of unaccounted time before the final southern bearing. This 4.9 NM is equivalent to ~42 microseconds RMS BTO error (i.e., larger than the ATSB estimated at KLIA). If you reduce the radial error limit to 1.5 NM (13 microseconds), the RMS speed variations rise, but only to 2 knots. So you can have a steady TAS on this route if the plane maneuvered for 25 minutes between 18:22 and 19:41 and was also flying south at 18:40 (to match the BFO then).
What happens when you pass the last programmed waypoint?
There are a number of blog posts that address the question of the FMS (or FMC if you prefer) action after the last programmed waypoint is passed. There seems to be a consensus that the aircraft will continue the same course.
Here is one posting:
http://www.reddit.com/r/MH370/comments/2973o2/how_is_autopilot_programmed/
which says:
“If there is an incomplete or inconsistent route programmed into the FMS, the plane keeps on flying the preprogrammed heading once it reaches the end of the programmed route.”
I don’t have a FMC manual from Boeing to verify this.
Higher air speed than LRC?
There are at least two independent methods to determine the TAS after the radar track. The first method is to calculate the speed from the diversion to 18:22. If you simply compute the distance along the ATSB’s radar track over that ~1 hour span and correct for winds aloft, you will get very nearly 500 knots. That is actually a very precise speed measurement, and it is a several % higher than what is typical for LRC, even on this particular night.
The second method is finding the constant speed that best matches the satellite data between 18:22 and 22:41. Again you get nearly 500 knots TAS.
If you read my post yesterday at this site (October 16 at 6:00 p.m.), item #6, you will see that I have compared the average fuel burn rate for my route with Boeing range tables. Even with a 3% allowance for the unreleased engine PDAs, there is still a 4% margin left to accommodate the increased fuel burn rate due to higher speed.
Speed Decline
Regarding the speed decline my route has at the last two handshakes, I can say there is nothing assumed or forced about this. That’s just how it came out. The speed from 00:11 to 00:19 is uncertain because the leg is so short that a few NM difference affects the calculated speed noticeably. If anything, my result was higher than I was expecting if the plane ran out of fuel completely about halfway through the leg. Perhaps it indicates simply that the plane was continuing to glide in the same direction and the horizontal speed did not decline much until the last part of the leg.
The speed between 22:41 and 00:11 was also a bit surprising to me in that it seems to show more of a decline than I was expecting based on the ATSB’s notion that the aircraft lost one engine perhaps only a few minutes before 00:16. If my proposed route is correct, then the average speed on this leg seems to imply that one engine quit sooner than currently suggested by the ATSB. I don’t know whether the aircraft (without pilot action) automatically transfers fuel to keep the two engines running the same length of time, or how accurate this is.
Does anyone know what speed reduction, and perhaps altitude reduction, will occur with loss of one engine?
@Richard Cole,
I will respond to your points:
1. Systematic errors in the AES frequency compensation may depend on aircraft position, course, and speed. BFO analysis of a different flight route than MH370, such as Amsterdam, is better than nothing, but the presence or absence of obvious systematic errors on an Amsterdam route is inconclusive regarding the presence or absence of errors on the MH370 route.
2. I am not saying that routes with low BFO errors are anomalous or should be ruled out, just that the relationship between BFO residuals and the likelihood that the route is correct is not as simple as “lower residuals means more likely.”
3. You are correct in noting that the limits of the ATSB’s search do not depend on BFOs. The southwestern limit only requires two numbers: fuel on board (which gives range, subject to some assumptions) and 00:19 BTO. I believe the ATSB also realized that BFO data did not lead to a unique route. It did serve to limit the turn time if you assume a single turn, and that moved them down the 7th arc.
Richard Cole (richardc10)
Posted October 17, 2014 at 2:17 PM
Hi Richard ~
Thank you so much for bringing this formal styling distinction to our attention. My sincere apology to Sir Tim.
http://en.wikipedia.org/wiki/Sir
Similarly, I have also learned from sources in Malaysia that the pilot of our subject flight MH370 is properly addressed as Captain Zaharie, not Captain Zac or Captain Shah.
~LG~
I am sure Sir Tim can survive. All very 12th century.
If we’re going Medieval, I call dibs on “Sir Rants-a-lot”…
@VictorI,
Thank you for your response.
“I believe there is a path from IGARI towards Port Blair (Andamans) to BEDAX and south that would work. The plane would need to be descending around 18:40 to match the BFO at that time.”
I find that very interesting. Do you remember where the plane would be at 18:40 during the descent in that model path? Would that line up with any airport approach?
” Of course, that would mean rejecting the primary radar data. For this reason, I have not studied this path in detail.”
Rejecting the primary radar data at 18:22 to be that of MH370 was exactly the inspiration for my question.
However, I think, rejecting that very “assumed truth” is actually an acceptance of the primary radar data as showing true and accurate returns of planes.
True and accurate, here, includes the sharp corner in the track just prior to the (assumed?) turn back, or possibly more accurately described as at the point of incorrectly grafting MH370’s track into that corner to the track of some other plane out of that corner.
Now, if we accept the above, the total scenario becomes a simple one. Some incident at the sharp corner (mid air collision, military exercise gone wrong, e.g. missing target drone and hitting MH370, or something of that nature), MH370 incapacitated, pilots turning west for emergency landing in Port Blair, later revising airport to some other and turning, then being overcome and the plane continuing on auto pilot to fuel exhaustion.
I think, using the sharp corner radar data as the initial condition for path modeling is a worthwhile exercise.
Cheers,
Will
@Bobby: You provided the following text: “If there is an incomplete or inconsistent route programmed into the FMS, the plane keeps on flying the preprogrammed heading once it reaches the end of the programmed route.”
As I see it, there are five possibilities for the navigation mode: constant heading or constant track, with a magnetic or true bearing; or a great circle. You have assumed the course continues as a great circle after the last waypoint is passed. This is not consistent with a constant heading, which is stated in the excerpt you shared. In fact, as you know, a great circle course requires that the true track is continuously adjusted (except for the special cases of a course due south or along the equator). I have not seen evidence that the great circle assumption is correct. I have seen statements suggesting either a constant heading or constant track.
You have allowed a flame-out of one of the engines well before the handshake at 00:11 UTC. This would require a higher than expected variation in performance for the two engines. I understand that your speeds “were not forced,” but you are using consistency of the TAS (and I think it really should be Ma number and not TAS) as one of your main criteria for judging the acceptability of a path. One could argue that if you find a surprising TAS for one of the legs, that would reduce the probability of the path being correct.
Your end point seems to exceed the maximum range calculated by a number of us, and you do this at a higher speed than the LRC speed. However, I do agree that there is imprecision in the fuel consumption numbers beyond the unknown PDAs.
Please do not take my comments the wrong way. I am not asserting that your path is incorrect. On the other hand, I don’t adhere to anybody’s belief that there is an absolutely correct path, be it yours, the Inmarsat’s, the IG’s, or mine. Every path that I have seen has its pros and cons as determined by the assumptions and constraints of the solution. At best, we can subjectively rank the likelihood of the various proposed paths, and there would be disagreements about those rankings.
One of the ironies is that while a number of us critically assess each others’ predictions for the path into the SIO, the general sentiment of the public is to group all the paths together and say we are all wrong because the lack of debris would indicate the plane did not crash in the SIO. Although I do not adhere to this belief, I cannot say with certainty that it is wrong. There is so few data in this episode and so much of the data is conflicting that little is known for certain.
To All,
Is it possible reset the nominal position of the satellite for the delta f comp ?
Victor – It’s an interesting disconnect isn’t it? The public have more or less moved on while the data contortions continue on. From the moment Chris – that plane went south – McLaughlin stamped his authority a veneer of certainty has been applied to the fate of MH370. Out among the punters though, no plane, no wreckage, no emergency beacons, no distress signal = 7 hours funny shit, and they are cynical enough to join a few dots and conclude that game is somewhere else.
LGHamilton / MuOne,
I take it MuOne has accepted the proposal as he is acknowledging an engagement!!!!! Yippeee!
I wish something beautiful could emerge from this tragedy and a wedding would be perfect! Don’t forget I am dancing at this wedding guys.
All The Best. Hope it really comes to be.
Nothing new to post, just following along on all the great technical expertise.
@VictorI: how do the IG fuel models which rule Bobby’s path out (for range) rule on Mike’s “decompression scenario” (for endurance)?
@Victori
>> As for other paths that ignore the radar
>> data, I will allow others to take a lead
>> in those calculations. Until I have
>> evidence to the contrary, I will accept
>> the radar data as valid, as both Thailand
>> and Malaysia reported to seeing the plane.
Not clear how the radar data is relevant to the question of whether the eyewitnesses, Malaysian police and US intelligence are indeed correct that the plane barreled in low during the overflight between Kota Bharu and Penang airports.
@Luigi: Yes, you are right. I mixed my responses to you and MuOne.
A lower altitude would be inconsistent with the 490-500 knots required for the radar track. Also, due to reflections and other atmospheric effects, I am not sure that it is impossible to have a very short-lived cell phone connection at 35,000 ft, even if a longer connection is impossible.
@Victori
>> A lower altitude would be inconsistent
>> with the 490-500 knots required for the
>> radar track.
Is that actually true, though? I recall it being disputed here the last time I raised this point, and we are talking about only a minor fraction of the trajectory. I haven’t been able to find any hard data on the dependence of max 777 speed on altitude, but I did find this statement which might be pertinent:
“At the most low altitudes, the Vmo airspeed restriction is usually determined by the windshield’s ability to handle the certification bird strike criteria. In many parts of the world, the ATC airspeed restrictions allow an IAS above 250 kts, so many transport jets are capable of a TAS as high as 400+ kts at very low altitudes, should they choose to operate at that speed.”
Source:
http://www.askacfi.com/3789/altitude-and-speed.htm
Bobby – 1. Systematic errors in the AES frequency compensation may depend on aircraft position, course, and speed. BFO analysis of a different flight route than MH370, such as Amsterdam, is better than nothing, but the presence or absence of obvious systematic errors on an Amsterdam route is inconclusive regarding the presence or absence of errors on the MH370 route.
You are right, but you are applying scientific rigour when the forensic exploration of satellite data has by now had to accommodate assumption at pivotal junctures – this being another one. There was no other way the sleuthing could progress. I look at it from Murphy’s law which means it has very likely gone off the rails by now. The Amsterdan flight didn’t have people in it playing ball with the electronics. If there is nothing in the boxes then it’s all wrong and we may never know how wrong – IF.
New York Times
September 14, 2001
”The fact of the matter is that cell phones can work in almost all phases of a commercial flight,” said Marvin Sirbu, professor of engineering and public policy at Carnegie Mellon University. ”An excess of caution prevents us from doing so, of course, because we are so worried about the safety of air travel.”
https://twitter.com/nihonmama/status/455278857902252032
@Victor,
I see your issue with what I believe is imprecise language (just as when the ATSB used “heading” in the table when they should have used “track”).
The gist of all the comments I have read is that the FMC is normally used only for great circle navigation (please provide a reference is this is incorrect). I know the MCP can be used to set the other four LNAV modes.
My assumption in modeling routes is that whatever LNAV mode is used coming out of the turn, it remains in that mode until the end of the flight. So for great circle mode, I assume once a great circle, always a great circle. For starting a True Heading, always True Heading. etc. In my modeling, I never allow LNAV modes to change on the southern route, nor have I seen any evidence that the aircraft does this on its own, nor have I seen any other route modeler do this.
I have not “allowed a flame-out of one of the engines” at any particular time. I’m not even sure how to do that.
Since the PDAs have not been released, we can’t estimate right now the difference in engine run times. I am not an aircraft performance analyst, so what may be surprising to me could be exactly as expected to them. That’s why I raised the question of speed/altitude changes when on a single engine.
By the way, my last Addendum shows that it does not matter whether you assume constant TAS or constant Mach number, the end point is virtually identical.
Unlike you, Victor, I do believe there exists one “absolutely correct path” – the one actually flown by MH370. My goal is to approximate it as closely as possible by modeling routes using the satellite data. I also believe that 9M-MRO will be found, and the combination of its location and the FDR data will allow determination of its “absolutely correct path.” It’s only a matter of time before we know it.
@nihonmama
So, what is the probability that the co-pilot’s cellphone reconnected at 35000 ft in the 16-minute hop from Kota Bharu to Penang?
@nihonmama
An interesting early (March 17) NYT article on the cellphone question, before the cell-tower reconnect story emerged:
http://www.nytimes.com/interactive/2014/03/14/world/asia/flight-mh370-map-update.html
This article also features a map of the plane’s flight path which is quite interesting and matches our current understanding. The map is annotated to say that the plane “descended unevenly to 23000 ft” before crossing the Malay Peninsula. The basis for this statement is not specified within the article — presumably it was not the eyewitnesses at Kota Bharu or the cell-tower reconnect, but some third source. The indicated altitude of 23000 ft is midway between a normal cruise and the more extreme values of 5000-10000 ft mentioned in some news reports.
@nihonmama
Sorry, that was the map URL. The article:
http://www.nytimes.com/2014/03/18/world/asia/questions-over-absence-of-cellphone-calls-from-missing-passengers.html?_r=0
@Brock,
@IG,
It is difficult to comment on someone’s analysis when they have not made it public knowledge. I’m referring to the IG’s fuel analysis. I have not seen any calculations presented which attempt to evaluate the feasibility of my proposed route. Did I miss this? If so, where are the numbers? The analysis presented by Barry Martin did not assess my route conditions. If the IG has done something beyond that, I would like to see the detailed calculations. As you well know, getting an answer out of a model with some assumptions is easy. Having a model that is validated to an accuracy better than 1-2% is quite a challenge even for Boeing.
The October 2nd update from the IG provides no details of their fuel model or of its assumptions, but says my end point “may be problematic in terms of fuel availability.” Perhaps this is what you are referring to, Brock. In any case, if the IG sends you their fuel model, would you be so kind as to make it available to everyone? Thanks.
In the comments section of an old Daily Mail story was this one as I have no idea of what’s relevant I am passing it on for someone to check .
“Jaxx, Langkawi, Malaysia, 6 months ago
My friend watched a large plane flying very low at 2.20am on the evening of the disappearance. It had no lights on apart from the undercarriage which was brightly lit. This was at Gungung Raya , Langkawi, Malaysia. Gungung Raya is a large mountain on the island. She thought at first it was going to crash but maintained its low flight through the mountains. This is an area of dense rainforest. The Thai plane sited on radar was at 2.13am at Butterworth around 10 mins flyiing time from here. My friend has lodged a report with the police”
Read more: http://www.dailymail.co.uk/news/article-2583807/Did-Malaysian-fisherman-missing-flight-MH370-flying-low-Gulf-Thailand.html#ixzz3GTJfd8w2
He’s back…the nerve of the guy.
http://www.watoday.com.au/wa-news/malaysian-defence-minister-to-inspect-mh370-search-vessel-in-fremantle-20141018-117v22.html
@All
Can someone provide a link to an altitude-speed flight envelope plot for the 777? I have been unable to find one. So far, what material I have found make me skeptical that inferred speed considerations should overrule multiple lines of evidence regarding a descent on the return to Malaysia. Since a plane will burn less fuel at cruising altitude, that suggests that the fuel burn calculations used in the various models people have been touting are significantly off. Maybe that’s why they haven’t found the plane.
I did learn that, unlike on the rival Airbus, the Boeing 777 has a neat feature which should appeal to enthusiast pilots. If you press hard enough on the stick on a 777, “flight envelope protection” disengages. You can then explore the full aerodynamic and structural limits of the aircraft, just as if you were at home in your den fooling around on your personal simulator.
http://en.wikipedia.org/wiki/Flight_envelope_protection
Interesting.
@Bobby: yes, I was referring to the quote you gave, as well as Victor’s yesterday:
“Your end point seems to exceed the maximum range calculated by a number of us”
…which implies the presence of rigourous fuel models.
My requests for their publication have been made to Mike (Oct.10), the IG generally (Oct.14), and Victor (yesterday). These are busy people, volunteering their time – and they owe neither me nor anyone else in the peanut gallery’s peanut gallery anything at all – so I would be very wrong to complain. But I sure am grateful to have you join me in this quest, Bobby: your credentials and dedication are outstanding, and dismissals of your PUBLISHED fuel model & analysis should, in my opinion, themselves reference a PUBLISHED fuel model & analysis.
@Luigi Warren
Thanks for sharing that. The option on 777’s to disengage the “flight protection envelope”, is an interesting point, and supports my emergency maneuver scenario.
Maximum speed for 777’s is listed at 590 mph. My nm calculation from Igari and over the Malaysian Peninsula indicates that airspeed was pushing 625 mph and speeds may have been even higher at some points, due to the onboard emergency.
Regarding the photo of the radar trace shown to the families, and this could be way off base, but I thought because the two listed distances were from Butterworth AFB, this trace was from “Butterworth” Radar. I put this in quotes because Duncan Steel’s post of radar (http://www.duncansteel.com/archives/930) showed a radar installation called Western Hill that is about 10nm from Butterworth but did not show a radar site exactly at Butterworth. If “Western Hill” radar and “Butterworth” radar are one and the same, using graphics from the referenced post, its range extends to MH370’s approximate location at 18:22. Note, this does not explain the missing contact circle shown on the photo.
JS – Good point about steering the antenna. My impression is there is general acceptance of second engine flameout around 00:15 with the RAT kicking in around 00:19. Others have posted the RAT supplies limited electric power i.e., only critical systems are repowered. I wonder if the antenna steering system and the inertial reference system are included. (I believe the inertial reference system is used to furnish speed and direction data to change the plane to satellite frequency bias.)
Dr. Ulich – Did your distance calculations include the squiggle in the flight path after 17:21? The radar trace in the ATSB report shows an easterly turn after IGARI towards BITOD before the left turns towards Penang. (The initial published sequence shows HCM-ATC reported losing radar contact when the MH370 was directly over BITOD.)
Brock & Luigi – In order for you to determine the burn rates for different speeds and altitudes, you should go to the table in my post on page 7 of the “What We Know” thread. That table came from a post on PPRuNE and was said to come from a B777-200ER manual. It provides the amount of fuel burned over the corresponding time period for three distances (600nm, 1200nm, and 2000nm) at four different elevations. (The fuel burn rates from the Delta manual were averaged from data taken over just a few minutes and therefore less precise that the values in the table.) You would then have to interpolate and extrapolate to get an approximate burn rate for the altitude and speed you wish to evaluate. Amendment #3 to Dr. Ulich’s white paper and the recent post go into detail on how the fuel burn rates that support the projected impact area were calculated.
I graphed the burn rates and got four nearly parallel lines, one for each elevation. The graph showed that as the elevation increased, the burn rate (in kg/hr) decreased while speed increased. My quick calculations (not nearly as precise as Dr. Ulich’s) indicate the a/c needed to be flying slightly above FL350 to achieve the required 6.13mt/hr average burn rate.
Nihonmama – The Delta manual shows ceiling values of 42,000ft and 43,100ft for a B777-200ER with Rolls Royce engines. I did not immediately see an explanation for that difference. However, I found posts in PPRuNE saying the ceiling is limited be the total weight. This means that just because the a/c can fly at FL420 does not automatically mean it can reach that altitude when heavily loaded. Some pilots stated that as fuel is burned over long flights, they might ask to increase the flight level 2,000ft every 2-3 hours.
@Lauren,
In response to your question “Did your distance calculations include the squiggle in the flight path after 17:21?”:
Yes.
If you look back at Table 5-1 in my original white paper of September 25th, you will see on lines 5 & 6 that the bearing has increased from 25 degrees (towards IGARI) to 49 degrees while turning east and then to 58 degrees (called “Begin Diversion Turn-Around). It appears that 9M-MRO turned from IGARI toward BITOD, but the diversion began (at 17:22) before reaching BITOD. My radar track analysis also includes the 180 degree turn. So everything you see in the ATSB’s Figure 2 is included in my distance calculations.
@Luigi Warren:
The speed limits of the B777-200 are VMO=330kIAS and MMO=0.87. These are maximum operating speeds that must not be intentionally exceeded. The cross-over altitude is 30477 ft. The AP will probably not accept selected speeds above those limits.
Unlike the Airbus’ hard protections, the envelope protections in the B777 are ‘soft’, they can be overridden by the pilot. The B777 flight envelope protections do not disengage, but manifest themselves by demanding an increasing stickforce to override them.
@Lauren H
>> Amendment #3 to Dr. Ulich’s white paper and the recent post go into detail on how the
>> fuel burn rates that support the projected impact area were calculated.
>> I graphed the burn rates and got four nearly parallel lines, one for each elevation. The
>> graph showed that as the elevation increased, the burn rate (in kg/hr) decreased while
>> speed increased. My quick calculations (not nearly as precise as Dr. Ulich’s) indicate
>> the a/c needed to be flying slightly above FL350 to achieve the required 6.13mt/hr
>> average burn rate.
Except that the burn rate is not a “known,” and (reminder!) the plane has not been found.
Nor is anyone suggesting that the plane was flying at low altitude for the vast majority of the ~8-hour flight. Rather, multiple lines of evidence argue that the plane came in low on its return to Malaysia, that is all. Per Ulich’s table, the plane took about 16 minutes to traverse the Malay peninsula from Kota Bharu (where eyewitnesses filed police reports on a low flying airliner heading west at 1:30 AM) to Penang (reported location of the celltower reconnect). Indeed, that is a somewhat intuitive sequence of events on several different models (political protest by the pilot, abortive kamikaze attack, catastrophic systems failure/fire/decompression). The person controlling the plane obviously had business to conduct in Malaysia — if anything, it would be counter-intuitive for him to return to the homeland and simply cruise around at 35K before zigzagging off to the tip of Sumatra and then down to the SIO. The actual fuel burn adjustment that would be required to accommodate a descent at high speed I cannot calculate. I doubt it would be that large on a percentage basis but, in terms of identifying a focused high-priority search area on any model of subsequent events, it is probably enough to be significant.
Basically, it looks to me like the model tweakers don’t want to get into the weeds on the evidence, so they have just plugged in a normal cruise altitude for the whole flight and said “good enough.” Only, so far, it is obviously not proving “good enough.”
@Luigi Warren
@Gysbreght
Thanks for the info about pilot capability of over-riding the “Flight Envelope Protection” system for exceeding maximum speed on 777’s.
The maximum speed listed for 777’s is 590 mph. However, my nm calculation of the speed between Igari and the west coast of the Malay. Penn., indicate it was pushing 625 mph and at some points may have been even higher.
This supports my rapid descent with an on-board emergency scenario, and the path 370 flew over the Penn.