In last month’s New York magazine article about Zaharie Ahmad Shah’s flight simulator, I cautioned against treating the recovered data as a smoking gun:
…it’s not entirely clear that the recovered flight-simulator data is conclusive. The differences between the simulated and actual flights are significant, most notably in the final direction in which they were heading. It’s possible that their overall similarities are coincidental — that Zaharie didn’t intend his simulator flight as a practice run but had merely decided to fly someplace unusual.
What I failed to question was the report’s assumption that the six points all belonged to a single flight path. On closer examination that assumption seems ill supported. Rather, it seems more likely that the six points were recorded in the course of two or possibly three separate flights. They were interpreted as comprising a single flight only because together they resembled what investigators were hoping to find.
The first four points do appear to show a snapshots from a continuous flight, one that takes off from Kuala Lumpur and climbing as it heads to the northwest. Between each point the fuel remaining decreases by a plausible amount. Each point is separated from the next by a distance of 70 to 360 nautical miles. At the fourth point, the plane is at cruise speed and altitude, heading southwest in a turn to the left. Its direction of flight is toward southern India.
The fifth and sixth points do not fit into the pattern of the first four. For one thing, they are located more than 3,000 miles away to the southeast. This is six or seven hours’ flying time. Curiously, at both points the fuel tanks are empty. Based on its fuel load during the first four points, the plane could have flown for 10 hours or more from the fourth point before running out of fuel.
The fifth and sixth points are close together—just 3.6 nautical miles apart—but so radically different in altitude that it is questionable whether they were generated by the same flight. To go directly from one to the other would require a dive so steep that it would risk tearing the aircraft apart.
The picture becomes even more curious when we examine the plane’s vertical speed at these two points: in each case, it is climbing, despite having no engine power.
The ATSB has speculated that in real life MH370 ran out of fuel shortly before 0:19 on March 8, and thereafter entered into a series of uncontrolled porpoising dives-and-climbs called phugoids. In essence, a plane that is not held steady by a pilot or autopilot, its nose might dip, causing it to speed up. The added speed willl cause the nose to rise, and the plane to climb, which will bleed off speed; as the plane slows, its nose will fall, and the cycle will continue.
Could a phugoid cause a plane to climb—663 feet per minute at point 5, and 2029 feet per minute at point 6? The answer seems to be yes for the fifth point and no for the sixth. Reader Gysbreght conducted an analysis of 777 flight-simulator data published by Mike Exner, in which an airliner was allowed to descend out of control from cruise altitude in the manner that the ATSB believes MH370 did.
A diagram produced by Gysbreght is shown at top. The pink line shows the plane’s altitude, starting at 35,000 feet; the blue line shows its rate of climb. Worth noting is the fact that the phugoid oscillation does indeed cause the plane to exhibit a small positive rate of climb soon at first. But by the time the plane reaches 4000 feet — the altitude of the sixth point — the oscillation has effectively ceased and the plane is in a very steep dive.
Gysbreght concludes:
As expected for a phugoid, the average rate of descent is about 2500 fpm, and it oscillates around that value by +/- 2500 fpm initially. The phugoid is apparently dampened and the amplitude reduces rapidly. I was slightly surprised that it reaches positive climb values at all. Therefore I think that 2000 fpm climb is not the result of phugoid motion.
Not only is the plane climbing briskly at the sixth point, but it is doing so at a very low airspeed—just above stall speed, in fact. If the pilot were flying level at this speed without engine power and pulled back on the controls, he would not climb at 2000 feet per minute; he would stall and plummet. In order to generate these values, the plane must have been put into a dive to gain speed, then pulled up into a vigorous “zoom climb.” Within seconds after point six, the simulated flight’s speed would have bled off to below stall speed and entered into an uncontrollable plunge.
Perhaps this is why Zaharie chose to record this particular point: it would have been an interesting challenge to try to recover from such a plunge at low altitude.
What he was doing at points 5 and 6, evidently, was testing the 777 flight envelope. This might seem like a reckless practice, but I think the opposite is the case. From time to time, airline pilots do find themselves in unexpected and dangerous conditions. For instance, as Gysbreght has noted, “On 7 october 2008 VH-QPA, an A330-303, operating flight QF72 from Singapore to Perth, experienced an In-flight Upset west of Learmonth, West Australia. The upset was caused by a freak combination of an instrumentation failure and an error in the flight control software, which resulted in an uncommanded pitch-down. The vertical acceleration changed in 1.8 seconds from +1 g to -0.8 g.” It would be better to experience a situation like this for the first time in a flight simulator in one’s basement, rather than in midair with a load of passengers and crew.
What Zaharie clearly was not trying to do was to fly to McMurdo Station in Antarctica, as some have speculated.
For one thing, while a 777 is fully capable of flying from Kuala Lumpur to Antarctica, it was not carrying enough at point 1 to make the trip. And if one were trying to reach a distant location, one would not do so by running one’s tanks dry and then performing unpowered zoom climbs.
The misinterpretation of the flight simulator data offers a couple of cautionary lessons. The first is that we have to be careful not to let a favored theory color our interpretation of the data. The investigators believed that MH370 flew up the Malacca Strait and wound up in the southern Indian Ocean, and they believed that Zaharie was most likely the culprit; therefore, when they found data points on his hard drive that could be lumped together to form such a route, that’s what they perceived.
A second lesson is that we cannot uncritically accept the analysis made by officials or by self-described experts. Science operates on openness. If someone offers an analysis, but refuses to share the underlying data, we should instinctively view their claims with suspicion.
@Richard Cole,
You said: “Without all the validation BFO data, external observers are speculating if they use small errors.”
I would take the opposite point of view: Anyone who does not use small errors is speculating.
Hers’s why I think so.
The Inmarsat paper makes several points. One is that BFOs from SEVERAL AIRCRAFT flying at the same time as MH370 (see their Figure 15 for one example) show small errors (within +/- 7 Hz peak error). Presumably they wanted to make the point that the non-aircraft errors (i.e., the ground system and satellite errors) were accurately characterized and compensated, so they will not significantly degrade the MH370 BFO accuracy. Using several simultaneous flights was a good means of demonstrating this.
The second point Inmarsat made was the statement “Validation was performed against several aircraft that were inflight at the same time as MH370, and for the MH370 aircraft in the days leading up to the accident, and good agreement between predicted and measured BFO was seen. . . . This suggests that ±7Hz is a conservative estimate of the typical accuracy BFO calculation achieves, . . . .”
Thus Inmarsat checked BFOs using 9M-MRO data for the days (plural) leading up to March 7th. In other words, they looked at multiple prior 9M-MRO flights. The clear meaning of their statements is that all the simultaneous and prior 9M-MRO flights were within the “conservative” +/- 7 Hz boundaries. It’s too bad they did not show the 9M-MRO BFOs from flights on previous days, but even if they did, we would still have the same conclusion, which is that the maximum expected BFO error is 7 Hz (not 10 or 20 or 40).
A third observation is that the Example Flight Path Inmarsat showed in Figure 16 has all of its BFO errors within +/- 7 Hz. Clearly Inmarsat believed in this upper bound on BFO error so strongly they would not allow the BFO error to exceed that threshold, even if it meant they had had to vary the track from 186 degrees to 180 degrees in order to do so (and with no explanation for the curving track).
Another point is this: Inmarsat is the only demonstrated “expert” with regard to BFO errors. Certainly the ATSB has no expertise in this area. Based on what is published in their book and on their limited (nonexistent?) previous knowlege of this satellite system, I don’t attach as much credibility to the DSTG’s BFO error analysis as I do to Inmarsat’s. The DSTG adopted a 7 Hz model standard deviation “to be conservative and allow for potential variation in the . . . bias value on the
accident flight . . .”. Apparently Inmarsat saw no bias variation during individual flights. DSTG showed one example of this (their Fig. 5.4). I wonder if the “kink” in Figure 5.4 could have been caused by a satellite eclipse?
Finally, the DSTG data show in Table 5.1 a actual 1-sigma near 4 Hz (Inmarsat does not quote a number, but it would be near 2-3 Hz). Any proposed route that has an 8 Hz RMS BFO error is clearly inconsistent with both the DSTG and Inmarsat analyses.
n.b. Nicobar islands are part of the Andaman islands group.
@DennisW
The whole thing stinks to high heaven because these are not waypoints entered into a flight plan that was deleted. These are snapshots in time of a flight (or flights) taken in real-time. The inclusion of fuel load, altitude, VSI, heading, turn rate, etc, means that certain moments within a flight (or flights) were selected and dumped to a drive as a snapshot output. Why? In what context? Through what mechanism? For what purpose?
For the life of me, I cannot understand why they’re not dated and timestamped and thus, by definition, capable of ending this discussion in a heartbeat.
The SIO coordinates, of course, are the most damning. But what is their context for Pete’s sake???
Just another aspect of the unending cluster**** that is MH370.
@MattM –
Maybe ZS was exporting his Flight Sim database to move to another installation before wiping the drives in preparation to reinstall the Windows OS and an new installation of a flight simulator instance. What we are seeing is the files recovered after being deleted from their post migration staging drives.
@Matt Moriarty
They were planted. Someone wants another seabed search.
If an Indonesian official states; ‘ANOTHER military radar suggestion detected it (MH370) IN the Andaman islands’, he also states there are more ‘military radar suggestions’ that detected it (MH370).
I think this would be very significant information.
Specificcaly about where the FMT could have taken place.
Why don’t they free this data??
@DennisW
>The use of term “one sigma” is highly misleading when referring to oscillator drift.
Indeed. My point was that the detail of the BFO error was not much discussed in the early papers. The DTSG paper doesn’t give much more detail, outside the sample in figure 5.4 and generally negative comments on its usefulness at the detailed, few Hz, level.
> “made it difficult for the official search to come to a conclusion”. Are you referring to humans i.e. SSWG?
Sorry, using the term ‘search’ to mean more than one thing. Yes, I as referring to the decision making process (however formed) taking a long time to decide whether the data optimisation models were useful, against the BFO errors. A ‘working group’ doesn’t sound like a decision making body, so I would imagine there is some board to which the SSWG reports. There were some press comments in late 2014 (which I can’t track down now) on unhappiness within the search organisation when the Autopilot Mode and Data Optimisation model solutions began to creep apart.
@Ge Rijn: “Indonesion radar range is just reaching the Nicobar islands. If thy saw it (MH370) in the Andaman islands they must have seen it at the time it flew out of their radar range.” And your following post.
I got the same impression. One wonders about these radars. I would guess there are or could be contractual agreements (geographically negotiated?) under international law stating how far radars are allowed to reach in respect to borders on land and in water. The flip side aspect being what limits them in time, topography and technology. “We are on guard 24/7 and we can’t see shit outside our territorial waters”. It sounds like one wants to turn that around. Depending a little on which country, I could guess that some would still communicate what actually was recorded, but through unofficilal means, or a third party, not to compromise International or bilateral contracts. In the case of Indonesia? If the military at the spot are not running their own show, I would lean towards that they have sent all info along a long time ago. But I am no expert on Indonesian-Malaysian relations.
@Richard Cole, DrBobby, Gysbrecht, VictorI, Dennis W et al. Wondering about the larger-than-noise one-sided BFO error with “geographic dependency” described by DSTG. Could it be that the automatic doppler correction on the a/c is working out its vector component velocities using spherical earth rather than WGS84? Or that the sat ephemeris “description” that is needed by the AES software is simplified to ignore eccentricity and simplifies Sat oscillation using inclination only? Or both (it would certainly simplify the calcs required if it was all spherical earth, mean volumetric earth radius, with no adjustment for satellite altitude change… )
@DennisW, Ge Rijn
It was one of the oldest bits of news that MH370 deliberately flew around Indonesian airspace.
Indonesia made it clear more than once that MH370 did not fly over Indonesian territory/in Indonesian airspace, but they were not very clear on whether they did detect MH370 outside of it.
http://www.antaranews.com/en/news/93270/indonesian-military-radar-did-not-detect-missing-airplane
Perhaps someone saw a radar blip outside of Indonesian airspace but didn’t therefore pay much attention to it? Is this regularly recorded or are we just relying on someone’s memory of that day? Indonesia did not provide radar data.
If the BFO value at 18:40 can be explained with a descent at that time, then the eventual FMT may well have gone unnoticed by Indonesian radar, for example, if it was flying close to waypoint BEDAX. That flight route would make sense if news reports are correct that MH370 skirted Indonesian airspace. Perhaps whoever was in control was aware that India did not regularly monitor the Andaman region. A possible route is reflected in the first ATSB model, from June 2014, terminating in the Broken Ridge area (from where debris is unlikely to turn up in Australia).
Here’s a map of the FIRs:
http://heresthenews.blogspot.co.uk/2014/03/mh370-flight-path-as-tracked-and-likely.html
Note that VAMPI is just outside the Indonesian FIR.
@Brock McEwen,
Thanks for your interest and probing questions from your 9/4 2:13 p.m. post. I will answer them below.
Your #1 description is not quite right. There is no “coincidence” as you put it. To explain what I meant, here is a more precise description of what I have found with some additional details added.
1. The best Great Circle route is at MRC (which is essentially constant Mach but not constant TAS because of temperature variation), but it is short on fuel and the BFO errors are too large at 4 Hz RMS. A Great Circle route at Long-Range Cruise has the same issues, only worse. So Great Circles are out (for these and several other reasons).
2. The True Track route at MRC also has the same issues: short of fuel and RMS BFOs of 4 Hz.
3. The True Track route at Holding LRC has slightly better BFOs of 3 Hz RMS (which I consider to be marginally acceptable), but it is also short on fuel (PDA = -1.9%). I hope my fuel model is within +/- 1.5%, but I don’t have any means at the moment to test it. A set of fuel readings during a previous flight would be welcomed (I asked ATSB, but they would not provide these).
4. There is a Magnetic Track route at Holding LRC that has RMS BFOs of 1.8 Hz (in line with Inmarsat) and a PDA of +4.0% (i.e., enough fuel even for well-used engines). It ends at roughly 33.4S, 94.9E.
5. There is a True Heading route at Holding LRC with 2.1 Hz RMS BFOs (in line with Inmarsat) and PDA = +4.7%.
6. I have read on a pilot’s blog that PDAs for used engines should be approximately 1.5 – 3.5%, but I have been unable to find anything official. If I adopt a 1.5% error bar on my PDA estimates, then any predicted PDA from 0.0% to 5.0% from my fuel model could be considered consistent with expectations. I would be more likely to believe a result near the high end than the low end simply because I suspect these engines had a fair number of operating hours. If anyone can shed light on this subject I would be grateful. I would say I think the +4.7% from the True Heading route is acceptable as being consistent with expectations and fuel model errors.
Within the limits of PDA acceptability, there is actually room to accommodate a significant descent near FMT (even down to FL100) and a climb back to FL360, which costs roughly 2% in fuel and therefore reduces PDA by 2% (from 4.7% to ~2.7%), if that would make you happier.
We don’t know what the PDA’s were in the early stages of the flight, and even if we did it would not tell us much of use. From the fuel data, we only know the DIFFERENCES in the PDAs between the two engines, and mostly in the climb phase. There is insufficient fuel data in the FI to determine a cruise PDA. We can see the difference in fuel flows between the two engines varies during the climb and transition to cruise. Several of us have estimated the difference in cruise PDAs to be 0.6% to 0.8%, with the right engine having a higher PDA. Unfortunately, we don’t have a reliable means of estimated the average PDA (which is certainly well-known to MAS, Rolls Royce, Boeing, and probably ATSB, but not by us). Bummer.
I am using a single set of interpolated wind data at 2100 UTC (exactly the same as your link). For the positions in my True Heading route, the winds are: 18:54 (11 kts from 89 deg), 19:41 (24 kts from 87 deg), 20:41 (13 kts from 86 deg), 22:41 (15 kts from 69 deg), 22:41 (27 kts from 274 deg), and 00:11 (45 kts from 261 deg). Currently I simply average the winds at each end of a flight leg to estimate the average wind effect. One might improve the bearing accuracy somewhat by integrating the wind along the actual route between the handshake times. Still, what I have now should be quite capable of finding acceptable routes if they exist. The terminus prediction accuracy could also be improved with a more refined wind model.
The True Heading route actually does not curve slightly east between 18:54 (end of FMT at 5.8N) to 21:41 (at 15.1S). The true bearings are 182.18 deg from FMT to 19:41, 183.43 from 19:41 to 20:41, and 182.18 deg from 20:41 to 21:41. All these legs end westward of their starting points (due to the initial westward heading plus the west-bound winds). As the wind varies along this portion of the route, the westward progress will vary also. In this part of the route, the average wind was strongest for the leg ending at 20:41; thus its westward progress was the largest, and its true bearing was also the largest.
All the information needed to create an accurate fuel model for cruise conditions is available in the Boeing flight performance manuals for B777-200ER with the Trent and GE engines plus the FI. It sounds simple, and it is in theory. Just take the remaining fuel at 17:07 from the FI, create a recurring model in Excel that estimates the fuel used every time increment (I use 1 minute) based on the altitude, temperature (you have to look that up based on location), air speed at that temperature (oh, by the way you have to convert the 250 hPa temperature to your altitude, too), and aircraft weight, and run that model until the fuel is exhausted. First you have to figure out how the accurately interpolate (and parameterize) the Boeing fuel flow tables. Then you have to correct the Holding FBR numbers because there is a footnote saying that 5% is added to account for racetrack turns (that we don’t need after FMT). Then you have to compensate for the effect of temperature (which is not mentioned for the Trent tables but appears in the GE tables as a footnote based on TAT which you have to convert to SAT, and you have to know what the ISA SAT is too). Then you have to account for the fuel used/saved in climbs and descents. Finally you get the pleasure of figuring single engine INOP fuel flows until both engines reach FE. Nothing to it. Only took me about 18 months to do after learning more than I ever wanted to know about Excel macros, circular references, etc. If you look at the graphs in my Figure 2 at the lower left one (I know they are small but I make up for that by having 11 of them!), it shows the fuel flow, also called fuel burn rate, versus time. You see several effects. First, the positive blip at 17:24 is the post-Diversion climb from FL350 (flying eastward) to FL360 (flying westward). Then the fuel flow slowly decreases (as the aircraft gets lighter) until 18:29 when the speed is reduced from LRC to Holding LRC. Then both the speed and fuel flow drop. The fuel flow continues to decline slowly as fuel is consumed and the aircraft gets lighter. At ~00:14 the right engine flames out and the fuel flow changes to engine inoperative holding, which is only slightly less than 2-engine Holding LRC. Finally at 00:17:29 the left engine exhausts the remaining fuel and the FBR drops to zero. Now scale the fuel flows over the whole route using a PDA parameter so second-engine FE occurs at exactly 00:17:29. Piece of cake. My fuel model is a single Excel worksheet, but it has links to/from about a dozen others, so it can’t be separated. Too bad it’s not a callable subroutine. You can roughly check the numbers, though, using the Boeing FBR tables with a weight at 18:30 of 208 MT.
PS
Here’s the article citing a Malaysian official who indicates that MH370 did try and avoid Indonesian airspace:
http://www.themalaymailonline.com/malaysia/article/untrue-that-mh370-avoided-indonesian-radar-hishammuddin-says
“The plane was then deliberately flown around northern Indonesia instead of flying over the country before it headed in a southerly direction towards the Indian Ocean where a multi-nation hunt is currently looking for it, Robertson said.”
It seems to confirm that MH370 did not fly in Indonesian airspace but that there is some information available to confirm that it did not do so other than that it wasn’t detected at all.
How is “airspace” defined? For example, Wikipedia says:
Dennis, Nederland,
May I insert my 2 cents? Back in February 2016 we discussed one interesting document released by Malaysian team. Citation from Jeff’s comment on February 5, 2016, 8:21 am:
—–
And interesting excerpt:
“CONSTRAINTS & CHALLENGES
• Sensitive Information from the military. Release of this information requires proper protocol and approval
• Radar Information from neighbouring countries is hard to come by due to sensitive nature of such information”
—–
The original document now requires password to access. Perhaps Jeff or Susie saved it:
http://www.icao.int/APAC/Meetings/2015%20AIG%20Workshop/02%20MAL%20-%20MH370%20INVESTIGATION%20TEAMS%20PRESENTATION.PDF
Anyhow, the point was that the investigators tried to obtain radar information from neighbour countries. The formulation clearly was made in the plural form. This is a kind of indication that both Thai and Indonesian military tracked MH370.
Johan,
“If you scroll down on wikipedia’s 777 page you will find maximum landing weight specs for the triple-7 variants…”
My point was that a possible mechanical failure in the area of EE-Bay could involve a possible damage of nose landing gear, which is located just under the EE-Bay. Thus the crew could anticipate the necessity or possibility of belly-landing. Hence it was too risky to attempt to land with > 30 tons of kerosene, during night and without immediate support of emergency services on the ground. The problem is not in the landing weight, but in the huge volume of flammable kerosene.
@Nederland. Although the media headlines at the time would have you believe otherwise, when you look at the statements quoted they DO NOT state that the a/c did not fly through Indo airspace. All they say is that THEY DID NOT SEE A PLANE WHERE IT WAS SUGGESTED THEY SHOULD HAVE by the investigators (NW end of Sumatra or points just beyond) and they also added that they WOULD HAVE SEEN IT if it had gone that way. Taken at face value, these statement would rule out any possibility of an FMT (on to south or marginally east-of-south heading) before reaching W of ~91E. They do not rule out the possibility that the Indonesians saw what might have been the a/c – but somewhere not predicted by investigators. In fact I believe that there is circumstantial evidence that some interested parties (who tasked Terra SAR-X satellite imagery over Sumatra and points immediately south of mid-island) did have reason to believe that the plane had gone that way – and not NW up the Malacca strait as the “authorised version” would have us believe.
Nice work, Dr Bobby
Indonesian airspace probably corresponds to Jakarta FIR:
http://ifonlysingaporeans.blogspot.co.uk/2016/03/singapore-flight-information-region.html
Pulau Perak island in the Malacca Strait belongs to Malaysia, but has no radar.
MH370’s route as known by radar was within Kuala Lumpur FIR. If it was later flying in Chennai FIR (Andaman region), as indicated above, the FMT must have occurred after 18:40 and it may well have avoided entering Jakarta FIR. This may appear more logical than flying over, or close to, Sumatra, given that MH370 was first flying in a northwesterly direction, although it appears its final destination was in a southerly direction. Probably there was some reason for it to fly in that direction first and the reason may have been to avoid trespassing Indonesian airspace. It is therefore logical to presume that MH370 continued to stay clear of Sumatra imo.
@Oleksandr:
I get it. My attempt at contribution, as I saw it, was the generalizability and possibility to get some input for a signature of what could have taken place there and in what order, but I realize you have been over this many times. One thing follows from another. If they tried to start dumping fuel immediately, as they perhaps should’ve, the electric supply for jettisoning was gone — according to your outline. So they tried to fix that and feared a fire onboard, and couldn’t land obviously etc.
Apart from the technical likeyhood and the proximity between cables (discussed by some) I guess there is also the timing dimension that is a “hotspot”.
@Paul Smithson
I think the actual verbatim statements only suggest that MH370 wasn’t flying over Indonesian territory:
“Indonesian Defense Minister Purnomo Yusgiantoro said the Indonesian military radar placed in the country’s western-most city of Sabang did not detect an airplane flying over Indonesian territory.”
“A spokesman from Djoko’s office, Agus Barnas, later elaborated on Djoko’s statement, saying the most likely route taken by the plane was via the northern Indian Ocean, west of Sumatra, then heading south to a location Najib indicated was the plane’s crash point.
‘Has there a single person who has confirmed the plane flew through our space? No.’ Agus told The Jakarta Post.”
Quoted from the sources above.
Obviously, they couldn’t identify MH370 as its tranponder was off.
@MH
I would rather want to see SIN to PER’s satellite data on that night.
For example Singapore airlines SIA225 takes off around midnight local time, takes about 4h40min and sometimes uses a B777…
***Disclaimer – wild speculation follows***
If we can imagine somebody clever enough to hack the SDU of one 777, why couldn’t he hack two 777 and copy/paste the ID no across.
One plane disappears, the other one lands as usual so nobody will care about that one.
And there you go @Jeff your neatly tied together by satcom ID satellite data is not any more 😉
This also would explain the apparent reboot of a SDU in flight.
First re-logon would be 2nd plane power-up.
One problem then is the BTO arcs don’t match such a flight. But could a different offset (different plane after all) explain that?
Second problem, how to explain the 2nd re-logon at the end of flight and the associated BFOs?
@DrB
re: your recent BFO post
Well, you certainly took a lot of space to make a simple point – that point being that the DSTG data is wrong. You come close to saying that without actually saying that below.
begin cut-paste//
Another point is this: Inmarsat is the only demonstrated “expert” with regard to BFO errors. Certainly the ATSB has no expertise in this area. Based on what is published in their book and on their limited (nonexistent?) previous knowlege of this satellite system, I don’t attach as much credibility to the DSTG’s BFO error analysis as I do to Inmarsat’s.
end cut-paste//
The BFO data presented by the DSTG is not an error analysis. It is a statement of fact. The measured BFO displayed a significant departure from the BFO calculated in the normal manner using satellite and aircraft data. The measured BFO is what it is, and the calculated BFO is not rocket science. It is, actually, a rather trivial calculation for anyone with the requisite math and physics skills. It has nothing to do with a background in satellite navigation which by the way is not the business Inmarsat is in. Inmarsat is a communication company. Not a navigation company. The whole idea of using these signals to produce a navigation solution is akin to turning a sow’s ear into a silk purse.
In any case, I have no doubt that the DSTG has the skill to accurately make these calculations. Citing Inmarsat’s greater satellite experience is just plain lame. Even Inmarsat admitted that these analytics were new and “ground breaking” for them. There is certainly not a well chronicled history of tracking aircraft by this means.
The DSTG made a significant effort to characterize these errors. The implication here is that the calculations that produced the errors were carefully checked, and peer reviewed. It took some guts to include figure 5.4 in the report, IMO. Certainly people from the DSTG did not anticipate that their competence (you used the word credibility above) would be questioned by providing puzzling data in good faith, and candidly admitting they are unable to explain it.
So along comes DrB declaring that he recommends ignoring this anomaly because it does not fit his model of the universe. The only possible explanation is that the DSTG fumbled the ball, and using low single digit BFO errors is the prudent approach to flight path modeling.
I regard your post as disgraceful. Rather than acknowledging the DSTG data for the issue that it is, you elect to declare it to be an artifact of inexperience. In effect what you are saying is to ignore this data, and move on with “analytics as usual”, which by the way have not been successful in locating the aircraft.
DrBobbyUlich
Could you possibly tell me if flying a great circle at constant Mach 0.81 at a constant 35,000ft, from an FMT at IGOGU at 18:37UTC to a 7th arc crossing at S37.62, E89.08 (total distance 2,712Nm) could be flown on the fuel available? Constant Mach 0.81 takes into account the estimated varying air temperature encountered during the flight south, as shown on earth.nullschool.net global map of wind speeds and temperatures.
@Bobby: thanks for responding. I don’t understand how the long descriptions of every flight mode answered my question about BTOs being satisfied by either…
1. Flying at 500 KTAS whilst FIGHTING the wind (in order to maintain a perfectly straight line as viewed from above, or “great circle”), and
2. Flying at 450 KTAS whilst ALLOWING sidewinds to buffet you wherever they may, because you’ve switched to a flight mode that merely keeps you pointed forever in one particular direction (constant HEADING).
In evaluating the above, fuel doesn’t enter the discussion. I’m trying to key in on the unavoidable inference that the winds on the night happened by fluke to blow the plane in such a way that the two changes (reducing speed by 50nmi, allowing sidewinds to buffet you off course) perfectly cancel, and still result in 6 arc intersections to within BTO error tolerances.
I’d be thrilled to receive even just the central spreadsheet containing your parameterization of the fuel charts. I can execute my mission (merely to educate myself on key drivers of fuel burn) without any of the linked files.
In assessing the effect of winds strictly between 5 and 20°S, I simply sighted along the path you plotted. If you used the same wind data I did, the plot clearly curves upwind. Whether the longitudes all decrease doesn’t enter into it – nor do bearings north or south of this range. I’d considered three possibilities: I chose a poor wind data source, the plot was a poor representation of your path, or your path reflected wind effects poorly. Your response eliminated only the first of the three.
I appreciate your thoughts re: PDA’s. I get that PDA’s are a much softer limit on the high side than on the low (erratic paths can always burn away otherwise problematic excess fuel). But I still caution: in your seminal work – which you hotly defended – your fuel burn rates were lower. Significantly lower. You argue the new model is better – which I have no technical basis to dispute. But even with all these refinements – and additional pre-FMT path circuity – you still back-solve to a PDA of 4.7 – some 36% less efficient than the least efficient results in the indicative range you supplied.
It is the COMBO of
– refinements which increased all burn rates AND
– a PDA result which at least nominally says you STILL have too much fuel by Arc 7
…that has me worried.
I’ve worked with detailed models of complex processes long enough to know that intelligence, training and model quality – even working together – are often no match for the insidious lure of confirmation bias. If you still assume each of the ISAT data, shoreline debris, and seabed search are authentic, then you will not stop until your model runs out of fuel at a point compatible with all three.
Since independent analysis has already cast deep aspersions on all three of those assumptions of authenticity, I ask you (and all other flight path modellers) merely to ensure your model is driven by inputs demonstrably calibrated to independently-derived, unbiased estimates – and NOT by the need to reconcile the ISAT data to the debris/search result.
@sinux
The second relogin – mh370 was still flying when SIA225 landed /parked/powered-down. This forced a relogin….
@ROB:
Have a care using nullschool earth, as the beautiful graphics are the results of a modelled pseudo real time system which cannot necessarily give the exact information you require. There may be better data for what you want to do available from here…
https://earthdata.nasa.gov/earth-observation-data/near-real-time/data-outages-and-issues
@Botis
Thank you for the advice Boris. I will take a look at the NASA website, tomorrow.
For once in a long time after you turned to namecalling you are back on track with the facts. That anomaly happened. Because of an MH370 split? It wants to tell us something. Don’t ignore a single thing. Every thing happened for a reason.
All major Austrian newspapers are reporting today, that the Vienna Airport has been targeted by hackers during the past couple of days (causing major disruptions for several days now).
Maybe MH370 was a hacker target, too ?
–> archive.is/2Ya6Q
Interestingly enough, the MH370 investigation was also hacked:
–> bit.ly/2c2IE0a
@DrBobbyUlich
The Inmarsat paper does not make any claim about the distribution of the BFO errors for any of the validation flights. Trying to extract a precise error distribution and standard deviation from the text and one graph isn’t quantitative.
The DSTG paper table 5.1 and figure 5.5 present the detailed data and indicate a standard deviation for the total in-flight BFO dataset of 4.3Hz. I don’t see how an analysis using the best data available today can use a smaller number. DSTG use a conservative figure of 7Hz in their particular autopilot mode analysis but note (sec 10.1) that the BFO data only becomes significant at the detailed level if the BFO error is reduced (artificially) to below 1Hz.
In my own work on models similar to Inmarsat’s I found that if a standard statistical assessment is made (Pearson’s chi-squared test, for example) and the standard-deviation goes much above 3Hz then the allowed range of destinations on the 7th arc gets very large: 4.3Hz is well above that threshold.
DennisW has commented on your point regarding the DSTG credibility.
@Richard Cole:
So we are starting to get an idea of why the plane has not yet been found in the search area. Where should the blame for this misunderstanding of the data and resulting futile search be placed?
@ALL
News.com.au are reporting on three items found recently in Mozambique. One of the pieces is the by now familiar red and white section of fin skin, the other two pieces appear to be new finds. The Director of the Mozambique Aviation said that one of the new pieces could be from a aileron, flap or elevator. The items are being sent to Malaysia for examination.
@MH
That SDU should have been shut down for good at Igari.
Maybe an attempt to restore the hacked SDU to default settings? To hide the hack…
If the payload came through loading a corrupted Sat table into memory, one would have to reboot once to trigger a new download, then a second time shortly afterwards to restore the normal ID.
Hi ALL
The link to the new debris.
Cheers Tom
http://www.news.com.au/travel/travel-updates/malaysia-airlines-flight-mh370-debris-exhibited-in-mozambique/news-story/61e7d9790b8baae6d4827a42d552e47c
@DrBobby,
Once again you are applying your own interpretation to the words in the manuals, rather than thinking about what they mean to an operational crew.
I first asked you for a definitive reference, in particular for your assertion that the aircraft can actually fly a Magnetic Track. You haven’t done so, rather you keep repeating the words in the manuals that we all know anyhow.
Heading hold is simple. No one is debating that, although holding a Magnetic Heading, over an extended time or distance is a different story.
I’m not interested in how you might implement the software. I can think of ways too, but that is all irrelevant.
You don’t have a definitive reference [and neither do I], so there is no point in further debate.
@Tom Lindsay @all
Two of the pieces are actualy earlier found items. The panel on 22 may and the other smaller piece before 6 june both named in the Malaysian list of items.
@Boris
I know that your question was not directed at me, but I will give my answer anyway.
There is no “blame” to be assigned. The assumptions made in the flight path modeling are all reasonable. The assumptions of AP flight dynamics and accurate BFO and BTO data are where anyone would start in the absence of additional information.
One could also ask “blame for what?” – not finding the aircraft? The survival period for the PAX and crew had elapsed long before the underwater search was even initiated. The only thing lost is the time and money spent in the search. I would argue that those considerations are insignificant in the grand scheme of things. No loss of life has occurred as a result of the search operation.
On this list they are items number 7 and number 9:
https://www.dropbox.com/s/o6h7m5qf4prvobm/mh370list1.png?dl=0
@DennisW:
Should have been clearer. I;m just very puzzled as to why it has taken so long to start querying the data. This has only surfaced since it started to look like the current search area could be wrong.
Who is to responsible for what would be a cock-up of massive proportions? The money is not too important, but 2 years of agony for the NOK must be almost too difficult for them to bear. Heads should roll if the assumptions made about the Inmarsat data are wrong, I just want to know whose.
@sinux, maybe MH370 changed their SDU ID for the remainder of their flight… I suspect it was changed to a military like ID. Inmarsat should have checked their logs for other unexpected data on their logs.
Bobby,
On September 1, 12:58 am you wrote:
“The maximum BFO at 18:27 with no climb is very close to 166 Hz on a due North course at LRC (using my BFO model which matches the Inmarsat tables within 1 Hz)…
No turns, circles, spirals, or any other maneuver than a climb (which increases the BFO) can more closely match the BFO data then.”
I disagreed with these statements. Later you suggested me to check my BFO model as you could not achieve values higher than 166 Hz with your model, even at higher speeds than LRC for zero heading.
So I inserted in Yap’s Calculator V2 the following parameters:
Time: 18:27:04
Lat: 6.899 deg (from my model output)
Lon: 95.674 deg (from my model output)
Alt: 10 km (insert any you like)
Ground speed: 500 knot
Course: 0 degree
According to it, the calculated BFO is 170.02 Hz, which is 4 Hz higher than your maximum. For the Ground Speed 570 knots (close to MMO for the respective ambient conditions in my model), I am getting 176.14 Hz from Yap’s calculator, which 10 Hz higher than your maximum.
Are you sure your BFO model is correct?
Brian Anderson, you state: “… holding a Magnetic Heading, over an extended time or distance is a different story.”
Why? The AFDS has all the information it needs. Magnetic heading comes from the ADIRU and is on the same ARINC 629 bus as true heading (as are magnetic and true track.)
Reference: Continental Training Manual, pdf p. 1841, figure captioned “ADIRS – ADIRU INERTIAL REFERENCE DATA FUNCTION – 2” and pdf p. 1843, figure captioned “ADIRS – ADIRU INERTIAL REFERENCE DATA FUNCTION – 3”. Magetic parametes are generated in the “Calculate Magnetic Parameters” block, which takes input from the Angles Integrator, the Velocities Integrator, and the Position Integrator.
@Dennis
“Yes, wiping and erasing are not the same which leads me to believe the terms were not used correctly. If the drives were wiped the data would not be recoverable. Right? I merely assumed, perhaps incorrectly, that the words selected were not selected with care. I still found no use of the term “main drive”.
My sense was both statements were referring to the six deleted points.”
Main drive was my term. I assumed if a drive was wiped it was not the same as the drive where the deleted files were. My bad.
I did some more browsing and think you may be right that many people including reporters use wiping and deleting interchangeably whereas technical people such as myself distinguish between the two.
Sometimes, but not always, the FBI and other forensic data recovery companies are able to retrieve data from a drive that has been wiped. Of course, they have extensive tools that enable them to do that.
@Aunt Bea
I heard the same thing relative to data recovery from wiped drives – sketchy put doable.
@Richard Cole,
You said: “The Inmarsat paper does not make any claim about the distribution of the BFO errors for any of the validation flights. Trying to extract a precise error distribution and standard deviation from the text and one graph isn’t quantitative.”
Inmarsat clearly says the errors are bounded conservatively by +/- 7 Hz. That by itself limits the standard deviation to a value significantly less than 7 Hz. In addition, inspection of Inmarsat’s Figure 15 reveals that of the 30 BFO measurements plotted on one flight, 25 are within +/- 3.5 Hz and only 5 are between +/- 3.5 and +/- 7 Hz. Although the data are limited, they clearly show a strongly peaked probability density function, and they imply a standard deviation less than 3.5 Hz (+/- 0.7 Hz).
@DennisW,
You are mischaracterizing my words. I clearly said “ . . . , I don’t attach as much credibility to the DSTG’s BFO error analysis as I do to Inmarsat’s.” You are free to assign credibility wherever you wish, but please do not criticise and misrepresent others who do the same.
@SK999,
I am aware of that reference too. But again, we are trying to infer what the software does from a block diagram of various inputs. I agree that the Magnetic parameters are calculated [with Mag variation determined from a look-up table]. But what then? For display on the CDU and MCP, as other references state.
One of the more interesting descriptions of what actually happens is here . . . http://www.ohio.edu/people/uijtdeha/theavionicshandbook_cap_15.pdf
Again, one must read the chapter from the point of view of an operational crew, but at least it makes many interesting statements, including . . “Not all terminator types can be used with all leg types. For example, a track leg can only be terminated by a fix since the definition of a track is the great circle path between two geographic locations (fixes).” . . . and “Conversion to magnetic is typically performed just prior to crew presentation.”
It is all too easy to misinterpret words and button controls to conclude that the aircraft can do things, when in fact there is no such software implementation, and often no operational need to ever require such mechanisms to even exist.
IMO, whoever was in control of the aircraft would be keen to avoid other aircraft as well. That person would need intimate knowledge of air trafic in the area. It is not just about avoiding Indonesian airspace to circumvent detection. Perhaps circumvention was also needed for other reasons.
Revisiting some of the other information about the flight.
Take Inmasat data analysis out of the picture.
12.41 MH 370 leaves KLIA
At 1.07am that the plane’s Acars signalling device sent its last message before being disabled some time in the next 30 minutes, apparently deliberately. A separate transponder was disabled at 1.21am but investigators believe the Acars was shut down before Hamid’s final, 1.19am farewell.
from the moment of sign-in at 12.36 when the plane was still on the ground, Hamid, a 27-year-old flying enthusiast, gave routine accounts of its location, ascent and altitude.
There is 54 minutes of communication with cabin.
Routine, lots of dialogue within the cabin from flight deck and from flight attendants. MAS pilots and crew make at least a dozen announcements during that time. They announce everything in three languages, Bahasa, English and Chinese. During these routine announcements they interrupt movies, the crew are busy serving drinks and dinner. It does not matter that this was a red eye flight. They are walking the aisles doing their routine for over an hour.
It is at about 90 minutes that the lights are turned off. Only a few can sleep during the first 90 minutes of these flights, there is just too much activity within the cabin.
Why did the pilot ascend to 43,000 to 45,000ft for 23 minutes? Was it to try to extinguish a fire or the passengers?
Brian,
The FCOM is unambiguous: “Heading Reference (HDG REF) Switch: Pushing alternately selects the heading reference for the PFDs, NDs, AFDS, and FMCs.” HDG HOLD in NORM mode means magnetic heading hold (a la KAL007). The only exception to NORM is when in polar region (not applicable here.)
The statement, “Conversion to magnetic is typically performed just prior to crew presentation” is given in the discussion of the flight management system, not the ADIRU or the AFDS. You hang a lot on this statement but it is meaningless in the context of “magnetic heading hold” for a B-777. which is an AFDS function, and in fact is not even mentioned in that chapter. As you are well aware, both the ADIRU and the FMC have magnetic tables (which need to be kept in sync!) Further, the chapter defines “track” as follows: “the definition of a track is the great circle path between two geographic locations”. At that point I stop reading.
You have still not answered the question about how it is that “holding a Magnetic Heading, over an extended time or distance is a different story.”