[Peetwo]

Recent analysis/review, on Duncan Steel's website, by members of the IG regarding the ATSB 3 December 2015 released document - titled MH370 - Definition of Underwater Search Areas & DSTG analysis :

Comments and Questions Regarding the ATSB Report on MH370 of 3 December 2015 

Response to the ATSB Report and DSTG Book on MH370

Richard Godfrey and Victor Iannello
9th December 2015

 
Introduction
The Australian Transport Safety Bureau (ATSB) published an update to their report entitled “MH370 – Definition of Underwater Search Areas” on 3rd December 2015. At the same time, the Australian Defence Science and Technology Group (DSTG; formerly the DSTO) through the ATSB made available a book entitled “Bayesian Methods in the Search for MH370”. Both may be downloaded as PDFs (1.76 MB and 8.76 MB respectively) via this webpage.

At first sight, the search area centred on a point located at 38.0S 88.4E aligns very well with the end point published by the IG (over 14 months ago, on 26th September 2014; see here also) at 37.71S 88.75E. In fact, these two points are only 24 NM apart.

However, when one digs a little deeper, there are obvious flaws in the Bayesian approach adopted by the DSTG team and the belated confirmation of the IG end point by the DSTG is more due to error than design. There appears to be a fundamental disconnect between the mathematics employed and the constraints imposed on possible paths based on the aircraft dynamics and satellite communication data. In this brief post we summarise how and why this is so.

Bayesian approach
The BFO and BTO data are stochastic in nature, so assigning random variables and calculating probability density functions for the derived end point from paths based on the uncertainty of these variables seems reasonable, especially if other constraints are considered such as human factors and aircraft performance characteristics.

The path travelled, however, was most likely not stochastic: there was a deterministic reason for whatever path the plane took, be it human intention or a particular failure mode. And that particular path could have a very low probability if it is just one of many possible random paths that are generated in the way the DSTG researchers generated their possible paths.

Bias toward straight paths and against curved paths
The DSTG posterior distribution has a high incidence of straight paths, which peaks the end point distribution around 38S. In fact, if all speeds and turn times are equally probable, the distribution should be much flatter. By virtue of the way the DSTG generate paths using random manoeuvres and random changes in heading and speed, curved paths (which require more turns) are not represented as often in the posterior distribution.

This bias for straight paths can be seen in Chapter 7 of the DSTG book. Figures 7.2 and 7.4 show paths for which the BTO and BFO are ignored completely. There is a hot spot for the end point in Oman, corresponding to a straight flight starting at 18:02 and proceeding northwest.

The assumption made, then, is that the plane flew straight between manoeuvres, so a flight with few manoeuvres flies relatively straight. This produces a bias in the analysis in favour of straight paths and against curved paths.

Bias toward faster speeds and against slower speeds
In the DSTG study, the speeds after manoeuvres are chosen to be randomly distributed between Mach 0.73 and Mach 0.84 based on fuel considerations. In this brief section we indicate why the choice of lower speed limit is erroneous if fuel consumption is the only consideration.

At FL350 and DTisa=0, these Mach speeds corresponds to True Air Speeds (TAS) of 421 knots and 484 knots respectively. At DTisa = 10K, the TAS range is 430 to 495 knots.

The Holding Speed of an aircraft is defined as the speed at which fuel flow rate is at its minimum, for any specified altitude. Over the interval from 18:28 to fuel exhaustion the average weight of the plane is about 192.3 tonnes. Using this average as a constant weight throughout that interval, at Holding Speed at FL200, and with DTisa = 0, the appropriate Indicated Air Speed (IAS) is 213 knots, which corresponds to Mach 0.47 and a TAS of 288 knots. At DTisa = 10K, Mach 0.47 corresponds to a TAS of 293 knots.

There is therefore a fundamental error made in setting a lower limit of TAS equal to 421 knots when speeds of around 288 knots would have sufficient endurance if the aircraft flew at a level around FL200.

Because fuel considerations allow lower speeds than M0.74, the DSTG model has an inherent bias toward faster speeds.

Cost Index
The Cost Index (CI) has finally been made public (a request for its publication was made in September 2014) and we are told that the flight was planned on the basis of a CI of ECON 52. The resulting MH370 Flight Path Model V16.0 is published elsewhere and shows an end point of 38.4S 87.7E.

Radar Data
The DSTG Bayesian approach is based on a prior defined by the Malaysian Military Primary Radar at 18:01 (penultimate 10-second radar capture).

The single capture from the Malaysian Military Primary Radar at 18:22 was not used by the DSTG because it is thought likely to be less accurate, near the radar system’s maximum range. However, this single radar capture has been shown to align with the satellite data a few minutes later.

The Malaysian authorities’ previous public releases of their Military Primary Radar data is not consistent with the Malaysian Military Primary Radar available to the team at DSTG, as it shows a detailed radar trace with a large number of data points after 18:01:49 up to 18:22:12. For example, see the photograph below.

We request the following:

1. Publication the 10-second Primary Radar data from 16:42:27 to 18:01:49.
2. An explanation of the source of the Primary Radar trace between 18:01:49 and 18:22:12 (as shown above) and whether the Malaysian authorities have made these radar-determined positions available to the ATSB and the DSTG.

BFO Bias
The BFO bias was found by the DSTG researchers to have a geographic dependency, but it was not possible for them to determine the cause. A geographic dependency can be ruled out, as the laws of physics do not change with location and time. As this uncertainty in the bias causes a large uncertainty in the overall BFO measurement, it needs to be better understood.

Conclusion
In summary, the recent DSTG report presents a generalized, Bayesian method to prioritize the search zone. Although the ATSB is using this report as justification of its current search area, the alignment between the new results and previous results are a result of an inherent bias towards straight, high speed paths. More generalized methods would allow a larger range of end points. This bias should be acknowledged.

MH370 Flight Path Model version 16

Richard Godfrey
9th December 2015

 
I have updated my MH370 Flight Path Model to include a simulation of the Cost Index (CI) at ECON 52, now that the CI has been made available in the latest ATSB report. This updated model is labelled as version 16.0, the preceding version 15.1 having been described here.

The model version 16.0 (Excel spreadsheet, 2.5 MB) may be downloaded by clicking here.
I retained a late final major turn (FMT) starting at 18:37:24 and 15.6 NM before the waypoint IGOGU.

I also retained a step climb to 39,000 feet.

The resultant end point is located at 38.4S 87.7E.

Update 10th December:

Version 16.1 of the Flight Path Model is now available here. Version 16.0 entailed a constant speed being used between ping arcs 6 and 7, whereas a deceleration would be expected (a value of -0.315 knots per second has now been used here in v16.1, in line with Brian Anderson’s analysis).

The resultant end point is now 38.2S 88.0E.

Thanks to Brock McEwen for pointing this out.

Misinterpretation of Aircraft Control Modes in the DSTG Book on MH370

Brian Anderson
14th December 2015 

 

1. Introduction 
   
   

The Australian Transport Safety Bureau (ATSB) published an update to their report entitled “MH370 – Definition of Underwater Search Areas” on 3rd December 2015. At the same time, the Australian Defence Science and Technology Group (DSTG; formerly the DSTO) through the above ATSB webpage made available a book entitled “Bayesian Methods in the Search for MH370” (hereafter, DSTG book).

Some of the statements made by the DSTG authors warrant further attention and explanation. Specifically, the statements made in relation to control angles and headings in 6.3.1 thru 6.3.5 of the DSTG book are inaccurate and misleading, although the way that these factors are incorporated in the analysis may not change the result of the study.
In this post I explain how and why those statements are misleading, why the errors made might be of fundamental importance in the search for MH370, and indicate what could and should be done in terms of correcting the analysis. The intent here is simply to identify mistakes made in order to assist the official search teams.
 

1. Lateral Control Functions
   
   

All the descriptions I have read of the roll functions for the B777 aircraft – and probably the Continental B777 Training Manual is the most detailed on the subject – indicate that magnetic parameters are calculated with reference to a look-up table of magnetic declination by using inputs of angles, velocities and position derived from the Inertial Navigation System (INS) and earth references (as opposed to the magnetic parameters being directly measured). For example:

Ref: section 34-20-00, page 41 of the Continental Manual (page 1839 of 3328)

Magnetic Heading And Magnetic Track
Rotational rates and linear accelerations go to the rotate accelerations to earth reference function. The rotate accelerations to earth reference function calculates accelerations in relation to the earth. The accelerations in relation to the earth go to the velocities integrator. The velocities integrator calculates velocities. The velocities go to the position integrator and the calculate magnetic parameters function. The position integrator calculates airplane position.

The calculations described above are performed in the ADIRS – ADIRU system, shown in block diagram form in section 34-20-00, page 43 of the Continental Manual (page 1841 of 3328), as below: .
The roll control functions are described as follows:
  
Heading/Track Hold
In this mode, the airplane holds either heading (HDG) or track (TRK).
If the HDG/TRK display on the MCP shows TRK, the airplane holds track.
If the HDG/TRK display on the MCP shows HDG, the airplane holds heading. 

Heading/Track Select
In this mode, the airplane turns to the heading or track that shows in the heading/track window.
If the HDG/TRK display shows HDG, the airplane goes to and holds the heading that shows in the heading/track window.
If the HDG/TRK display shows TRK, the airplane goes to and holds the track that shows in the heading/track window. 

It is not explicitly stated in the manual whether the aircraft will (at least over a period of some hours) hold the heading or the track if the heading display is in Norm (i.e. displaying a magnetic heading, and thus follow a curved path).
(With regard to the distinction between the terms “Heading” and “Track”, please see the addendum at the end of this post.)

Further evidence that magnetic angles are in fact calculated and not measured can be ascertained from Aero Quarterly (issue: Qtr 04/09), as published by Boeing, discussing Correcting the Effects of Magnetic Variation, and emphasizing the importance of airlines updating their inertial reference systems to the latest magnetic variation (MagVar) tables in order to avoid potentially hazardous magnetic heading-related navigation errors.

The Heading- and Track-Select functions serve only to turn the aircraft onto a new Heading. Once that Heading has been acquired the roll control functions serve to maintain the Heading or the Track automatically.

Considering the possibility that the Heading or Track may be displayed with reference to True North [True], or Magnetic North [Norm], there are therefore four possible combinations of Heading and Track Hold that might ultimately determine the path of the aircraft, as follows:

(i) TRK Hold [True]
In this mode the track of the aircraft is maintained constant with respect to True North, and small instantaneous adjustments to the aircraft heading are performed to compensate for wind vectors so as to achieve this. The resulting flight path describes a loxodrome.

(ii)        HDG Hold [True]
In this mode the heading of the aircraft is maintained constant with respect to True North. No adjustments to the aircraft heading are performed to compensate for wind vectors. The resulting flight path would describe a loxodrome in zero wind conditions, but is otherwise continuously modified by wind vectors.

(iii)       TRK Hold [Norm]
In this mode the track of the aircraft is maintained constant with respect Magnetic North, and small instantaneous adjustments to the aircraft heading are performed to compensate for wind vectors to achieve this. The resulting path will be curved. Since the magnetic heading is calculated in the ADIRS – ADIRU system the aircraft heading (with respect to the earth) must be continually modified and updated by incorporating the declination obtained from the MagVar tables relating to the current aircraft position, and the wind vector.

(iv)       HDG Hold [Norm]
In this mode the heading of the aircraft is maintained constant with respect to Magnetic North. No adjustments to the aircraft heading are performed to compensate for wind vectors. The resulting path will be curved as in (iii) above in zero wind conditions, and is otherwise continually modified by wind vectors. Again, the aircraft heading (with respect to the earth) must be continually modified and updated by incorporating the declination obtained from the MagVar tables relating to the current aircraft position.
 

1. Lateral Navigation
   
   

Lateral Navigation (LNAV) is a specific control mode whereby the aircraft typically follows a series of waypoints previously stored as a flight plan in the Flight Management System (FMS). The track followed by the aircraft between waypoints is (a segment of) a Great Circle. The aircraft heading is continuously adjusted for wind vectors to maintain this track.

Waypoints can be defined in many ways. They are not limited to predefined Instrument Flight Rules (IFR) waypoints, and can be entered into the FMS (more specifically, the Control Display Unit, or CDU) to modify the flight plan at any time. For example it is possible to describe a waypoint simply as a set of latitude and longitude coordinates (such as “N071.2E0953.15”, which represents a waypoint at  N7° 1.2’ E95° 3.15’, and is displayed as N07E095).

Section 6.3.5 of the DSTG book states, in reference to the LNAV control mode:

“Expert advice indicates that if the autopilot system is operating in lateral navigation mode and it reaches the final programmed waypoint, then it reverts to the previously selected heading hold mode.” 

This statement is at variance with the description of LNAV operation in the Honeywell B777 FMS Manual, and with the description found in a variety of other B777 FMS references.

The Honeywell description of the LNAV action after overflying the last programmed waypoint is this:

“. . . DISCONTINUITY is displayed in the scratchpad, and the aircraft maintains its existing track.” 

Other references state:

“LNAV maintains current heading when passing the last active route waypoint.” 

The different terminology used in different references (Track vs Heading) may simply be one of common usage of the terms when describing an instantaneous point in flight (i.e. the instant that the discontinuity occurs).

It is interesting to note, from the Honeywell manual, that:

“Only the active waypoint course can be referenced to magnetic north because the ADIRU can provide magnetic variation only for present position.”

This would suggest that the course followed after passing the last waypoint, which by definition cannot be an active waypoint course, cannot therefore be referenced to magnetic north.
 

1. How is this relevant
   

Considering the aircraft path southwards from a position in the northwest of the Malacca Strait:

(i) If this path was being managed in LNAV mode (with or without manual input), then the final turn south must have been at or about a predetermined waypoint, and furthermore the track from that point must be a Great Circle to another waypoint that had previously been manually entered into the FMS via the CDU. Clearly, manual input of at least two waypoints is therefore necessary at some time before reaching the final turn in the Malacca Strait. After overflying that final waypoint (if indeed it was actually reached), then the aircraft would be expected to continue on the same track. At that point the track may become a loxodrome.

(ii) There is no evidence to exclude the possibility of some flight manoeuvres towards the end of the track in the Malacca Strait, which would have delayed the commencement of the flight south. For example it is entirely possible that the aircraft completed an orbit of greater than 360 degrees, perhaps 460 degrees even, before proceeding south. Any delay or manoeuvres such as this can be shown to still fit the available satellite data, in particular an intercept with an arc meeting the BTO at 19:41 UTC. However it seems most unlikely that such manoeuvres would have been manually entered in to the CDU to become an active flight plan.

(iii) An alternative to the final turn south being a turn at a waypoint is the possibility that the turn was initiated manually using the Heading- or Track-Select controls. After selection the aircraft rolls onto the desired heading or track, and then holds the selection, as described in 2(iii) and 2(iv) above. (Note that this selection also effectively overrides the LNAV function.) It is therefore necessary to consider flight paths consistent with these control modes.

(iv) We do not know the control mode in use at the time of the final turn south. However, one might be biased toward Norm (magnetic reference) rather than True, since Norm is the de facto standard for most flight manoeuvres where these control modes would be used. For example, virtually all tracks and headings described on aviation navigation charts are referenced to Magnetic North, as are virtually all instructions from Air Traffic Control (ATC).

(v) While TRK Hold [True] would result in a relatively straight path (a loxodrome) from the final turn, each of the other three control modes would result in curved paths, in fact curving significantly eastwards as the flight progressed, due to increasing westerly winds at more southerly latitudes, and the magnetic declination increasing more rapidly.

(vi) The DSTG analysis seems heavily biased toward straight paths after the final turn south, and it is therefore not surprising that a “hot spot” at about 38 degrees south results. On the basis that it seems equally likely that curved paths may have been followed, or indeed the turn south commenced further to the NW in the Malacca Strait, or occurred at a later time, it is of concern that termination points further NE on the 7th arc have not been shown to be viable, and perhaps even equally probable, given our lack of knowledge of the aircraft control modes in operation during MH370’s flight southwards.

(vii) It can be shown that curved paths, such as might result from control modes described in 4 (ii), (iii) and (iv), can also fit the Inmarsat-derived data, and end on the 7th arc, but further to the northeast on that arc than the “hot spot” indicated in the DSTG book and on the cover of the latest ATSB report update. Such curved paths, however, are not at constant speed. Speed variations, and in particular significant slowing of the aircraft towards the end of the flight, are necessary in order to fit the Inmarsat data with these paths.

(viii) While TRK/HDG Hold [Norm] have their place in managing aircraft navigation – and in particular in respect of standard procedures, navigation chart information and under instruction from ATC – the requirement in these situations is for only relatively short times and relatively short distances. In such situations the magnetic declination does not change significantly and the resulting aircraft azimuth remains essentially fixed.

Furthermore, that azimuth would be precisely the same whether the displayed track or heading was referenced to Magnetic or True North. In contrast, maintaining TRK/HDG Hold [Norm] over lengthy times or long distances has absolutely no value. The fact that either of these two modes may have become the default lateral control mode over something like the last six hours of the flight of MH370 serves only to compound the problem of identifying the final resting place.
 

1. Conclusion
   

The analysis in the DSTG book should be reviewed after correcting the misinterpretation of the aircraft control modes as presented in Chapter 6 of that book.

Also of interest is this article from Planetalking today:

New details about MH370 search come to light

Ben Sandilands | Dec 17, 2015 1:22PM |

Just a tiny part of the seventh arc search zone

New insights into the search for missing flight MH370, and what happens the moment the jet is found, were explained at a briefing for members of the Surveying and Spatial Sciences Institute in Perth recently.

Paul Kennedy, MH370 Search Project Director, Fugro Survey, took the audience through the same induction video shown to recruits before they joined the three vessels that are sonar scanning the priority zones in the southern Indian Ocean that are believed most likely to contain the sunk wreckage of the jet.

This YouTube of his presentation gives but a distant low resolution view of Mr Kennedy’s graphics, but it is what he says that is of interest.

At the outset he reminds the audience that the so called seventh arc along which the priority zones are located is different from the earlier arcs which represented places MH370 could have been when it send automated pings to an engine performance monitoring site at Rolls-Royce in Derby via an Inmarsat satellite and ground stations.

This last and unexpected transmission from the jet was initiated by the failure of its engine generated electrical supply and the automated deployment of a ram air turbine that popped out of the fuselage to generate emergency power as the jet fell toward the sea.

With the help of audience members Mr Kennedy unrolled a large scroll of the seabed in the seventh arc priority zone that stretched across the room and out one door which had mapped its features and their depths with sufficient accuracy to prevent deep sea ‘tow fish’ or sonar scanning platforms being towed into the side of underwater obstacles like volcanic craters while surveying the terrain from a height of around 100 metres.

On that map MH370 would only have been a small dot, and it is of course covered in dots, and often convoluted terrain, with deep fissures and troughs and sea mounts and cliffs.

Mr Kennedy said the Boeing 777 would only have subtended half a millimeter in size on the waist high scroll unfolded across the room.

He explained that the raw data was not just being recorded on the search vessels in real time, but being uploaded to ‘cloud’ computers via specially targeted reception beams directed to them from satellites to provide a high bandwidth tunnel.

That data was being simultaneously but separately reviewed “by multiple sets of eyes in Australia and the US”. (He made no reference to Malaysia in relation to the uploads.)
This was occurring in sea states in which peak waves had been recorded as 16.5 metres at their highest, but more commonly of around 11 metres , and often around 6-8 metres.

(There is a great deal of information in the presentation on the working environment, and the health and safety precautions taken on ‘swings’ that included around 13 days total sailing time from Fremantle to reach and return from the priority zones).

Mr Kennedy said there had been prolonged periods when the search vessels had to maintain a heading into winds of around 150 kmh, unable to turn to either side in the sea states that prevailed. He said the crews working shifts covering continuous operations were holding onto the sides of their bunks while trying to sleep, causing considerable fatigue toward the end of periods of duty.

He said the seabed search includes ‘nightmare’ zones when towering complex cliffs and pot holes within volcanic craters may hide the wreckage of the Malaysia Airlines 777-200ER .

The Norwegian Hugin autonomous underwater vehicle or AUV was being used to explore the otherwise impossible to scan wrinkles which could contain parts of MH370. This had included descending deep into features inside volcanic cones, and continued to look into a 200 kilometres long 70 degrees slope that was around 1000 metres in height above the adjacent sea floor.

This slope contained ledges and fissures that the regular tow fish could not look into.
By coincidence on the first anniversary of the disappearance of MH370, a towfish had come across an unnatural set of objects which looked as though they could have come from the missing flight.

However when it was studied in closer detail by the AUV it was found to be an unchartered, and so far, untraced shipwreck.

Mr Kennedy said that some maritime detective work had established that the anchor was probably made around 1820. The actual wooden ship had disappeared over time, and all that was left was the anchor, the ship’s bell, a spread of metal nails and fixtures, and a large sea chest.

What happens when MH370 is found?
Once it is clear that MH370 has been found the satellite data links to the vessel will be shut down and immediate steps made to notify the next of kin of the 239 people on board the jet.

Mr Kennedy’s brief comments about this made it clear that the search intends, commendably, to ensure that those who lost their loved ones on MH370 will be first to learn that the jet and thus their final resting places, have been found.

They will not learn of it through the media. (However it is highly probable that once next of kin have learned of a discovery, word will reach the media, possibly before but most likely only shortly before, an official media briefing is called. These are this reporter’s observations, not Mr Kennedy’s).

Other revelations in the briefing include a sea floor duration of up to 32 hours for the AUV and a very important checking process that validates the performance of the equipment being used in the priority search area. The AUV doesn’t upload in real time, but sends tracking information to the surface, and on recovery after a dive then downloads its data.

Boeing has set up a test and calibration field on the sea floor that each vessel passes through on its way out of and into Fremantle with each tour of duty.

This field includes objects that would correspond to parts of the sunk wreckage of MH370, such as the engines and other heavy components, and replicates their expected resolution to sonar scanning tow fish.

The scientist who made this YouTube record of the presentation said,
“This is incredibly important as when you do surveying you must have confidence that all the work you have done is verified.

“If upon returning to port and doing a closing check on the way back, the items in the test field were unable to be sighted in the detail they had expected this would not only detect a fault with the equipment, but almost certainly void ALL of the past several weeks of deep ocean scanning carried out.

“You wouldn’t know if your equipment stopped working at 100 percent two weeks or 2 hours ago!

“So doing a return scan over the items off the coast of WA would be a great relief to all of the crew on board that all their work is valid.”

MH370 vanished early on a flight between Kuala Lumpur and Beijing on 8 March 2014, when the jet was over the Gulf of Thailand and about to enter air space under Vietnamese control.

Here is the presentation on Youtube, definitely well worth viewing...

MTF..P2