MH370 – Now for the facts

1886
August 02, 2016

If there’s one group of people who desperately want to find the wreckage of Malaysia Airlines flight MH 370, says the man guiding the search effort, it’s the experts who have been involved over the past two years.

There are still 200 people looking for the Boeing 777-200ER and its 239 lost souls and The Australian Transport Safety Bureau’s Peter Foley says none of them will get a decent night’s sleep until it is found.

Foley has seen firsthand the angst of families in Australia and recently met Malaysian and Indian families in Kuala Lumpur to talk about the search for their missing loved ones. “You talk to any one of my guys and they are absolutely fully focused on finding that plane because they want to take those folks home,’’ he says.

The ATSB has been under fire lately for not dropping its current search plans and sailing off to chase the theory that plane’s captain, Zaharie Ahmad Shah, ditched the aircraft under full control to keep it intact and make it harder to find.

The bureau has been accused of wasting money, ignoring the so-called “rogue pilot” theory for political reasons and one Australian newspaper even questioned whether the search was “a sham”.

The vilification has been frustrating for the Australians charged with finding the plane and who still insist there is no evidence collected so far that indicates a control input after the plane turned southward on its final leg to the Southern Indian Ocean.

Malaysians ignored evidence

Airline chief says its a coverup

Most in the aviation industry believe there was human intervention at the start of the flight, with the pilots the most likely candidates, but international regulations mean it is up to the Malaysian lead investigators to determine and comment on what led to MH370’s mysterious disappearance.

In terms of the controversy about the end of the flight, Foley says the ATSB has been doing what any crash investigator around the world would do: following the facts and the science.

Like fictional detective Sherlock Holmes, aviation investigators eliminate the impossible to get to the truth. They sift through wreckage and data, gathering as much information as possible to construct a hypothesis that best fits the facts. If the facts change, they adjust the hypothesis.
While not perfect — the bureau was heavily criticised over a flawed report over the 2009 crash of a Pel-Air jet off Norfolk Island – the ATSB is highly regarded by its peers and the global airline community as being very good at its job.

When it came to the biggest aviation mystery in the 100-year history of commercial flying, however, there was no wreckage and the facts with which the air safety sleuths had to work were few and far between.

So how did the ATSB come to decide that the 120,000 sq. km, search still under investigation was the best place to look?  And why do they think that’s still the case despite a barrage of often noisy criticism?

When the Malaysia Airlines Boeing jetliner disappeared while travelling from Kuala Lumpur to Beijing on March 8, 2014, nobody knew where it had gone
Someone, possibly Zaharie, had turned off the communications between the aircraft and the ground, including its transponder, and effectively rendered it invisible except to primary radar.

Bungling in Malaysia and some doubtful assumptions meant the initial search lost precious time looking in the wrong area and it was not until a breakthrough analysis of satellite signals that attention focused on the Southern Indian Ocean.

This was despite the fact, as revealed in an interim report released last year, that the aircraft was tracked by both military and civilian radar after it was cleared to take off from Kuala Lumpur at 16:43 co-ordinated universal time (UTC), or 12.43 am Malaysia time (UTC is the global standard for time and, for most purposes, is equivalent to Greenwich Mean Time).

The report showed the aircraft’s Automated Aircraft Communication and Reporting System (ACARS), a digital data link between the aircraft and ground stations, made its last transmission at 17:07 UTC. The aircraft’s mode S transponder, which allows secondary radar to track its position, dropped off the radar display at 17:20, about a minute after the last radio transmission of “Good Night Malaysia Three Seven Zero’’.

It was seen by military radar to be turning right but then almost immediately made a constant left turn in a south-westerly direction. Contrary to initial reports that the aircraft climbed to a height of 45,000ft, (it was actually too heavy at the time to reach this altitude), it was later found that the radar was not calibrated for accurate altitude data.

About 13 minutes prior to the plane passing slightly to the south of Penang, it was around 33,000ft with a ground speed of 529 knots. The aircraft dropped off civilian primary radar after it passed the island at 17:52 UTC but military radar continued to track it headed west-north-west towards Pulau Perak, a small island over the Straits of Malacca, which it reached at 18:03 UTC.

The military radar tracked the aircraft for another 19 minutes until it disappeared abruptly about 10 nautical miles after a waypoint called MEKAR to the north-east of Banda Aceh in Indonesia.  At that point, it appeared to be tracking north-west along an air route known as N571.

It was a breakthrough analysis of satellite signals by British company Inmarsat that gave the first inkling that something totally unexpected had happened. The satellite data consisted of seven “handshakes’’ between the plane, an Inmarsat geostationary satellite over the Indian Ocean and a ground station in Perth.

Although the plane had been rendered effectively “dark’’, a satellite system designed to transmit aircraft and telephony data was still active.It initiated a log-on request at 18:25 UTC and again at just before 00:19. Between those two handshakes the ground station initiated five routine log-on requests — it automatically interrogated the plane if it hadn’t heard from it for an hour —  and there were also two unanswered ground-to-air telephone calls.

It wasn’t much but it was enough to work out through the delay in the signals between the plane and the satellite — a phenomenon known as burst timing offset — a series of circles on which the plane must have been when the handshakes took place.

Investigators combined this with the known fuel and operating characteristics of the Boeing to reduce this to an arc, but this still cut a vast swathe across the surface of the Earth and extended as far north as Kazakhstan.

A second breakthrough came when experts were able to use changes in frequency, or burst frequency offset, to determine which direction the plane had headed. This involved a highly sophisticated treatment of Doppler shift, the same phenomenon that results in increase and decrease in pitch people hear when a fire engine goes past.

The ATSB, which had been asked by the Malaysians to lead the underwater search, brought in local and international expertise to help it check and refine the calculations to work out where best to look. This included The Australia’s Defence Science and Technology Group (DSTG), the UK Air Accidents Investigation Branch, plane maker Boeing, the US National Transportation Safety Board, Inmarsat, and French aviation technology company Thales.

Various agencies were asked to check the calculations and when they did, the results were similar but not exactly the same. To make sense of the differences, the DSTG was asked to provide expert data analysis to provide the most appropriate method.

They decided on a statistical approach that combined satellite data with various permutations of aircraft dynamics and available environmental information to produce a probability distribution encompassing 300,000 potential flight paths.

They checked this against the real world by performing 60 validation experiments using 10 satellite data sets from six flights. The result was a probability distribution curve with the 120,000 sq. km search area corresponding to 95 per cent of the probable impact points in the southern part of search zone and 90 per cent in the north.

“It’s absolutely the best analysis we’ve got and indeed it’s based on a fair bit of hard science and heavy maths done by DSTG,’’ Foley says. “Their modelling is a good as it gets with the data we’ve got, refined to the point of no further refinement possible really.’’

One key to the hypothesis the aircraft was uncontrolled in its final moments was the final logon-request from the aircraft.  While it was routine for the ground station to interrogate the plane, the reverse was highly unusual and it happened twice on this flight.

The first coincided with the loss of communications with the plane and may have been due to a power disruption when the systems were disabled.
When power was restored, the satellite data unit on the plane initiated a log-on request that included the inflight entertainment system setting up a ground connection for an SMS/email application.

In the final transmission, the plane sent a log-on request and successfully logged on eight seconds later but without the expected entertainment system request.

The log on request came at 00.19:29 UTC when the aircraft had been airborne for 7 hours and 38 minutes and investigators believe it coincided with the aircraft running out of fuel.

A 3 per cent difference in engine efficiency made it probable the right engine flamed out before the left and investigators estimate the left engine would have continued working for a further 15 minutes. During this time the autopilot would have compensated for the uneven thrust but investigators say it would not have been able to maintain an altitude above 29,000ft.

Electrical power in the aircraft was provided by generators on the engines but there was third generator, generally used on the ground, called the auxiliary power unit (APU).

When the left engine flamed out, the autopilot would have disconnected and a ram air turbine (RAT) unit designed to supply power to basic controls would have dropped down underneath the plane.

Power to the satellite data unit would have been cut and it would have taken the APU, which is designed to automatically start when power is lost from both engine generators, about 60 seconds to come on line.

Even then, its contribution would have been short lived with investigators estimating it had just 30lbs of residual fuel available to it.

But allowing for changes in the pitch of the aircraft and accelerative forces, it was calculated there would be enough to fuel to supply the APU and provide power to the satellite unit for the two minutes it took log on.

Contrary to some claims, the researchers looked at the possibility of a controlled glide at the end of the flight but found it was inconsistent with the satellite data.

Investigators estimated the jet could have glided an additional 100 to 125 nautical miles under pilot control but also noted a controlled ditching would require engine thrust to properly control direction and vertical speed at touchdown as well as hydraulic power for the flight controls.

“The final SATCOM transmission was considered by the satellite working group to be due to a power interruption to the SDU,’’ a report released in December said. “Given the performance analysis by Boeing, it is entirely reasonable to assume that engine flame-outs triggered the APU auto-start that restored power to the SDU. “This evidence is therefore inconsistent with a controlled ditching scenario.”

Boeing used sophisticated engineering simulators to test various end-of-flight scenarios involving a right then left engine flameout with no control inputs. But even this had its limitation due to a lack of information about basics such as starting altitude and systems configurations,

The simulations initially pointed to the aircraft descending in spiral with a low bank angle left turn and hitting the water close to the seventh arc some 10 nautical miles (18.5km) forward and 10 nautical mile to the left of where the left engine flame out occurred and resulted in the searchers adding 10nm to the search width on either side. It also found the aircraft may have exceeded its design specifications as it descended in a wave-like “fugoid’’ involving oscillations in altitude.

At the request of the ATSB, Boeing recently ran further simulations based on different altitudes, electrical configurations and models of turbulence to better gauge the impact on how the aircraft descended.

“That resulted in quite a bigger box than 10 by 10 nautical miles ahead and to the left of the track which has, in a sense, helped us determine how far west of the seventh arc that we need to search,’’ Foley says.

“It tells us that the aircraft in all likelihood performed a series of fugoids in a spiralling path, it didn’t in fact become rapidly unstable and quickly crash. Had that been the case we would have found it by now because it would have been somewhere pretty close to the seventh arc.’’

An analysis of the burst frequency offset at the last arc indicated the aircraft was descending at between 5400 ft per minute and 10,000 ft per minute when it sent the log on request.

When the acknowledgement comes from the ground station eight seconds later, experts calculate it was plummeting at between 12,000ft per minute and 20,000ft minute.

Foley says this scenario was consistent with the Boeing simulations and was a further argument against a controlled ditching.

Another piece of key evidence is sitting in the ATSB laboratories in Canberra. Failure analysts are looking a large piece of the aircraft’s right main wing flap, a moveable panel used to increase the surface area of the wing when the plane slows for a landing, to see if it was extended on impact. If this proves not to be the case, it will serious blow to the controlled landing theory.

Foley says it would rule out the availability of power at the end of the flight to the main hydraulic pump used to extend the flaps. While this does not rule out a controlled glide without power, Foley says it’s hard to understand why anyone would do that with a large fly-by-wire aircraft such as the B777.

“If you had an intention to try and land on the water to try and minimise damage, you’d want to do it with power and hydraulics so you could extend flaps and slats,’’ he says. “So it makes it much more likely that the aircraft was in an uncontrolled state if we find that the main flap was not extended.’’

The ATSB executive says investigators have a sense of what they will find but need to do the work to validate their suspicions. “The observations at this point are pretty preliminary, we’ve only had it in bits for a few days,’’ he says. “So we’ll let out failure analysts do what they do without any pre-conceived notions about an outcome, we’ll just see what the science tells us basically.’’

Debris continues to be found and some already points to the violent break-up of the plane. A small section of interior panel from the front of the aircraft was found in Mozambique and Foley believes a part found by debris hunter Blaine Gibson is likely to be the back of an economy class seat “that’s quite a way back in the fuselage’’.

Other pieces of wreckage determined to be almost certainly from MH370 include a segment of engine cowling found on a beach in South Africa two pieces of debris found in Mozambique, including one thought to be part of the horizontal stabiliser.

A new claim centres on the flaperon – a moveable wing surface that combines the attributes of an aileron and a flap – found on La Reunion Island near Mauritius a year ago and confirmed by French officials as coming from the plane.

The former chief investigator with the Canadian Transportation Safety Board, Larry Vance, told current affairs program 60 minutes that damage to the wing panel’s trailing edge indicated it was extended and subjected to water erosion as part of a controlled ditching.

Mr Vance contended that the force of the water hitting the flaperon was the only thing that could have caused the damage and the only way the panel could be extended was if the pilot selected it.

A similar argument applied to the flap. “Someone was flying the airplane at the end of its flight,’’ Mr Vance told the program.

Not everyone is convinced:  an experienced former B777 captain told AirlineRatings he believed the same damage could occur with the flaperon retracted.

Foley also urges caution, pointing out that the right flaperon is driven by the central hydraulic system powered by the windmill-like RAT. Even with the autopilot disconnected, he says, the flaperon would still be working.

“The roll stabilisation mode that the flaperon is operating on is one of those lower level autonomous things that the RAT, from all accounts, can power,’’ he says. “So it gets commands still, even with the RAT, and it gets hydraulics from the RAT.

“It could well have been working at the end of flight stabilising the final roll of the aircraft so we can’t conclude with certainty from the flaperon and the failure of its trailing edge that was someone in control.’’

Proponents of the controlled ditching have also been excited by a recent leak outlining details of a flight path into the Southern Indian Ocean plotted on a flight simulator program owned by Zaharie.

That Zaharie plotted a course into the Southern Indian Ocean on his home computer and later wiped the data was reported two years ago along with the fact the simulators hard drives had been sent to the FBI for analysis. However, those reports did not have the data points revealed recently by New York magazine and showing a course similar to the one taken by MH370.

While Australian officials agree this could show planning and intent, they note it does not tell them the location of the aircraft or what happened at the end of the flight. They also say the Boeing did not have enough fuel to reach the end of Zaharie’s flight path.

The discovery of the wreckage has also opened the way for various drift studies using ocean currents to attempt to determine the origin of the pieces.
A problem with the drift studies is that they are blunt instruments in terms of accuracy and a great deal of time elapsed between the plane’s disappearance and discovery of the wreckage.

Australia’s Commonwealth Scientific and Industrial Research Organisation modelled likely source points for the wreckage found near Reunion Island and also looked at where debris drifting away from three separate areas on the seventh arc might end up after 500 days.

The modelling produced results that were consistent with the search area, although it found a slightly lower probability that the source points were in the southern part of the priority zone.

The CSIRO is about to embark on a new study using six model flaperons built by the ATSB to compare the drift characteristics of the wing part with “drifter buoys’’ used to record ocean temperature and salinity information.

“The flaperon remains the most interesting of all the pieces that washed up because it was a first arriver and it arrived La Reunion after about 500 days,’’ says Foley. “It was pretty clear from the extent of the marine growth on it that it hadn’t sat on a beach somewhere for any period of time.’’

The drifter buoys are designed to drift with surface currents and are initially fitted with drogues, although these eventually detach. CSIRO scientist David Griffin has plotted the course of 66 drifter buoys without drogues that passed through the search area, or close to it, over the past 30 years.

CSIRO is hoping to use the historical information on the way wind, wave and currents affected the buoys to construct a model that will give them an indication of the general origin of the flaperon.

Foley says the CSIRO has updated its surface current models for the Indian Ocean for 2014-16 so the searchers are dealing with the best ocean current model for the area possible.

The CSIRO is not the only body doing modelling and recent study by Euro-Mediterranean Center on Climate Change looked at five of the pieces of debris found so far and attempted to reverse model how they had travelled and their origin.The study’s simulation overlapped with the northern half of the search area.

“However, our simulation shows that the debris could also have originated up to around 500 km further to the north,” said centre researcher and lead author of the study, Eric Jansen.  “If nothing is found in the current search area, it may be worth extending the search in this direction.”

Foley says there has been a big variation in the results from reverse modelling. “The forward modelling is far more accurate, that narrows it down quite a bit but once again it’s 20 degrees of latitude,’’ he says “What we hope to do with Griffin’s work is narrow it down to about five degrees of latitude.’’

Three ships remain available to search the last 10,000 sq. kms of the priority search area. Once that is done, the governments of Malaysia, Australia and China have agreed to suspend the search if no further credible evidence emerges. There are strong arguments the search should continue in some form but it seems the governments are reluctant to commit more resources after spending at least $180m — $100m from Malaysia, $20m from China and $60m from Australia.

There is still hope among the searchers that the plane will be found in the few months it will take to sweep the remaining section of the search zone.
The ATSB intends to look more closely at some 20 points of interest and has sonar experts trawling through the data from areas already searched in an effort to make sure nothing has been missed.

Attempts are also underway to analyse marine growths on the parts to better understand where they may have come from and which areas they travelled through.

New ATSB chief commissioner Greg Hood observes that the investigation is still within the timeline it took to find the missing wreckage of an Air France flight 447, an Airbus A330 which plunged into the Atlantic Ocean in 2009. That aircraft was still transmitting information until the last moment, debris was found within days and searchers had good idea of the impact site.

He says The ATSB wants to be transparent and is looking at another printed report and releasing the Boeing modelling.
“We’ve got nothing to hide, we’re doing our best,’’ he says.

What is certain is that even if the search vessels are withdrawn later this year, somebody, somewhere will be working on what has become modern-day aviation’s greatest puzzle.