UPDATE: December 4, 2015
IT TOOK ABOUT A YEAR, but the findings are out regarding the crash of AirAsia flight 8501, the Airbus A320 that went down during a flight from Surabaya to Singapore last December, killing everybody on board.
The crash had nothing to do with the weather, as had been suspected in the immediate aftermath (see earlier posts below). What happened, essentially, is that a combination of mechanical malfunction and human error led to the pilots losing control of the plane.
I was hoping the pilots would be vindicated, but the report doesn’t paint a pretty picture.
The trouble started with a failure of the plane’s rudder limiter, a system that helps regulate control inputs at high speeds and high altitudes. As part of their troubleshooting, the pilots pulled and then re-engaged a circuit breaker that supplies power to this system.* This had the unintended consequence of disconnecting the jet’s autopilot and autothrottles.
In most circumstances, dealing with these automation disconnects should have been straightforward: take control and fly the plane by hand. On this occasion, though, there was a wrinkle: because of reasons tied into the limiter failure, when the autopilot quit, the plane’s rudder deflected sharply in one direction. This resulted in a very hard turn to the left, at one point exceeding 50 degrees of bank (for perspective, a normal turn seldom exceeds 20 degrees of bank, and is usually around ten or fifteen). This would have been startling, but nothing they couldn’t recover from in just a few seconds.
What happened next is the critical part — and the part that I don’t get. We have the captain, who’d been troubleshooting the malfunction, instructing the first officer to “pull down” on his control stick, a request to which the first officer enthusiastically complied — with so much vigor that he sent the jet into a rapid, 5000-foot climb and a high-altitude stall.
As the report sums it up, “The manual handling resulted in the aircraft entering a prolonged stall and upset condition, which was beyond the capability of the crew to recover.” Why the captain asked for the “pull down” maneuver I don’t understand. But the bigger mystery is, why did the copilot make such a drastic input? And why was he unable to recognize and deal with an impending aerodynamic stall?
I just don’t know. And this is shaping up not unlike the Air France flight 447 disaster — the Airbus A330 that stalled and crashed off the coast of Brazil in similarly murky circumstances in 2009. One can read and re-read every analysis and black-box transcription of that accident, yet there are aspects of it — the crew failing to recognize and respond correctly to an ordinary aerodynamic stall — that are forever perplexing.
And does this in some way tie into the design of Airbus aircraft, or into the design of modern jetliners in general? As a corollary, are the skill sets that modern airline pilots have come to rely on inadequate in some fundamental way? Have pilots “forgotten how to fly?” Maybe it’s a bit of both, but it’s simply not right when minor malfunctions lead to a deadly loss of situational awareness.
For instance — if we can do this without turning it into a Boeing-versus-Airbus thing — could somebody explain to me the rationale behind Airbus’s independent side-stick control concept? In traditional airplanes, the control columns and wheels (the yokes, if you must, to use a term I’ve never liked) are physically linked, and they move in unison. When the pilot on the left turns, pushes, or pulls, the controls on the right turn, push, or pull as well. And vice-versa. This makes pretty obvious sense, allowing either pilot to understand exactly which inputs the other pilot is making. On Airbus models this doesn’t happen. When either pilot moves his or her side-stick, the other stick remains in place. Only when contradictory inputs are made is an alarm triggered. In rare circumstances this can lead — as was seen in the Air France tragedy, and possibly in the AirAsia accident as well — to a situation of dangerous confusion. I’m told there are valid engineering and human factors reasons for this design. I’ve never piloted an Airbus, and I’m loath to criticize without learning more, but thus far I fail to comprehend what those reason are.
* The resetting of cockpit circuit breakers is common practice, by the way. It is not, however, at the pilots’ whim. Protocols vary a bit carrier to carrier, but generally, resets are only to be be performed in accordance with written procedures (a checklist for example) and/or under the guidance of maintenance personnel.
Headquartered in Kuala-Lumpur, Malaysia, AirAsia is the largest low-fare airline in Asia and one of the biggest in the world. It operates about 70 aircraft, all of them A320s, on routes around Indonesia, Thailand, Malaysia and beyond. (AirAsia X is the airline’s long-haul affiliate, and operates the larger A330.) I flew AirAsia between Bangkok and Phuket a few years ago. For what it’s worth, except for a delay on the outbound leg, its operation struck me as no more or less professional than that of any other major airline.
Asia is now the world’s biggest and busiest air travel market, having surpassed both North America and Europe. Low-cost competitors like AirAsia are gaining ground throughout the region.
UPDATE: January 6, 2015
AS OF TODAY, Indonesian authorities are blaming the crash on “ice.” What that might mean, exactly, I don’t know. I’m unaware of any evidence linking the accident to icing, and it’s irresponsible, if you ask me, to be publicly announcing such sure-sounding theories, particularly with a media that is off and running with the slightest new speculative thread.
Indeed, what the media needs right now is a proverbial chill pill. It’s much too soon to be trying to nail down a cause. A properly executed air crash investigation takes months, usually. I was on CNN earlier today, and to the interviewer I said the following (I’m paraphrasing for clarity):
The media’s fixation with this accident is not helpful. With every crash now, it seems we get into this cycle of marathon coverage. I don’t know if it’s because, or in spite of the fact that major plane accidents have become so rare that we’re lavishing so much attention on them. Thirty years ago, when we were dealing with five, ten, even a dozen serious accidents every year, we didn’t obsess like this. Of course, in those days there weren’t multiple 24/7 news sources starving for attention across multiple platforms. And that, more than anything, is what all this focus is about: feeding the news machine.
You won’t see that part of the interview, however, because they edited it out.
UPDATE: January 2, 2015
LET’S SEE if we can square away some of this “tracking” business that has been getting so much attention.
Christopher Drew, in a December 29th op-ed in the New York Times, said that “Airlines use satellites to provide Internet connections for passengers, yet they still do not stream data in real time about a plane’s location and condition.” Two days later, in a similar Times op-ed written by the editors, it was stated that airplane location is updated only in fifteen minute increments.
Neither of these things is true, usually. It depends where the flight is operating, what equipment is on board, and which air traffic control (ATC) facility the crew is working with. As a general rule, flights are constantly tracked and monitored. By regulation a flight must always be in contact, one way or the other, with both air traffic control and company dispatchers on the ground. This is true in domestic airspace, and over the remotest points of the ocean as well.
In the busiest airspace, such as over the continental U.S. and Europe (and many other regions), planes are generally in radar and VHF radio contact, which makes tracking a cinch. ATC and airline dispatchers can easily monitor a jet’s location, altitude and speed (plus other parameters, depending). The same is true even in some oceanic airspace, such as over the North Atlantic, where cockpit equipment such as CPDLC and SATCOM datalink allow more or less real-time monitoring of a flight’s progress. In addition to basic position data, newer aircraft can transmit data about engine performance and the mechanical status of certain onboard systems.
In some areas of the world, however, position reports are sent only intermittently, at designated waypoints rather than continuously. This is the “tracking gap” that the media has been so fixated on ever since the disappearance of Malaysia Airlines flight 370. There is room for improvement here, I feel, particularly for long-haul aircraft that operate routinely in non-radar airspace. Planes could and perhaps should be equipped with a relatively simple, inexpensive, and fail-safe technology that allows continuous location tracking, no matter where.
It was a little startling for the Times to begin with a pair of premises that are only partially true. It’s easy to extrapolate: if the writers don’t know what they’re talking about with respect to aviation, should we trust them when it comes to law, politics, healthcare and so forth? Shouldn’t there be a vetting or review prior to the publication of pieces covering specialized and esoteric topics?
I notice that the December 31st op-ed includes a quote from an “aviation consultant.” Not that he said anything wrong or stupid, but I held my breath. This is one of my ongoing pet peeves: when airplanes are the subject, the media loves to cite aviation academics — aerospace researchers, aeronautics professors, etc. As I’ve noted before, these people are bright and their work is important, but they tend to have limited knowledge about the day-to-day operation of commercial planes.
Another question that keeps coming up is why the various black box data — the data recorded by the cockpit voice recorder (CVR) and flight data recorder (FDR) — can’t also be monitored via satellite, radio or wi-fi, in real time, rather than stored away on a piece of aircraft hardware. In other words, we could be constantly aware not only of a plane’s position, but any malfunctions and mechanical problems it might be having.
The main reason why is because it would take immense mounts of bandwidth, multiplied by the thousands of airplanes in the air at any one time, to upload all of the hundreds of parameters monitored by the FDR and CVR. And for what practical purpose, exactly? For the one airplane every 25 years or so that is temporarily missing, out of the 40,000 or so commercial flights that operate every day? Such a thing is certainly possible, but it would be technologically challenging and highly expensive. Is it really needed, in practical terms?
This issue comes up all the time. To me, it’s symptomatic of a culture in which people are accustomed to instant explanations and instant access to everything. “Why can’t we have the answers right now!”
December 31, 2014
FIRST THINGS FIRST, we need to trundle out the boring but critical post-crash disclaimer: It is a bad idea to speculate too broadly on the how-and-why so soon after an air disaster. Almost always the initial hunches and theories end up totally off-base or at best incomplete. We live in an age when people want and expect instant answers, but that just isn’t possible with plane crashes. It often takes months or even years before a cause is nailed down. In some cases we never learn for sure what happened.
That said, a seeming red flag in Sunday’s AirAsia crash is the weather. Could the Airbus A320, flying from the busy Indonesian city of Surabaya to Singapore, have wandered inadvertently into a violent thunderstorm and suffered some kind of catastrophic malfunction or structural failure? It’s possible.
I’ll point out that flying into thunderstorms is about the biggest no-no in all of commercial aviation. The crew had asked for a weather-related altitude change shortly before the disappearance, a request that was denied by air traffic control — presumably because of traffic constraints. This isn’t terribly unusual; pilots ask for altitude changes and route deviations all the time, and not always are they granted. However, that does not mean the AirAsia crew had no choice but to plow headlong into a storm. Worst-case, the crew always reserves the right to do what it needs to do, with or without permission. I cannot imagine the pilots willingly flew into what, on the radar screen, would have been a bright red splotch of potentially dangerous airspace. Perhaps a patch of weather that the pilots presumed would be manageable turned out to be otherwise? We don’t know.
Some are drawing comparisons between this incident and the 2009 Air France tragedy. They occurred under somewhat similar circumstances, and the media is eager to link these recent incidents together and wring some scary significance out of them. Some commentators have noted, for instance, that both planes were built by Airbus. I understand the temptation here, but this is extremely premature, and it’s unlikely that the aircraft model played a significant role. Remember that basically half of all the commercial jetliners in the sky are Airbus models.
An even bigger red herring is the fact that the pilots made no distress call. Several news outlets have brought this up. Effectively it means nothing. Communicating with air traffic control is pretty far down the task hierarchy when dealing with an emergency. The pilots’ priority is to control the airplane and deal with whatever malfunction or urgency is at hand. Talking to ATC comes later, if it’s practical.
Whatever caused the crash of flight 8501, the year appears to be closing on a tragic note. That’s a shame, seeing that 2013 was the safest year in the history of modern commercial aviation. Not to sound flip, but we can’t expect every year to be the safest, and it’s important to look at the broader context. This year will be something of a correction, but over the past two decades the rate of fatal accidents, per miles flown, has been steadily falling. The International Civil Aviation Organization (ICAO) reports that for every million flights, the chance of a crash is one-sixth what it was in 1980, even with more than double the number of planes in the air.
Whenever people bring up the less-than-stellar accident record for 2014, I remind them of how bad things used to be. In 1985, 27 — twenty seven! — serious aviation accidents killed almost 2,500 people. That included the JAL crash outside Tokyo with 520 fatalities; the Arrow Air disaster in Newfoundland that killed 240 American servicemen, and the Air-India bombing over the North Atlantic with 329 dead. Two of history’s ten worst disasters happened within two months of each other! That was an unusually bad year for any era, but throughout the ’60s, ’70s and ’80s, multiple major accidents were the annual norm. Today, large-scale air disasters are much fewer and farther between. You almost wouldn’t know it, of course, switching on the TV: the media’s fixation and round-the-clock coverage of what, in times past, would have been only short-lived stories (or in some cases complete non-stories), messes with our perspective and gives many people the idea that air travel is a lot more dangerous than it actually is.