Wednesday, November 11, 2009

"The pilot is the weak link"

According to a new book by William Langewiesche, reviewed here by the NYT, the true hero of the incident earlier this year when a US Airways plane was landed safely on the Hudson river after losing both its engines was not the pilot, Chesley "Sully" Sullenberger. It was Bernard Ziegler, a Frenchman who perfected the "fly-by-wire" technology used by Airbus. Langewiesche asserts that the Airbus was nearly capable of landing itself, even after losing its engines, and while Sullenberger made the right choice to land in the river, the landing itself required only moderate skill and any decent pilot could have done it.

It is an interesting claim, because the other headline-making air accident this year was the loss of Air France flight 447 on June 1, en route from Rio de Janeiro to Paris. On that occasion, there were several suggestions that fly-by-wire, and the lack of manual pilot overrides on Airbus aircraft (in contrast to Boeing), were responsible.

Langewiesche further claims, credibly, that being an airline pilot is such an incredibly monotonous job that the best and brightest do not want to do it today. (Some pilots may find unusual ways to alleviate that boredom.) Michael Moore says that pilots in the US are so poorly paid today that many of them work second jobs. No wonder so many foreign pilots are now working for Indian carriers, who not only continue to pay well, but Air India pays expat pilots more than Indians. In fact, their annual bonuses (up to $15,000) are comparable to the total annual pay ($17,000) of some pilots for major US airlines, if Moore is correct.

Air India, of course, is in a financial crisis, as -- to a lesser extent -- are all Indian carriers; so such generous pay may not last long. But good pilots are still required, even if Airbus planes require little skill to fly.

Unless Airbus invents a pilotless aircraft (Boeing doesn't seem very impressed with fly-by-wire), I think interesting times are ahead for the airline industry, and for passengers.

11 comments:

Rahul Basu said...

'Interesting' is an interesting use of the word :-) You mean the passengers not knowing whether it's the pilot or the airplane flying the aircraft.

Rahul Siddharthan said...

I am trying to track down a very heated article after the Air France crash, blaming European control-freakishness for Airbus's fly-by-wire which it said caused the crash, and claiming that American individualism meant Boeing would never go that route. But I can't find it now. The Boeing 777 does have fly-by-wire, apparently, but allows pilot overrides for all functions, which the Airbus does not. See also this older article.

Rahul Basu said...

BTW this article mentions the first Airbus 320 crash as the one in France. But there was one soon after in India - one of the last major crashes in India - when the newly acquired A320 by Indian Airlines hit the wall or some obstruction at the end of the Bangalore airport. I remember reading at that time this debate between the pilot trying to do something (based on his previous experience with older aircraft) and the aircraft having a mind of its own. (Interestingly I also recall some physicist friend of mine comparing it to TeX -- you type in the paper and TeX decides how to format it correctly and it is non trivial (other than using verbatim) to override the way TeX believes it is to be formatted and it fact is not recommended).

Rahul Siddharthan said...

Yes, I remember all the IA A-320s being grounded for a while after that. Perhaps insufficient training was the problem, since the planes have proved quite safe since then...

Anonymous said...

Rahul,

There seems to be some confusion over fly-by-wire technology. In the simplest aircraft the pilot moves the control surfaces by moving levers that are connected to the control surfaces with a network of cables and pulleys. More advanced/complex aircraft connect the control levers to the surfaces through hydraulic lines, pumps and pistons. In a fly-by-wire system, the control levers manipulate the surfaces through an electric circuit of relays, motors, sensors and actuators. An onboard control system, flight computer if you may call it, ensures that the pilot doesn't by accident or deliberately effect maneuvers that will crash the plane. Hydraulically controlled aircraft are supposed to be more difficult to maintain, what with the leakage, and power being wasted in runing high pressure pumps. FBW controls are lighter, are easier to maintain, and perform consistently across a range of temperatures, etc., IIANM Boeing began working on FBW about a few years after Airbus and produced its first FBW with the 777. Pre-777 Boeings do have some sort of intermediated control. On paper there could be an FBW with almost no flight envelope protection, although I am not sure, why an airframe designer would let all that data generated by the sensor network go waste. Cars now feature drive-by-wire systes - elctrically operated power steering, brakes, and throttle, and even suspension.

The A320 IINM is the first commercial airframe to do away with the control column and wheel. With the more advanced on-board computer networks, a number of flight control laws (or control modes) can be used to restrict catastrophic pilot actions. The A320 does have some sort of manual override as far as I can see from Wikipedia. And the Boeing 777 too implements similar protections. Here's Boeing document on the 777 control system http://www.davi.ws/avionics/TheAvionicsHandbook_Cap_11.pdf. The 777 control system implements protection by limiting, i.e., increases feedback should the pilot try to exceed flight envelope limitations. While this means that the pilot can violate angle-of-attack or bank angle norms, the feedback on the control lever would e so great that it would warn the pilot, or scare the wazoo out! Boeing calls this an electronic system designed to work as a manual system. The reasons for this compromise aren't clear, but there is little that indicates, that this stems from the famed American virtues of rugged individualism, free enterprise, democracy, choice, gun rights and what have you! Compared to my inelegant description of FBW, is the Boeing manual's crisp description.

Here's a video on an A320 test flight - http://www.youtube.com/watch?v=WCc-R4xXZPU&feature=player_embedded

Rahul Basu said...

incredible video ! So are some of the others on the same you tube page.

Rahul Siddharthan said...

jbeck - thanks for the very informative comment. So if I understand right, there are two aspects to fly-by-wire. One is replacing hydraulics by electronics, which is similar to what has been happening in cars, except much more sophisticated (since a plane is much more sophisticated). The other is computer control, and taking control away from the pilot. Boeing is beginning to go down the first route, but not so much the second; and that, it seems to me, is indeed due to philosophical differences (Boeing thinks that, ultimately, pilots should have complete control; Airbus doesn't). Whether that difference comes from American versus European culture, I don't know... The 2000 article I linked above sounds pretty even-handed, and says pilots in the US are split evenly on which system is better, and the debate will continue for the next "five to ten years"; it's now nine years and counting...

Rahul Siddharthan said...

ps - the idea of feedback on the control lever as a warning system is interesting, but it seems a bit complicated to me: why not just have a flashing warning on the screen, "You are trying to do something dangerous! Are you sure?" and if the pilot says yes, let him do what he likes, and if no, the computer resumes full control... From a well-trained pilot's point of view, the two should be equivalent, but from the implementation point of view, surely the latter would have been easier?

Anonymous said...

The salaries are suspiciously low and maybe entry level pilots at smaller companies draw $17k.

As per
http://www.pea.com/imd/airline-pilot-salary.asp

the lowest staring salary among the listed cos is $28,536
That said, pilots have been seeing decreases in pay. There was a recent article in NYT about a young pilot who has seen his pay drop by 50% after being re designated as first officer from captain.
http://www.nytimes.com/2009/10/14/business/economy/14income.html?pagewanted=all

Experienced pilots get upwards of $100k.
My guess is expat pilots being employed by AI/IA are experienced who would be making pretty much the same here or elsewhere or slightly more. Not substantially more.

Sridhar

Rahul Siddharthan said...

anonymous - the pea.com data is from 2004. According to Moore, there have been severe cutbacks since then (as your NYT article also says). "Sully" Sullenberger was also quoted saying no airline pilot wants his children to join the industry, if I remember right.

Anonymous said...

It's taking me some time to get over the unlikeliness of a business type like me talking about an engineering topic like controls, with a group of theoretical physicists from MatScience! Airbus has done many things to overtake Boeing over the last few decades. Among them are its innovations in aircraft control systems. There's a good reason why experts talk of aircraft control philosophy and laws. Because the control philosophy used to develop the control system and its attendant laws, is an axiomatic idea of the way of the pilot. European control philosophy it would seem has a less heroic and rational picture of the pilot. There are other problems that held back Boeing during the last 25 years or so, its legacy investment in technology development during the 60s, compared to Airbus's start with a clean slate in the 70s. And of course the notorious US NIH (not invented here) obsession - if it is NIH, it's not good.