Tuesday, July 21, 2009

Airlines And Moral Systems Failure

Soon after returning home from my last trip to Southern California, I got into a short conversation with a married couple who are my next door neighbors about my trip. They too had recently traveled to So. Cal., not for work, but to visit relatives. Whereas they flew there and back, I drove. I think they have come to the conclusion that I'm a slightly peculiar character, so they saw my choice to drive as just one more proof of oddity. When they asked if I would be returning anytime soon, I told them that I wouldn't be going back for a couple of months, and that when I returned, it would only be to visit family. One of them asked, “Are you going to fly this time?” “No way,” I said. “Planes crash!” “So do cars,” they both chuckled.

But in all seriousness, I have given up flying for the time being. It's not because of an “irrational fear of flying.” Rather, as I have said before, it's because during the last oil price super-spike, we all witnessed the death of several airlines and the consolidation and extreme cost-cutting measures of several of the survivors. I also know that the surviving airlines are still being squeezed by rising oil prices and falling revenues caused by the general economic collapse. This is hindering their ability to maximize profits. Knowing how big business operates nowadays, I expect that their cutbacks have extended far beyond such visible things as carry-on luggage charges and elimination of in-flight snacks, and that they have been cutting back on things essential to keeping their planes safe in the sky. I therefore expect that as our economic collapse progresses and oil prices go through further spikes, there will be a significant increase of in-flight safety incidents (including crashes).

Is there any evidence to support such an expectation? I haven't done a rigorous statistical categorization and analysis of airline in-flight incidents over the last few years. Such an analysis, while valuable, would take a significant amount of effort, and I don't have the time right now to undertake such an effort. But it is interesting to note that over the last month and a half, during the time when I drove to Southern California twice, there have been at least five crashes of planes used either for passengers or cargo.

One particular crash that caught my attention is the loss of Air France Flight 447, an Airbus A330 that fell out of the sky into the Atlantic Ocean on 1 June 2009. According to initial reports, the plane was in international airspace under the watch of Brazilian air traffic control when it flew into a band of thunderstorm activity. As it entered the storm zone, it experienced a rapidly cascading failure of its highly complex, computer-controlled avionics, leaving the pilots with progressively less information about flight attitude and speed, and less ability to control the plane. There came a point when something critical happened, and the plane fell out of the sky.

What's interesting is how independent investigators drew certain initial conclusions from the crash data, how the French corporation Airbus Industrie responded to these conclusions, the French oversight of the debris recovery operation and their inability to find the plane's “black boxes,” and the preliminary conclusions of the French Government's BEA (Bureau d'Enquêtes et d'Analyses pour la Sécurité de l'Aviation Civile), an agency similar to the American FAA.

Initially, the conclusion of many independent investigators was that the plane broke up in midair. The injuries seen on recovered bodies and the fact that many of those bodies were missing clothing, supported this conclusion, as well as the fact that the crash debris field was several miles long and several miles wide. Airbus spokesmen then began to blame the pilots for flying the plane in a way that exceeded its design limitations, secondarily blaming flight sensors for giving the pilots faulty speed and attitude information.

But this provoked further questions about the safety of the heavy reliance of Airbus on composite-fiber components in their aircraft, in places like wings, rudders and other control surfaces. These suspicions were amplified when the nearly intact tail of the plane was found near the extreme end of the debris field, suggesting that it had sheared from the aircraft while in flight. Such incidents had happened before with Airbus aircraft.

The voicing of these suspicions prompted many vehement assurances from Airbus that their airplanes are well-designed, and perfectly safe to fly. In this Airbus was joined by an unlikely ally, namely Boeing, who are in the process of producing a next-generation plane, the 787 Dreamliner, that will be made up of over 50 percent composites.

Why are composites so important to passenger jet manufacturers right now? It boils down to fuel prices. Those who build the lightest jetliners capture the biggest share of the market, because the person who buys and flies those jetliners can carry the most cargo or the largest number of people for the lowest operational cost.

But aircraft manufacturers have reached a point where their craft are so lightly built that they must be outfitted with extremely complex computer-controlled avionics, to prevent pilots from overstressing their aircraft while in flight. Safety margins have been shaved to the minimum, since generous margins would add weight and operational cost.

A word about composites is in order. It is well known that composites tend to delaminate and develop voids when subjected to the cyclic loading and extreme temperature variations experienced by passenger jets. While the military uses composites in many of its most advanced jet aircraft, it is also true that the military subjects its planes to very rigorous maintenance and inspection regimens. This is not true of planes maintained by airlines, whose manufacturers have convinced the operators of these planes that only visual inspection of components is necessary, or at most, a “tap test.” Hardly any airline does more extensive testing, such as ultrasound scans.

In light of these issues with composites, the outcome of the crash investigation by the BEA is quite interesting. In a report issued on 2 July 2009, the BEA concluded that Flight 447 was brought down by undetermined factors, and that it hit the ocean intact, without breaking up first in midair. They also stated that it was at least six hours from Flight 447's last known transmission before an emergency was declared. This directly contradicts the Brazilian government's assertion that it was less than half an hour after the flight's last transmission before Brazilian air traffic control issued an alert. It is also interesting to note how all the major media outlets have regurgitated the BEA report without questioning the contradictions between its conclusions and the earlier conclusions of investigators not connected to Airbus or the French government.

I believe that Flight 447 and the handling of its aftermath is an indicator of the sorts of things we will see as high-value systems and providers of high-value services are stressed and squeezed by oil depletion and economic collapse, and as these providers struggle to maintain profits. Meanwhile, here are some links for you to enjoy (or maybe not, if you have to fly somewhere anytime soon).

For Further Reading...


Stormchild said...

Well said...

I would have to add my own concerns, regarding over-reliance on computerized systems.

I no longer fly at all, but when I was still required to fly on business I did my best to avoid ever getting on an Airbus.

When this product was first introduced, there were a shocking number of crashes that, on investigation, appeared to be caused by the computerized control system overriding pilot commands. Planes were flying into hillsides and mountains while the pilots were doing their frantic, doomed best to get more altitude, and the idiot IT system was stubbornly refusing to let them.

Washington DC residents were recently treated to another sample of this, on the ground, in the horrific Metro subway crash that killed at least 9 people, including the Commander of the DC National Guard... and his wife. One train suddenly picked up speed and rammed into another from behind.

Blame-the-operator-of-the-ramming-train was the first, cowardly, predictable response. She was texting. She was new. It had to be her fault, just because...

NTSB investigations quickly revealed that the operator was doing everything humanly possible to stop her train. But, again, it was being run by the IT system and she could not override it.

NTSB investigations have also now revealed that it was - again - the IT system, essentially, that killed her and 8 of her passengers. It was failing to detect trains at multiple locations in the Metro track web, and the train she rammed 'vanished' from the sensor system. Whereupon the computerized control system picked up her train and smashed it, full speed, into the train it didn't see.

What's more, this 'loss of trains' from the sensing system has been a problem for at least two years; and Metro KNEW about it. Did they do what could be done to permit operator override in danger situations? Pretty obviously not.

Even more disturbing is the fact that Metro has now, finally, contracted to develop a system that will actually produce alerts when such detection failures occur, and the press release about this calmly states that NO SUCH SYSTEMS CURRENTLY EXIST.

I'm sorry, TH, I know you work in the IT field and I also know you are a TRUE engineer, which means you try your best not only to anticipate failures, but to build in backups and failsafes. If every engineer and programmer was a TH, I'd hop on any plane, train, or automobile anytime. But they're not.

Any program is only as good as its programmer. Any system is only as good as its engineers. There are too many corner-cutters in the design area, too many corner-cutters supervising them, too many bean counters putting profits above anything else.

After all, it's nobody THEY know getting killed on these things. THEY fly Learjets and Gulfstreams...

... which brings me to the sad fate of Payne Stewart, and all aboard his private jet, en route to Palm Springs in1999 ...

but I'll let the link above retell that story. My point is that we live in an economic system that rests, ultimately, on human sacrifice, and almost none of us realize either this fact or its implications.

TH in SoC said...

Thank you, Stormchild. I am thinking of making your comment into a complete post, as I think it illustrates both the diminishing returns of system complexity and the dangers of trying to lower human capital costs by excessive system automation. You are also quite right about how our present economic system sacrifices human life for the sake of profit.

I think a growing number of people are giving up flying...

Stormchild said...

Thanks TH. I have a very minor blog called Potemkin's Office and have been posting there very sporadically about management myths and things like this. I am thinking of putting my comment up there, expanded a bit, and linking back to your post here. Feel free to use anything I have said here, as you see fit, with my blessings.

I am no engineer, and certainly no programmer; just an engineer's kid - who got into science for the love of it, and thinks like a scientist pretty much all the time.

I would be very very interested in your thoughts on this subject; you ARE an expert in this area.

By the way, I coined the term 'thanatoeconomics' back in 1997, in several realspace conversations, to describe our economic system, because it truly is built on human sacrifice.

In order to function, our system actually requires a permanent underclass, AND needs a certain number of people to be deprived of their livelihoods at regular intervals. [Consider the obscene fact that a company's stock price goes up when it indulges in mass firings.] This same system places little or no value on preserving human lives; business schools in this country will actually teach you how to determine when liability exceeds profitability - aka, how many people you can afford to kill before it gets too expensive.

I have long intended to write about thanatoeconomic theory [it's not a theory, of course, it's stark reality], but most of my blogging time has been invested in Gale Warnings for the past 3 years, more or less as a prelude. Thanatoeconomics, after all, is the ultimate consequence of an abusive society.

I've seen the term used once recently by a European writer, who coined it independently. Feel free to use it yourself; I can think of no more fitting name for the system in which we find ourselves.