Monday, December 29, 2008

A Newer, Safer NASA?

Regular readers of this blog—all five of you—know that in the past I have had less than kind words for NASA regarding safety issues. The agency has become something of a poster child in engineering ethics circles, mainly because of the Challenger and Columbia space-shuttle disasters. Recently, however, under the leadership of Michael Griffin, NASA has shown signs of getting its act together. It has stuck to long-overdue plans to retire the current fleet of space shuttles by 2010, which by Thursday will be only next year. And it has embarked on ambitious but generally well-considered plans to develop a new way of getting to the moon and beyond: the Constellation program, which includes the new Ares series of partially reusable rockets and space capsules. In these plans, NASA is at least trying to do what's right, which is to get away from the increasingly antiquated and hazardous shuttle system toward newly designed systems that take advantage of thirty years of aerospace progress, while embracing safe tried-and-true technologies.

For example, the personnel-transporting Ares I rocket uses liquid-fueled J-2 engines, the same kind that the Gemini and Apollo programs used. The solid-fuel boosters employed by the space shuttle were a bargain-basement compromise that led directly to the first major shuttle disaster, and it will be good to see them go.

But the best-laid plans of mice and men, at least those employed by the government, are subject to political winds. News reports of a few weeks ago carried a story about a dust-up between Griffin and Lori Garver, a former NASA public relations officer who now heads the Obama transition team in charge of deciding what to do with NASA.

According to Time, the dispute arose when Garver and Griffin met in a NASA library during a book signing. Although details of the encounter vary by source, one issue was the possibility of either canceling the Ares I program altogether or switching to a cheaper alternative using existing rockets such as the Atlas and Centaur, currently used for unmanned space projects. Griffin reportedly questioned Garver's engineering qualifications and said that Atlas and Centaur rockets are not safe enough to use for people.

Of course, it is within the prerogative of a new administration to replace Griffin with someone more cooperative. But if NASA has finally learned something from its past mistakes and is headed in a good direction, it would be a shame if the incoming administration forced it into the same nickel-and-dime mode that led to the twenty-eight-year dangerous compromise called the space shuttle.

The environment that NASA operates in now is very different than the situation three decades ago. In the late 1970s, the only credible competitor in space was the old USSR, which had pretty much thrown in the towel when the U. S. reached the moon in 1969. Today, all kinds of folks are getting into the act: China, Japan, and India all have serious space programs, and Russia, far from having abandoned theirs, will be our only link to the International Space Station from the time the shuttle is retired until we develop something better—assuming there are funds and political support to do so.

The appeal of space flight—especially manned space flight—has always been more emotional than rational, more tied to political prestige than to economic realities. From a strictly business point of view, the market is very thin—there simply aren't that many super-rich people willing to pay $20 million for a joyride in space. And while there is money to be made from orbiting satellites, that end of the business has become truly a business, in which the cheapest rocket that will do the job is used. For the foreseeable future, flying people outside the atmosphere is something that will always be a money-loser. So the agencies charged with doing it have to justify their existence on grounds other than profit.

For NASA, this means two things: science and the romantic appeal of space exploration. As far as science goes, the Hubble Space Telescope has proved very fruitful, but note that the only time we have to send people up to it is when it breaks down. With advances in autonomous robotic instruments, even exploration of other planets can be done much more economically with unmanned probes. So sending people into space doesn't really further science, except to add to our knowledge of what happens to people when they spend long periods of time in space. (The short answer to that is, nothing much good.)

So that leaves romance and a kind of quasi-religious feeling as the real basis for manned space flight. Griffin is reportedly a believer in the idea that mankind's long-term destiny is to colonize space, whether other planets in our solar system or planets in other solar systems. I have had some minor dealings with people who think this way, and it really amounts to a kind of religion. If you don't believe in the supernatural, and you think we are well on the way to trashing the only planet we can live on, that leaves space exploration as the only hope for immortality of the human species. After all, the sun is going to run out of gas here in a few billion more years, and then what?

I am generally well disposed toward idealism, if it is directed at worthy goals. And I'm all in favor of the scientific aspects of space exploration. But sending people into space costs a lot more than unmanned projects, and (needless to say) are more dangerous to those participating as well. In a democracy, public expenditures have to be justified by the economic or political good that they can achieve. These political goods include maintaining what prestige we have in the community of nations, satisfying the desires of those who see space as the final frontier, and making heroes, without whom few nations make it for long. It is up to the Obama administration and the new Congress to decide whether NASA's plans are worth the cost. Whatever happens, I hope that we end up with something to replace the shuttles that is orders of magnitude safer and will serve the public in the best way possible—whatever that is.

Sources: Reports of the Griffin-Garver controversy I used can be found at the Time Magazine website http://www.time.com/time/nation/article/0,8599,1866045,00.html
and the Orlando Sentinel website http://blogs.orlandosentinel.com/news_space_thewritestuff/2008/12/nasa-has-become.html.

Monday, December 22, 2008

A Riveting Story: The Titanic's 94-Year-Old Mystery Solved

The sinking of the White Star Line's luxury passenger liner Titanic on April 14, 1912 has got to be one of the most famous engineering failures in history. Everybody knows the story: how the ship ran full speed into an iceberg despite warnings relayed by the then-new wireless, and sank less than three hours later with the loss of over 1500 lives. Since the discovery of the wreck in 1985, researchers have been able to recover hundreds of artifacts and subject them to modern forensic analyses. Two of these researchers—metallurgical experts Jennifer Hooper McCarty and Tim Foecke—have written a book about their discoveries. What Really Sank the Titanic clears up a long-standing mystery about the tragedy and points the finger of blame in a surprising direction.

Boards of inquiry held immediately after the disaster obtained enough information from survivors to piece together the following story. At about 11:40 PM, as the Titanic moved at about 22 knots through a near-freezing sea "as smooth as glass," lookouts spotted an iceberg in the path of the ship. The steersman had just begun to turn the bow to port when the berg scraped along the starboard side of the ship, making a long-lasting noise that was described variously as tearing, jarring, or ripping. Although the ship's six watertight compartments were immediately sealed when it was discovered that water was coming in, there were enough holes in different parts of the hull that eventually all six compartments filled up, and the ship sank. The ship's designer, Edward Wilding, said at one inquiry that a long, narrow series of slit-like openings about two hundred feet long and only an inch or so wide would have accounted for the fashion and speed with which the ship foundered. But since steel is much harder than iceberg ice, he could not explain how such openings could have occurred. There the mystery lay at the bottom of the North Atlantic for over eighty years, until salvage expeditions began to bring pieces to the surface.

In a decade-long investigation, McCarty and Foecke, respectively graduate student at Johns Hopkins University and staff member at the Gaithersburg, Maryland office of the National Institute of Standards and Technology, obtained samples of the Titanic's hull, which consisted of large steel plates held together by rivets (electric-arc welding was not to become the standard steel-fabrication method until World War II). McCarty's archival research in England revealed that Harland & Wolff, the ship's Belfast builders, used two kinds of rivets: the more modern machine-formed steel rivets for the central part of the ship, and the old-fashioned hand-formed wrought-iron type for the stern and bow sections, where much of the collision impact probably occurred.

The making and installing of wrought-iron rivets was largely a manual operation. The Titanic needed over three million rivets in all, and this huge demand led flocks of entrepreneurial ironmakers to enter the field. The hand-stirred "puddling" process then used to make wrought iron from ore required strong and highly experienced workers, of which there were not enough in 1912. So it turned out that Harland & Wolff bought wrought iron from a wide variety of suppliers, some of whom were much less experienced than others. McCarty and Foecke have proof of this in the form of long, stringy slag inclusions they found in some of the recovered rivets. These inclusions tended to make wrought iron, an already a less satisfactory material than steel, even weaker.

Why weren't steel rivets used throughout? Besides reasons of cost, steel rivets had to be formed with hydraulic riveters—large U-shaped steel machines upwards of six feet high that had to be laboriously positioned on either side of the plate to be riveted. Then the rivet, shaped much like a blunt round-headed nail, would be heated, inserted into its hole through the two overlapping hull plates to be joined, and squeezed between the jaws of the riveter. This squeezing formed heads on both ends, and as the rivet cooled, the resulting shrinkage provided tension that held the two steel hull plates together in a watertight joint.

At least, that was how it was supposed to work. The problem was there was not enough room in the bow and stern areas to maneuver the hydraulic riveter. So the builders resorted in those areas to the older hand-forming way of riveting, which couldn't use steel rivets because of reasons to do with the different ways steel and iron cool. Wrought iron was more forgiving to the delays and variations involved when a boy tossed a red-hot rivet from a portable stove to the rivet gang, which placed it in its hole and pounded it in by hand.

When the Titanic embarked on her maiden voyage on April 10, 1912, her bow hull plates were held together by wrought-iron rivets. The iron itself had probably never undergone any systematic quality testing, and the only quality tests done on the finished riveting job was a hurried hammer tap by an inspector, who listened to the sound it made. All this inspection could detect was loose rivets, not those made from defective wrought iron.

Then came the iceberg. While ice itself will crumble if forced against solid steel, the typical iceberg was a lot heavier than the Titanic. So in a glancing collision, the iceberg exerted tremendous localized force against only a few hull plates at a time. While even poorly-made rivets can withstand the mainly sideways stress that uniform pressure causes (e. g. hydraulic pressure on a water tank or a ship's hull), some of the forces that the iceberg caused tended to pull the hull plates apart, causing tensile stress. And the researchers found that wrought-iron rivets made of bad iron with lots of slag inclusions pop their heads off much more easily than either steel rivets or wrought-iron rivets made with better material. Significantly, many of the steel plates recovered from the wreckage were missing their rivets altogether. And riveting is not a gracefully-degrading fastening method. Once one rivet in a row pops, the ones next to it get much higher stress and are likely to fail as well, leading to a kind of chain-reaction zipper effect.

That is exactly what McCarty and Foecke say must have happened as the iceberg bounced repeatedly along the side of the ship, popping rivets and opening up long, narrow slots between hull plates—exactly what designer Wilding said in 1912 must have happened, though he couldn't explain exactly how. The researchers also show in detail how a rival theory—one that says the cold Atlantic waters made the plates themselves brittle enough to shatter like glass—is full of holes, so to speak.

So the roots of the Titanic disaster prove to go in several directions: to the heedlessness of the captain who failed to slow down in a known field of icebergs, to the rulemakers who didn't require lifeboats for everybody, and, surprisingly, to little mom-and-pop wrought-iron puddling operations that sprang up all over the United Kingdom in response to increased demand for wrought iron. McCarty and Foecke conclude that if all the ship's rivets had been steel, the ship still might have sustained serious damage, but not so much as to sink it in less than three hours. Even a few hours longer afloat could have given time for nearby ships to arrive and save most or all the passengers. But that was not the way it happened.

Sources: What Really Sank the Titanic was published in 2008 by Citadel Press. I also thank my wife Pamela for her thoughtfulness in this birthday-gift selection.

Monday, December 15, 2008

Explosion in Tyler: More Questions than Answers

Over the weekend I heard a short news item about an explosion and fire at an oil refinery in Tyler, Texas. Here is the fruit of less than an hour's web research on what happened:

Tyler is a town of about 100,000 in East Texas. Among its industrial facilities is a smallish oil refinery (it's the 94th largest in the U. S.) owned by an Israeli company called Delek USA. On Thursday afternoon, Nov. 20 of this year, a part of its saturated gas unit exploded and caught fire. Four employees of the company were hospitalized, and two of them died later of their injuries. About 2,000 gallons of gasoline spilled into a nearby creek during the accident, but didn't catch fire. Five employees have filed suit in a Houston court over injuries suffered in the blast. The U. S. Chemical Safety and Hazard Investigation Board has not yet posted a notice of investigation for this accident.
The plant has been shut down since the explosion and no one is saying when it might start up again.

As oil refinery accidents go, this one is not in the major leagues. The 2005 BP refinery accident in Texas City, Texas that killed 15 people is probably the most recent leader in that regard, if we judge by the number of fatalities. But the death or injury of even one person in any engineered facility is the result of something that shouldn't have happened.

I wish I could present you with a complete story of exactly what went wrong in Tyler that day and how it could have been prevented. But alas, it is not my job to gather such information from primary sources. Investigators from insurance companies or perhaps the Federal government will undertake that arduous and exacting task, equipped with tools and knowledge that such specialists have. In a few months, perhaps, the truth will emerge about what caused the accident. If history is any guide, human error will have turned out to play some role.

There are a lot of good reasons why refineries are such dangerous places. Just handling millions of gallons of highly flammable liquids and gases involves considerable risks on its own, even if you're not doing anything to make matters worse. But that is how refineries operate: you take combustible crude, run it through pipes surrounded by intense flames, squeeze it under tremendous pressure that will use any slight excuse to shoot flammable stuff out into the air where it will spontaneously burn because it's so hot, and then subject it to all sorts of chemical indignities with catalysts, further heating, pressure, toxic acids, and so on.

It would be bad enough if it was simple, but refineries are some of the most complex large systems on Earth: thousands of pipes, pumps, valves, tanks, sensors, actuators, and other stuff, all having to be operated just so or else you're in big trouble fast. The fact that we don't have refinery accidents every day is a testimony to the incredibly disciplined management and training that has to go into good refinery operations. Refinery employees have hard, dangerous jobs, and most of them do their jobs well. But not always, especially if they or their managers get careless or squeezed by cost issues into deferring necessary maintenance or safety practices.

Equipment failure, while not unheard of, is something that can almost always be prevented. The physics and chemistry of how steel fails and how chemicals behave are well enough understood that chemical engineers can predict what's going to happen in almost any given case. The problem is making sure that knowledge is applied at the right time in the right way.

The fact that the Tyler plant is owned by an offshore firm may not have any bearing on the accident. But responsibility is a funny thing: like radio waves, it tends to weaken over long distances. This is not to say that all companies based in the U. S. act more responsibly than any foreign firm—that is silly. But the point is that if nationalism means anything at all, and there is abundant evidence to show that it does, people are going to be more conscientious about protecting the lives and wellbeing of fellow citizens before they will look out for the interests of foreign nationals. It is a natural human tendency, but one that must be fought against when the situation of foreign ownership of plants arises. Years ago, when it was more common to find American firms owning offshore factories, the temptation was to neglect safety for non-American native populations, and tragedies such as the Union Carbide Bhopal disaster in India were the result. Now that ownership of manufacturing facilities seems to have become anathema to U. S. businesses, the shoe is on the other foot. Such refineries and plants that we have here are increasingly owned by foreign firms, whose owners may or may not be as careful to ensure safe working environments as U. S. owners might be.

We will simply have to wait to find out more about what happened in Tyler last November 20. But let's hope that any lessons we can find out will be learned well by everyone who has anything to do with oil refinery safety or engineering.

Sources: I have used information from the following sources: http://www.haaretz.com/hasen/spages/1044426.html, http://www.tylerpaper.com/article/20081120/NEWS08/811200292, and http://oilspot2.dtnenergy.com/e_article001272178.cfm?x=b11,0,w.

Monday, December 08, 2008

Michael Polanyi and the Purity of Science

Everybody agrees that engineering and technology these days depend on science. Modern engineering is inconceivable without the advances made possible by the Scientific Revolution, both in our state of knowledge and in our approach to knowledge itself. One of the twentieth century's most profound thinkers about the scientific way of knowing and its connection to technology was Michael Polanyi (1891-1976).

Born of Jewish parentage in Hungary, he obtained a Ph. D. in chemistry and began a career as a research chemist in Germany. He married a Catholic there and left with his family for England shortly after Hitler came to power in 1933. At the University of Manchester, his interests turned gradually from chemistry to the philosophy of science. When he was invited to give the prestigious Gifford Lectures for 1951-52, he focused on the personal nature of supposedly "objective" scientific knowledge and published his thoughts in his best-known work, Personal Knowledge, in 1958. Fifty years later, his arguments about the purity of science and its relationship to technology are probably worth heeding even more now than when he wrote them.

Personal Knowledge is a deep and wide-ranging book, and all I want to do today is to give you a small sample that touches on the relationship between science and engineering, and how we allow the practical needs of technology to dominate the way we do science at our peril.

If you write a proposal these days to the U. S. National Science Foundation for even the most abstruse and theoretical project, you will have to address two questions in your proposal. The first, appropriately, is "What is the intellectual merit of the proposed activity?" That is, how will this work advance the state of scientific knowledge? This is an entirely reasonable criterion, and one which has been in place since the early days of the Foundation in the 1950s. However, the answer to a second question is now given equal weight in the funding reviews: "What are the broader impacts of the proposed activity?" In an explanatory paragraph, the Foundation expands on this second question thus: "How well does the proposed activity broaden the participation of underrepresented groups (e.g., gender, ethnicity, disability, geographic, etc.)? . . . .What may be the benefits of the proposed activity to society?" Unless an investigator can muster adequate answers to both main questions, the proposal stands practically no chance of getting funded.

Now the NSF distributes taxpayers' money, and the taxpayers have a right to know what they're getting. But contrast that second question with Polanyi's words about why scientists should do science. In describing how science fares in developing countries, he says that "it suffers from a lack of response to its true values. Consequently, the authorities grant insufficient time for research; politics play havoc with appointments; businessmen deflect interest from science by subsidizing only practical projects. . . . Encircled today between the crude utilitarianism of the philistine and the ideological utilitarianism of the modern revolutionary movement, the love of pure science may falter and die. And if this sentiment were lost, the cultivation of pure science would lose the only driving force which can guide it toward the achievement of true scientific value. The opinion is widespread that the cultivation of science would always be continued for the sake of its practical advantages. . . . The scientific method was devised precisely for the purpose of elucidating the nature of things under more carefully controlled conditions and by more rigorous criteria than are present in the situations created by practical problems. These conditions and criteria can be discovered only by taking a purely scientific interest in the matter, which again can exist only in minds educated in the appreciation of scientific value. Such sensibility cannot be switched on at will for purposes alien to its inherent passion [e. g. for writing answers to question about the "broader impact" of research]. No important discovery can be made in science by anyone who does not believe that science is important—indeed, supremely important—in itself."

I wonder how far we in the U. S. have gone in the direction that Polanyi warned about. I see several signs that maybe we've gone quite a distance. Corporations have long since shuttered their basic-research divisions: Bell Laboratories is no more, IBM and Xerox are no longer spending large fractions of their income on basic research, and even before the present economic downturn, any CEO who authorized spending that couldn't promise a return in one to three quarters was asking for trouble from the stockholders. Following the fiasco of the Superconducting Supercollider in the 1990s, the axis of high-energy physics research left the U. S. and returned to its birthplace, Europe. We are relying on the Russians for the next few years of access to the International Space Station once the Space Shuttle is (finally!) terminated. And native-born Americans seem to be allergic to almost any kind of graduate study that doesn't promise quick financial rewards. That eliminates science and engineering graduate study for most of them, which explains the highly international flavor of most graduate schools, and by now most engineering and science college faculties.

That is great news for the folks who come to the U. S. to better their education, of course. We are 99% a nation of immigrants anyway, and I do not begrudge anyone a place in the great adventure of science and technology, no matter where they came from. But we can't rely exclusively on immigration from places which, in another generation or so, will have graduate schools as good or better than ours, partly because people in those countries seem to have learned what we're in the process of forgetting: that some science, in fact all true science, has to be pursued for philosophia—the love of knowledge, not for what it can do or how much money it can make. Some engineers and politicians may not like to hear that, but unless we're prepared to do without new science, we should at least recognize the problem, which, as in any engineering ethics issue, is the first step toward solving it.

Sources: The Polanyi quotations are from pp. 182-183 of Personal Knowledge: Toward a Post-Critical Philosophy (Chicago: Univ. of Chicago Press, 1958). The questions an NSF applicant must answer can be found in the Grant Proposal Guide on the NSF website, http://www.nsf.gov/pubs/policydocs/pappguide/nsf08_1/gpg_3.jsp.

Monday, December 01, 2008

Engineering Social Capital

Social capital is the network of personal relationships—memberships in associations, personal friendships, even people you say hi to at the grocery store—without which a society becomes just a collection of isolated individuals. Of course, even a highly dysfunctional society has some social capital, unless everyone is living as a hermit or a Robinson Crusoe with no human interaction of any kind. But social capital is at least as important to a society's well-being as the more familiar financial and physical kinds. In his 2000 book Bowling Alone, Harvard professor of public policy Robert D. Putnam showed the vital importance of social capital for all sorts of things, ranging from personal health to national prosperity. He also exhibited tons of evidence that the U. S. has suffered a long-term decline in social capital beginning around the 1960s and progressing through the rest of the twentieth century.

As I read Bowling Alone, I kept finding little pieces of my life being explained here and there, and I'm sure nearly everyone who reads it will as well. That Junior League thing that my mother belonged to—social capital. Those bridge games and backyard barbecues my parents were always having—social capital. (Did you know that in 1958, one of every three adults were bridge players?) The stubborn decline in the percentage of eligible U. S. engineers who belong to professional societies such as the Institute of Electrical and Electronics Engineers over the last twenty years—social capital again.

Much of the change is generational. It turns out that the baby-boomer-parent generation which came of age in the 1940s and 50s also participated in a peak in social capital, as measured by everything from memberships in voluntary organizations, to voting and political action, to union membership and writing letters to newspapers. Those who came afterwards do less of all these things. In a series of state-by-state studies that compare the degree of social capital with public health, efficiency in government, economic strength, and so on, Putnam shows that the right kinds of social capital (there are bad kinds, it turns out) benefit these other socially desirable factors as well. In other words, if you live in a state with higher social capital, you're likely to make more money, live longer, and even be happier with your life, on average. So the fact that social capital in the U. S. is sinking is cause for concern.

What has this got to do with engineering? Only nearly everything. Many of the sources of the decline Putnam describes have their roots in technological changes: the rise of television in the late 1940s, the increased urbanization (and suburbanization) of America, even (although he doesn't mention it) technologies like air conditioning, that makes staying inside with the windows closed on really hot days a viable option. Although the Internet was just gathering steam as Putnam finished his tome, I'm sure he'd have a lot to say today about the rise of computer-mediated relationships at the expense of face-to-face encounters.

Far from being a technology-is-bad Luddite, Putnam acknowledges that technology has helped to increase social capital in some ways. The automobile made travelling for social reasons easier, especially for young people. Having online relationships is better than no relationships at all. His main point is that many technologies have externalities—unintended consequences, basically—that often adversely affect social capital. When your choice of what to do in the evening was to go to a movie with friends or play a pickup basketball game, both activities involved you with other people. But now the options might be between watching your latest Netflix CD or trying out your new videogame—typically both solitary pursuits.

Every engineer whose work affects the way people spend their time or money—which is nearly every engineer—should read this book. It's a big book, some 400 pages and notes, but it's easy to read, entertaining in spots, and toward the end Putnam goes on a little Sherlock Holmes quest to find the "culprits" responsible for the decline of social capital. Technology isn't the only one—the shift of women into the workplace, urbanization and sprawl, and several others helped as well—but it is a major enough player to justify the attention of any engineer whose product or service might tend either to bring people together or to isolate them from others.

A good example of a newly engineered product that potentially leads to beneficial social capital is Nintendo's Wii video game console. Unlike other games which require only rapid finger movements, the Wii console encourages the use of games that involve the player's whole body and can more easily involve two players at once. I haven't tried it myself, but if the reports are true, a video game that encourages two people to be playing the same game in the same room produces more social capital than otherwise.

So the next time you stop what you're doing and think about the social consequences of your work, try thinking along the lines of social capital: Will this product bring people together? Or will it be one more reason for a guy to sit alone in his room with his machinery? It's not the only factor, but it is a factor that engineers seldom consider. And Bowling Alone makes it clear that something needs to be done soon.

Sources: Robert D. Putnam's Bowling Alone was published by Simon & Schuster in 2000. For an explanation of externalities, see my blog "The Ethics of Externalities" on Nov. 3, 2008.