Monday, July 27, 2009

Smashup at an Atom Smasher in Europe

Accidents at high-energy physics labs do not normally make the news. But when the lab is the European Union's crown jewel consortium CERN, and the accident puts their latest and greatest accelerator, the Large Hadron Collider, out of commission for up to a year, the incident is worthy of wider attention than it has gotten so far.

America used to dominate the field of experimental physics, as our first-out-of-the-gate development of nuclear weapons proved. But with the political collapse of the Superconducting Supercollider project, a Texas-based particle accelerator that was cancelled in 1993, the world leadership in high-energy experimental physics began to move to Europe, where last September the Large Hadron Collider was being put through its initial tests.

State-of-the-art particle accelerators are some of the most complex experimental systems ever built. The LHC resides in an underground tunnel 17 miles in circumference under the border between France and Switzerland. The basic idea is to shove atomic nuclei around a huge evacuated pipe with the aid of strong electromagnetic fields and the guidance of monstrous superconducting magnets that have to be submerged in tons of liquid helium. Such enterprises are of course very costly in comparison to more modest laboratory equipment, but a $5 billion expenditure these days when the U. S. Congress is cogitating about health-care programs costing 200 times that much doesn't seem like a lot of money. Nevertheless, you want to protect your investment even if it is only $5 billion, and that seems to be what got damaged the most in last fall's accident.

Superconducting magnets are used because once you get current running in them, no electrical power supply is needed for the magnet itself. To produce the tremendous magnetic fields needed with conventional magnets would require prohibitive amounts of power. But running a superconducting magnet is not a trivial task. The LHC's magnets have to be cooled down to about two degrees C above absolute zero, and only large amounts of liquid helium can do that. So CERN became the world's largest user of helium in its attempt to cool down all the magnets for their inaugural run of the LHC last September.

Nine days after the tests began, an electrical connection between two superconducting magnets apparently failed. In a normal magnet, this would not be that much of a big deal, since an open connection would just cause maybe a transient electrical discharge and then the whole thing would shut down. But with a superconducting magnet, a bad connection causes heat. Heat causes a superconductor to abandon its main desirable property, which is to be superconducting with no resistance. Once some resistance shows up, that generates more heat because the current cannot stop instantaneously in a large magnet. The heat boils off more helium, more of the magnet heats up and gets resistive, and you have a great big vicious circle called a "quench."

Being very large magnets, the LHC units quenched in a big way. The helium pressure was so high it blew all the way into the vacuum tunnel and spread insulation and trash everywhere. Since CERN policy is to clear the tunnel of personnel any time tests like this are going on, no one was injured. But successive press releases in the weeks and months after the accident cited longer and longer delays before the system could be up and running again. As of this writing (July 2009) repairs are still being made, and hopes now are that the LHC can go online again sometime this fall.

While the CERN managers are to be congratulated that this accident didn't hurt or kill anyone, having a multi-billion-dollar machine damage itself to the extent that it takes a year to fix is not exactly wise use of resources either. Any system as complex as the LHC will do unexpected things when first fired up, and CERN has decided to install fault detectors as part of the repair process that will give advance warning of a possible quench condition in the future. This is wise, prudent, and consistent with the highest engineering ethics principles. We just hope it will prevent such accidents in the future.

Which brings us to the larger question: what is all this billions of dollars of machinery and personnel good for? The holy grail of high-energy physics research right now is a thing called the Higgs boson, named after one of the six or seven theoretical physicists who thought of the concept in 1963 and 1964 (somebody had to be first, and it was Peter Higgs). According to the most widely accepted model of subatomic particles, called the Standard Model, the Higgs boson somehow gives mass to all elementary particles (such as electrons, I suppose) that have mass. No one knows how heavy this Higgs boson is, but guesses range from about the weight of a silver atom's nucleus on up. The problem with making really heavy particles like this in a particle accelerator like the LHC is, you need a lot of energy per particle. Current machines can get up to 120 billion electron-volts of energy (120 GeV) into a particle, and the LHC is supposed to reach values about ten times higher. When it finally works. So hopes are high that at long last the elusive Higgs boson will show its face, or tracks, or however they plan to catch one.

Fortunately, nobody has yet had to die for the chance to discover the Higgs boson. But a lot of people have put a lot of effort and money into looking for it, and now after the accident they're having to wait an extra year. Some may look at such work and criticize it on the basis that hey, people are starving in Rwanda and we're spending billions on chasing some pencil-pusher's pipe dream? But the fact that humankind can contemplate the universe and expend lots of energy searching for purely abstract products of the intellect such as the Higgs boson, is one of the things that distinguishes us from animals. I for one wish the CERN workers first safety, then success in their search. But being a Texan, I also wish we'd finished the Supercollider and made Higgs bosons first right here in Waxahachie.

Sources: Several news articles are available on the vicissitudes of the LHC accident. I used material from the U. S. LHC website http://blogs.uslhc.us/?p=393, an article in Discover Magazine online at http://blogs.discovermagazine.com/80beats/2009/02/10/until-next-fall-lhc-smashes-only-hopes-not-particles/, and one at the Big Science News blog at http://bigsciencenews.blogspot.com/2008/09/lhc-quench-stops-cern-re-start-delayed.html. For an instructive news video of what an unplanned superconducting-magnet quench looks like, see the YouTube version of a report on a hospital's MRI magnet that blew up at http://www.youtube.com/watch?v=1R7KsfosV-o.

Monday, July 20, 2009

The Third Pole of the Health-Care Debate

In the current debate over how health care in the U. S. should be funded, no one seems to be talking in a first-principles way about the moral underpinnings of the system. As I've said before, technology and engineering are so heavily involved in modern health care that I think engineers are disingenuous if they say they don't have a dog in this fight.

Sometimes situations can be clarified by going to extremes. I will pick three extremes which I think represent the poles of what we could do about health-care funding, and then see what conclusions we can draw about the moral bases of each.

The first pole is the extreme-libertarian solution: get government totally out of health care and let people figure it out on their own. This would mean shutting down Medicare and Medicaid, selling off all government-run hospitals, and letting the chips (and the patients) fall where they might. I think even the most extreme libertarians might blanch at ending such long-established practices as the licensing of doctors, but maybe not. One of my cousins is a libertarian, and he hasn't blanched in years. Surprisingly, this solution might not make much difference to those who are employed at firms that provide good health benefits. But for retired people who have come to rely on government-paid health care, there would be big problems, as well as for many who can't afford private care or insurance for it. This solution maximizes individual freedom from government actions, but if you're poor or retired or both and can't afford medical care, you would be out of luck. This shows the radical individualism of libertarian philosophy, and why it is inferior as a total guide to life.

The second pole is the extreme-socialist solution as is done in Cuba, for example: free health care for everybody. Of course "free" means the government pays for it all, and Except for the sick folks that documentary filmmaker Michael Moore took along with him to Cuba as an example of how much better their health-care system is than ours, I'm not aware of a huge surge of people who travel to Cuba for their marvelous medical institutions or services. The fact is that any government, especially Cuba's, has finite resources, and when resources are allocated by bureaucrats rather than markets, the results are often less than optimal, even if there are enough resources to begin with, which in the case of Cuba is doubtful. This solution removes a person's economic status from the equation, at least in theory, but requires rationing, waiting lines, and a pretty low average level of quality. Besides which, it seems that government-run health care systems tend to encourage a psychological dependency on the state which some think is mentally unhealthy in the long run. As with most "free" offers, free medical care often comes with political enslavement.

Most debates about the subject seem to be focused on just where between these two poles we ought to land. But I think that leaves out a critical factor, which I will dignify with the designation of my third pole. It's not easy to think of a name for this pole that will not evoke negative connotations. Perhaps "charity" in the older sense of "love" would cover it. Many private and public hospitals provide services to indigent patients whose costs they absorb, which means that everybody else who uses the hospital helps pay for it. Historically, the idea of caring for the sick without consideration of cost was a founding principle of many medical institutions with religious backgrounds. Many hospitals were staffed by nuns who took vows of poverty. And I think when this motivation is present, it forms the best of all foundations for individual careers and institutional principles. Would you rather be treated by a doctor who went into the business because he wanted to help people, or because he could make a pile of money? Yes, skill is part of the equation, but skill is more than mere technical proficiency. Being a quality of character, charity does not fit easily into economic calculations or political structures. But the first two poles either discount it totally or regard it as an unreliable and suspect motivation that is best ignored in favor of government-run solutions to the problem.

Many religious leaders, up to and including Jesus, made healing a vital part of their ministries. I do not have all the answers to our health-care problems, but I think we should consider making more room for and encouraging those who provide care in the neighborhood of the third pole—people and institutions who help patients because it is the right thing to do, not just because they can make money at it or because the government compels them to. If the debate can center more around this idea, I think the outcome, whatever measures it takes from the other two poles, will be better than otherwise.

Monday, July 13, 2009

Cyber-Security and North Korea: Worth Worrying About?

Beginning on July 4, numerous government and private websites in both South Korea and the U. S. succumbed temporarily to attacks by a shadowy entity suspected of connections with North Korea. Although the damage and disruption were temporary, this sort of thing may be a small wave of a big future for web-based warfare. But unlike conventional warfare, which has at least the restrictions of the Geneva Convention, cyber-warfare is so new that there are few international agreements about it, and even less agreement among those responsible in the U. S. as to what should be done to defend against it.

First, the attacks. According to the Wall Street Journal, they were "relatively unsophisticated," but that doesn't mean they weren't effective. The distributed denial-of-service attacks were carried out by large numbers of computers which harbored a virus that ordered them to flood the targeted websites with requests for service. It takes time for website operators to notice what's going on, get a fix on where it's coming from, computationally speaking, and devise work-arounds to restore service to legitimate users. In the case of these attacks, the time involved was as long as three to four days during which service was interrupted or degraded to some degree. Besides several government-operated websites in North Korea, U. S. websites operated by the Defense Department, the Federal Aviation Administration, and private entities such as the New York Stock Exchange and Amazon.com were attacked.

Although positive identification in these types of attacks is difficult, the timing and other characteristics of the attacks makes it likely that North Korea is responsible. North Korean dictator Kim Jong Il is in poor health and has not made public his plans for a transfer of power. That makes the normally volatile country even more unstable and likely to pull malicious pranks such as nuclear-weapons tests and missile firings, which have also occurred in recent months. But when should we quit calling these web attacks pranks and take them more seriously?

Cyber-warfare is the ultimate in modern conflicts. It resembles conventional terrorist actions such as suicide bombings in that its effects are large in proportion to the resources required, its perpetrators wear no uniforms and can blend into the civilian populace easily, and identifying even so broad a category as the country of origin for a cyber-attack is difficult, let alone finding the people responsible. As far as I know, no one has died as a direct consequence of a cyber-attack, although as more and more life-critical systems from medical care to power grids depend on the Internet, this may soon change. Websites accessible to the general public are the easiest targets, but the harm caused by disabling them is generally limited to loss of revenue or public access, which is inconvenient but not life-threatening.

We can expect that attackers will grow in sophistication and focus as time goes on. There is already some concern that critical infrastructure systems that use the Internet are more vulnerable to attack than they should be. But if history is any guide, we will coast along in blissful ignorance until someone wreaks real harm—death or destruction of large amounts of real property—before steps are taken to remedy these vulnerabilities.

Conventional wars were fought over physical objectives such as territory, natural resources, or lives. As much as many terrorists would like to, no one has yet figured out a way to kill you by means of your own computer, unless you count grabbing your laptop and lamming you over the head with it. There is a cautionary lesson here for those who would like to integrate their own bodies with the web by direct implants of nerve-stimulating devices in the brain and so on. If a computer does something I don't like, I can always walk away. But if it's wired permanently to my brain and some hacker gets his hands on it, I won't have that option. There's a sci-fi movie for you, but one I wouldn't want to watch.

President Obama, to his credit, appears to be the most web-savvy occupant of the White House so far. But his plans to name a cyber-czar in charge of government internet security have hung in limbo as he searches for a suitable candidate. It's not like the President has nothing else to do, but this may be one of those cases where we will wish he'd paid a little more attention to a low-profile matter at the expense of one more town-hall meeting on health care, for example.

All the same, I hope that such a czar will wear his or her authority lightly. One of the strengths of the Internet, and the cyberworld as a whole, is the way that highly distributed expertise works very effectively most of the time to remedy problems as they occur. It is an example of engineering ethics at work that is quiet, not flashy, but worthy of our attention nonetheless. The great majority of computer and networking experts have a vision of the way things ought to be that is both moral and efficient, and tend to work most of the time in cooperation with each other to keep things running well. But the strength of such distributed expertise is also its weakness, in that it takes only a few malicious people who grab the controls to mess things up. Let's hope that we can mount organized defenses against such attacks in time to thwart them before they cause the kind of headlines that 9/11 received.

Sources: I used information from the following articles on the recent cyber-attacks: a piece by Donald Kirk of the Christian Science Monitor at http://www.csmonitor.com/2009/0708/p06s24-woap.html, and an article in the online edition of the Wall Street Journal by Siobhan Gorman and Evan Ramstad at http://online.wsj.com/article/SB124701806176209691.html.

Monday, July 06, 2009

Exclusive: Why the Austin Scaffolding Fell

On June 10 of this year, three construction workers died when a section of scaffolding collapsed at a construction site near the campus of the University of Texas at Austin. Because the accident resulted in deaths, personnel from the U. S. Occupational Safety and Health Administration began an investigation shortly thereafter. While it is the normal policy of OSHA not to divulge information on an investigation that is incomplete, leaks do occur. If you keep reading, you will be the beneficiary of one such leak. But first, some details about the accident.

Photographs published at the time showed that the scaffolding was a cantilevered box frame that extended several yards on either side of an elevator-type mechanism that allowed the whole thing to move up and down the side of the high-rise apartment complex under construction. Photos of units on the manufacturer's website show that the scaffolding can extend as far as 25 or 30 feet either side of the mast that supports it. The workers using the scaffolding walk along the frame to do whatever operations the construction job requires.

There's no particular mechanical-engineering magic to cantilevering. Like a branch of a tree that sticks out sideways, the box frame's unbalanced weight has to be supported by the central elevator. In particular, the bolts that keep the top of the box frame attached to the elevator unit are under tremendous tension, since the whole weight of the frame tends to pull the bolts away from the elevator. If these bolts aren't strong enough and fail, the scaffolding will fall away like a branch cracking off a dead tree. That is apparently what happened on June 10.

The following information is what the newspeople call an unconfirmed report from an anonymous source. But I believe it to be reliable. According to the source, the bolts used in the scaffolding that failed were not the required Grade 5 type. If you have ever bought bolts at a hardware store, you may have noticed little patterns of lines on the hexagonal bolt head. These are not just random forging marks. They indicate the quality of steel in the bolt, and in particular, the maximum stress the bolt can withstand. Grade 5 bolts have three embossed lines in a "Y" pattern on the bolt head, indicating that they can take up to 120,000 pounds per square inch of stress under certain defined conditions. Cheaper bolts with no head markings are available. For example, Grade 1 bolts have only about half the stress capacity of Grade 5.

Assuming my source is correct, how did the wrong bolts get there? To answer that question requires that we move from the realm of science to the realm of human behavior. The problem could lie anywhere from the manufacturer of the scaffolding, to the general contractor for the site, to the subcontractor who installed the scaffolding, to the subcontractor's employees who put it together, to someone else who might have lost the Grade 5 bolts and substituted others without knowing the danger it would cause. Presumably the OSHA investigators are also working on this angle of the problem, which will require extensive interviews and inquiries which may never reach a satisfactory conclusion. But this shows the vital importance of apparently minor details, such as three little lines on a bolt head, to the safety of construction workers who probably had nothing to do with the assembly of the scaffolding.

A subsequent news article in the Austin American-Statesman pointed out that Texas has one of the worst records in the U. S. for construction-related fatalities. Pay rates are low, some construction employers opt out of workers' compensation insurance, and since OSHA primarily investigates complaints lodged by workers, the workers have to be their own safety inspectors. Since many speak only Spanish and some are undocumented, their reluctance to undertake this role is understandable. Simply identifying the next of kin of the workers killed in this accident was a challenge to authorities, since their families lived in small towns in Mexico.

How can accidents like this be prevented in the future?

This incident shows how vital the "soft" technologies of training, inspection, and good management can be to safety improvements. Even if we give the benefit of the doubt to the contractor and subcontractors by assuming the use of the cheaper bolts was accidental, the accident reveals a grave deficiency in the way supplies are inventoried and assembly procedures are carried out. No human-operated process can be made 100% foolproof, but if certain bolts have to be Grade 5 in order for a scaffold to be assembled safely, there must be a way to make sure that only Grade 5 bolts go into the system. Simple things like matching spray-paint colors or other wordless guidelines can go far to prevent tragedies like this, especially where non-English speaking employees are involved.

Such measures come too late for the three workers who died when the less expensive bolts apparently failed. But I hope as more details emerge, the lessons of how to avoid this kind of accident won't be lost on those who are in a position to make sure their employees and subcontractors have a safe working environment.

Sources: My source for the bolt information is anonymous for the simple reason that I do not know his name. Published information on the scaffolding collapse can be found on the Austin American-Statesman website at http://www.statesman.com/search/content/news/stories/local/2009/06/17/0617construction.html. Information on the grades of bolts available can be found at http://www.engineershandbook.com/Tables/boltgrades.htm.