Toured National Petascale Computing Facility

Jackie and I got a tour of the NCSA’s National Petascale Computing Facility at the University of Illinois today, where they’re getting ready to install the Blue Waters supercomputer.

This picture shows just the power stations—all the space between these units will, over the next few months, be filled with rack after rack of water-cooled POWER7 modules. (A big part of the building houses cooling towers to dissipate that heat).

There are a couple of supercomputers already installed at the other end of the room, including the EcoG, designed and built by students to enter into a contest for energy-efficient supercomputers. (It took 3rd place overall , and was declared the “greenest self-built cluster.”)

It was build on ordinary commercial-grade racks, which turned out not to be quite strong enough to support all the hardware they were installing—you can see where they braced it with two-by-fours.

A week earlier, we’d attended a tour of the NCSA’s Data Visualization Lab, where we’d been treated to a bunch of 3D videos (shown on a very large, very high-res screen) produced on various supercomputers. It was pretty cool, but I didn’t get any photos worth sharing.

Because I’m a big geek about security and related topics, I was particularly interested in the facility’s secure entry. Employees need to swipe a proximity card and submit to an iris scan. Only after the cylinder closes behind them does it open in front—and it won’t do that if a weight sensor suggests that there’s more than one person in the cylinder.

Secure entry at the National Petascale Computing Facility

Those of us on the tour just walked in through a door next to the secure entryway.

Nuclear power, terrorism, and accepting baseline risks

People are bad at comparing risks, and people like to point this out by making comparisons to risks that people tolerate on a daily basis. For example, pointing out that many more people die in car accidents than are killed by terrorists, or pointing out that providing electricity by burning coal kills and injures more people than providing electricity by fissioning uranium.

At one level, I find these arguments compelling. I find it preposterous that we spend so much money on homeland security. That money would be much better spent (in terms of lives saved per dollar) on traffic safety, or probably a lot of other things. I gather, based on the fact that people keep pointing this out without producing any visible change in funding priorities, that most people don’t find this a compelling argument.

I’ve always wondered about that, and perhaps I’ve figured out why in the “coal versus nuclear” argument, which I don’t find compelling.

Plenty of people die to provide us with power from coal. Miners die from accidents. People die in road accidents moving coal from the mine to the power plant. Workers die in ordinary industrial accidents at power plants. People die from respiratory problems caused by or exacerbated by pollution from burning coal. People die in severe weather—which is becoming more common, probably because of global warming.

Except for that last, this is our baseline status. We know the costs and risks, and we accept them. Some people work to improve things—better mine safety, better level crossings for trains, lower emissions from coal burning—but the baseline is accepted. Importantly, an individual can do a lot to reduce his or her risk, such as by not making a career in coal mining, by exercising due care at rail crossings, and by living some place with clean air (and not smoking).

With nuclear power, things are different. The baseline status is safer. Deaths in uranium mining are very small, because the volume of uranium ore needed is so small compared to the volume of coal. Deaths from industrial accidents are small, because the number of workers is small (and, perhaps, because some additional attention is paid to safety at a nuclear plant for reasons having to do with greater regulation and particular concerns about public perceptions of safety). Deaths caused by the release of radiation are very, very small, because we go to vast effort and expense to avoid them.

But although the baseline status is relatively safe, the contingent risks are huge. The problems that led to the catastrophe we’re seeing now at the Fukushima Daiichi plants are replicated all over the world. It’s not just plants built on fault lines and plants built places where tsunamis can occur. It’s things like redundant safety systems that aren’t really redundant. Most especially, it’s committing to providing active safety over a period that’s much longer than human institutions reliably persist.

On the former issue, I have an oddly relevant memory. As a boy I attended public hearings in Kalamazoo on the licensing of the nuclear power plant at Palisades. At one hearing, a lawyer opposing licensing pointed out that a line carrying backup power for the plant ran through the same conduit as a line carrying the regular power. In some clever showmanship, he snapped a pencil in two to illustrate the fact that this produced a common point of failure. Learning that the backup generators at Fukushima were in basements where they would be lost in a tsunami produced an odd echo of that memory.

The latter issue is really more to the point. We are relying on corporations to actively manage the safety of these plants and the spent fuel—corporations that will cease to exist if the cost of this management burden ever grows to the point that it consumes the corporation’s profits.

I think the degree to which these safety issues needed to be actively managed has surprised a lot of people. I’ve many times heard people suggest that managing nuclear waste was no big deal—just put the stuff in a concrete vault and put a fence around it with signs saying “If you come in here you’ll probably die.” I always knew that was dumb, but I was mostly worried about people deliberately coming in to use the waste to make dirty bombs and the like. I didn’t quite realize to what an extent the spent fuel rods depended on a whole complex system of cooling equipment to keep them from bursting into flame and spreading radioactive smoke and steam wherever the wind blew.

So that, I think, is why we accept coal power and think of nuclear as dangerous. We could give up coal power anytime we, as a society, decided that the cost was too high. If we were willing to cut way back on air conditioning, electric lighting, and all the other things we run with electricity, we could just quit the whole thing. The only dangers left behind would be some moderately dangerous holes in the ground, some toxic heaps of ash, and the pollutants that are already in the air. With nuclear power that’s very much not true. We could give up nuclear power today and we’d be on the hook for decades of active management of the high-level waste and generations of (mostly passive) management of the low-level waste.

I think maybe the issue with risks from terrorism is the same. People know what the trade-offs are for driving. If we, as a society, decided to give up driving, we could cut deaths from road accidents almost to zero. But terrorism isn’t like that. There’s nothing we could give up to prevent terrorism, and the contingent risks are huge. An endless stream of terrorist acts that killed tens, hundreds, or thousands of people seems very different from the many other activities that we engage in that cost tens, hundreds, or thousands of lives.

It’s a bummer about nuclear power, though. It would be cool if a network of high-speed electric trains could provide transportation in a post-peak oil world, and I’d begun to think it might be a reasonable alternative. A mere twenty-five years with no major nuclear accidents was enough to make nuclear power start seeming pretty safe again. This is a good reminder that it really isn’t—and that we need to think carefully about the difference between accepting risks for ourselves now, and accepting risks for everyone stretching off into the future.

[Update 2011-03-23: There’s a lot of  misinformation about whether very low doses of radiation are harmful. Here’s a paper with a survey of what we actually know about the effects of low doses of ionizing radiation (from the Proceedings of the National Academy of Sciences).]

Making doom funny

In my review of Dmitry Orlov’s book Reinventing Collapse, I talk a bit about how everyone says that the book is funny, but no one ever quotes the funny bits. There’s a reason: The humor sneaks up on you, building on previous bits. All the really funny bits are only funny if you’ve read up to them.

For those of you who want to read something really funny about peak oil, but were unconvinced that such humor was worth shelling out the cost of a book (or taking the time to read it), there is now an alternative: Dmitry Orlov’s latest article at Culture Change, Peak Oil is History.

Once again, it’s tough to quote a sentence or a paragraph that’s funny, but that’s okay: Just click on over and read the article. It’s free, and it’s much shorter than a book.

Whether you’re one of the people who understood peak oil some years ago or one of the people who just figured it out, Orlov wants to make sure that you understand that the reality of life on the declining side of the oil production curve won’t look like the mathematically smooth logistic function that’s usually displayed. Rather, it will look something like the front side of the curve, with spikes and dips that map to wars and recessions and other catastrophes. Further, he wants to make sure that you know those little jerks up and down—especially the jerks down—matter to you.

It would be theoretically possible to ride the downward curve of oil production in a fashion that would look like the reverse of riding it up. In fact, if we’d spent the thirty years since Jimmy Carter warned that our “intolerable dependence on foreign oil threatens our economic independence and the very security of our nation” preparing to do so—improving our rail infrastructure, switching to wind and solar energy, and generally becoming much more efficient—we’d be in a position to do that pretty comfortably.

In practice, though, things are going to suck.

Things will, however, suck rather differently than people expect—which is Orlov’s point. People expect that the rich will go on much as they have, while the poor will get squeezed by high prices—and there will be plenty of that. But after laying out the reasons why it won’t work that way, Orlov concludes by saying, “it becomes difficult to imagine that global oil production could gently waft down from lofty heights in a graceful smooth and continuous curve spanning decades. Rather, the picture that presents itself is one of stepwise declines happening in more and more places, and eventually encompassing the entire planet.” A stepwise decline that quickly results in even rich people having “no access to transportation fuels and severely restricted transportation options.”

Orlov makes doom just about as funny as possible, perhaps even a little funnier.

What renewable energy really looks like

Tobias Buckell just posted about Portugal’s push into renewable energy. He links to an article claiming that 45% of Portugal’s grid electricity now comes from renewable sources, and that they’ve managed this with just a 15% increase in electricity costs. Making the (somewhat unlikely) assumption that one could get another 45% increase for another 15% increase in price, he suggests that it would be totally worth it:

I’d take a 30% hike for energy independence and no money being sent to terrorists in a fucking heart beat.

Frankly, I would too. In fact, I’d be willing to pay a lot more than that. Unfortunately, I’m afraid it would cost a lot more than that—more than most people would pay.

First of all, Portugal was already paying about twice what we pay in the US for electricity. The 15% bump was on top of that. Second, Portugal had substantial untapped sources of hydro power. The US doesn’t.

Either of those, I expect, would doom the project. The first makes it unaffordable—I’d be willing to pay 30% over double what I’m paying now for electricity, but I doubt if very many other people would. The second makes it impossible—we have a lot of untapped wind power, but that comes and goes. Use of wind power will grow, but even with a much better grid (to distribute power from where the wind is blowing to where people are using it), you need something more reliable for baseline power.

But neither of those is the real problem, which is that the US uses three times as much electricity per person than Portugal does. (13646 kWH  versus 4663 kWH per capita in 2005, data from the World Bank.) If you look at the historical per capita energy use in each country, you can see that both countries have shown steady growth—but Portugal is only up to about where the US was in the early 1960s. (And, sadly, following right in our footsteps.)

So, to shoot for the Portugal model we’d have to:

  1. Cut our energy use by two-thirds,
  2. Double the price (plus 30%), and
  3. Either invest vast additional sums in the grid (perhaps $100 billion) or accept brownouts when the wind wasn’t blowing.

Again, I’m totally up for that. My electricity consumption is probably already two-thirds below the US average. My typical electric bill runs just about $30; I’m sure I could stretch my budget to cover $70 if the payoff was no more carbon in the air and no more sending buckets of cash to people who hate us.

But based on the way people actually behave, I’m forced to assume that most people would rather burn the planet and fund terrorists than turn off the AC, downsize the car, and pay up for organic, locally grown food.

Responsibility for oil spills

I’m as outraged as anyone at the incompetence that led to the Deepwater Horizon oil spill in the gulf: both the slipshod regulation by the government and the incompetence and criminality of BP, Transocean, and Halliburton. I wouldn’t mind one bit if all three companies were broken by cleanup costs, restitution to injured parties, and civil and criminal penalties. But I’m a bit sad to see all the blame being laid at their doorstep.

The fact is, spills like this are an entirely predictable result of consuming 85 million barrels of oil per day. If you consume that much, you have to produce that much. And if you produce that much, you will have accidents. Some of the accidents will kill people. Some will contaminate huge swaths of the ocean.

Sure, BP et al deserve much of the blame. But there’s plenty of blame to go around. A good share of it belongs to every one of us who drives a car, heats their home, or buys anything made out of plastic.

What did you think was going to happen?