Stardate 20020924.1903

(Captain's log): I may as well run through some of the other alternate-energy pipe dreams and point out their flaws:

Geothermal: It works in Iceland, but Iceland is one big volcano. The number of actual sites in the US where geothermal is practical is small, and they're all isolated, and most of those are in national parks. (The best geothermal source we own is Yellowstone. The second best is Mauna Loa.) Geothermal releases huge amounts of sulphur gases, and if you use equipment to remove that it also consumes most of the energy. Even if fully developed, the energy generation would be miniscule by comparison to our total consumption.

Solar: I talked about that here. Solar can also be used for water heating and home heating in some areas, but the total energy involved is not that great in the grand scheme of things, and it requires an impressive amount of equipment to be purchased and installed per megawatt.

Wind: It isn't where we need it, and it isn't when we need it, and there ain't enough of it. The power grid has to adjust its energy generation to match consumption, and we can't turn the wind on when we need more energy. The source is diffuse and it requires a massive investment to make and install all the windmills. There are not all that many appropriate sites where the wind is regularly strong and a lot of the places where that's true (e.g. the Columbia River Gorge) are protected areas. Windmill farms are an eyesore, and they kill a lot of birds. (A lot of birds.) The equipment is large, complicated and will require a lot of repair to keep working; the resulting energy will be inadequate and unreasonably expensive per unit energy yield. And I'm still not convinced that it won't take years before any given windmill finally yields as much total energy as it took to make it, transport it and install it. Ireland is making a massive investment in wind power, but when they're finished and have fully deployed all sites it's only going to generate 520 megawatts, when the wind is blowing. That's one eighth of the power generated by The Dalles Dam.

Fusion: wake me when it actually works.

Solar satellites: In about a hundred years this will be fine. But until we have solved the minor problem of cheap bulk cargo delivery to orbit, it isn't feasible. (The only other way is to make them in orbit, using material from the Moon, and that is an even more massive problem. It would require establishment of a Moon colony, for one thing. And even that requires bulk cargo delivery from Earth to orbit.) As it currently stands, there are physical reasons why the cost in energy of putting such a satellite into orbit is huge, and it would take a long time before such a satellite paid for itself in energy terms, let alone financially.

Tides: Harnessing the power of tides isn't all that easy. The original proposal was to create pools which filled at high tide, and then drained out through fairly standard turbine electrical generators as the tide fell again. The energy generation is intermittent and not timed to usage. And there's a much larger problem having to do with how turbines work. As a general rule, efficiency of a turbine increases as a function of the pressure being fed through it, and in most parts of the world the tide only goes up and down a few feet. That doesn't provide enough head to actually make turbines work well even if you wait until absolute low tide before beginning to drain your pool. (The turbines in The Dalles Dam are designed to utilize 81 feet of head, which is to say that the surface of the reservoir should be 81 feet above the turbines. No place in the world has tides remotely like that except the Bay of Fundy.) Which means that to generate any kind of significant power this way you need a really big pool and a heck of a lot of turbines, because the conversion process will be extremely inefficient. I don't believe this can be scaled up to the point of creating enough energy total to actually make any difference, especially considering the ecological damage it would do to coastal waters where it was deployed. Another big problem with tides is that it's saltwater, which is highly corrosive to both metal and concrete.

Fission: Completely practical from an engineering standpoint. Completely useless from a political and economic standpoint, at least in the US. The cost of a new fission plant is stratospheric now because of the amount of liability insurance it would have to carry, even if for no other reason. And the regulatory burden is beyond belief. It's been something like twenty years since the last design start for a fission plant in the US, and since the WPPS fiasco, no investor will be willing to take the chance again until something drastically changes.

Coal: Coal works really well. That's why it's our primary source of energy.

UPDATE 20020925: Different people have written in to point out two major omissions from the above.

Hydrogen: (non-fusion, hydrogen as chemical energy) The problem here is that hydrogen is a fuel but not an energy source. Gasoline is both. But there's no substantial natural source of hydrogen which we can tap, so any hydrogen we use can only be created by utilizing energy from some other energy source. Hydrogen is like electricity, a way of moving energy from one place to another. That's why discussion of conversion from internal combustion engines to fuel cells in vehicles may well be important when you're concerned about air pollution or changes in industrial policy but isn't when you're talking about energy sources.

Hydrogen currently has some nontrivial engineering problems though, the Hindenberg disaster notwithstanding, excessive flammability isn't one of them. In fact, gasoline is more dangerous than hydrogen in those terms. The film of the Hindenberg burning is quite impressive, but hydrogen is in reality not very flammable. When released it rises extremely rapidly because it is so light (much lighter than helium), and it also tends to thin out so fast as to no longer produce a flammable mix with the air.

The biggest problem with hydrogen now as a fuel for vehicles is that it's really hard to store an adequate number of joules in a small space with hydrogen without liquefying it. The best answer so far seems to be certain metals which spontaneously form hydrides and release the hydrogen equally readily, but the energy density doesn't appear high enough yet to be practical, and no one will want a vehicle that has to be refueled every fifty miles or less. Any attempt to switch to hydrogen away from gasoline for our vehicles would require an immense change to the physical plant. Not only would a significant number of vehicles have to be replaced, but we'd also have to build major installations to create the hydrogen using energy from some other source, and we'd need to create an entire infrastructure to distribute it and make it available at service stations. Irrespective of any technical problems involved, the capital investment would be immense, and since it isn't actually a source of energy, it is irrelevant within the context of reducing our use of Arab petroleum.

Biomass: I'm a bit embarrassed that I forgot this one to begin with. Biomass is, at its core, an extremely roundabout form of solar power. The idea is to use farmland to grow greenery, and then to burn the greenery to generate energy, but depending on who is making the proposal the details can vary wildly. Ethanol as a fuel is one example of this, but it is exceedingly inefficient because it is based on corn and only utilizes the grain, and wastes most of the energy in that and uses none of the energy in the rest of the plant.

A more efficient form of biomass is methanol, which can be created from the entire plant. The most efficient form is to burn the entire plant in a big power facility, for instance as a substitute for coal in electrical generation. There are some engineering issues involved, such as the fact that the biomass has to be dried or somehow have its water content reduced substantially, but that's a detail of the process.

It's an attractive idea, but I'm not sure the numbers make sense. I'm not sure I believe it's possible to actually supply a significant portion of our current energy use this way. You're only talking about actually harvesting greenery from the fields once or at most twice per year, and you're only going to get a few tons of dried fuel per acre each time you harvest. The US uses about 60 million short tons (about 55 million metric tonnes) of coal per month, or about 650 million metric tonnes per year.

According to this page at Oak Ridge National Laboratory, anthracite and bituminous coal (which make up most of the coal we use) contain 27-30 gigajoules per metric tonne. Agricultural residues are 10-17 gigajoules per metric tonne (as a function of water content). As an approximation, that means about 2 tonnes of biomass would be needed to replace each tonne of anthracite. Are we actually capable of producing, collecting and transporting 1.3 billion tonnes of dried biomass per year? The US currently has about 200,000 square kilometers of irrigated land; can we generate seven thousand tons of biomass per square kilometer? Or even a tenth of that? Not easily, if it's possible at all.

One of the reasons we can produce that much coal is that it's concentrated; a given coal mine can produce millions of tonnes of coal with a relatively small amount of machinery. But biomass would be extremely spread out, and you'd need an impressive infrastructure investment for all the trucks to collect it and bring it to rail yards for transport to the power plants. And a non-obvious part of the problem is that right now our agricultural practices are using that same biomass, partly to stop soil erosion and partly to reduce fertilizer usage. (Also, a lot of it is used as animal feed.)

A general comment on gasoline and oil: Another point which one of my correspondents made is that a refinery doesn't take oil and make gasoline for a while, and then switch to making diesel, and then switch to lubricants. They do not have the ability to arbitrarily control how much of each product they produce. Oil refining yields certain proportions of each product, and they all emerge simultaneously in ratios which can to some extent be adjusted but not to the extent that many think. Refining isn't really a manufacturing process; it's more like large scale fractional distillation, which takes what's already in the oil and separates it.

You're always going to get at least some gasoline out, quite a lot in fact, if you need the other things coming out of the refinery, like heating oil and diesel fuel. It does little good to reduce consumption of gasoline unless you equally reduce all the other needs by about the same amount. (In fact, the reason that gasoline was originally selected in the late 19th century as the fuel of choice for vehicles is that it was a waste product of the petroleum industry at the time which was making its money producing other things, mainly kerosene and lubricants.) The reality of the process is that the gasoline is going to be produced whether it's consumed or not; there's hexane and heptane and octane in there, and it's going to come out when you refine the oil. If the market for it is removed, they'll burn it at the refinery as a waste product.

Actually, what would happen is far more prosaic: with reduced demand, the price would drop because supply is relatively inelastic, which would give consumers more of an incentive to keep using it, which would work to stop the process of reducing demand. Simultaneously, the price of the other petroleum products would rise, somewhat reducing demand for them, and thus overall reducing the total usage a bit. But nothing like as much as you'd expect, in part because the demand for many of the other products is also now quite inelastic.

Update: yet more here.

My new blog: Chizumatic