2026 Electricity Cost, by Source

These are the fully burdened costs of each electricity source, counting capitalization, construction, maintenance, fuel depreciation, etc., as of today

Coal                        .092 to .23 per kWh

Natural Gas            .065 to .10 per kWh

Hydro                     .02 to .06 per kWh

Nuclear                    .037 per kWh

Utility Scale Solar  .03 to .05 per kWh

Residential Solar    .14 to .18 per kWh

Wind                       .024 to .075 per kWh

 

The bottom line: as of today (a couple of years ago, actually), utility scale solar was cheapest and shows signs of continuing to come down.  Hydro is #2, where you can get it, and that price has been stable for some time.  Wind is #3.  Nuclear is potentially competitive, but there are a lot of complicated externalities which make the cost in the table nearly impossible to achieve.  The recent standby, Natural Gas, is now close to the bottom of the heap, and the old standby, Coal, is significantly more expensive.

There are very few externalities from Solar.  The initial cost of the panels, once very large, is approaching trivial and the cost of the land is dominant.  They have very few environmental consequences. (although not zero: the shade changes hunting patterns for predators, for example, and productivity of farmland is somewhat reduced).   The panels last for decades, although recycling them when they eventually do wear out is a small problem.

There are more externalities for wind.  Perhaps the worst is that they interfere with the flight of some birds.   They make a noise which can be annoying if you're close, and some people find them unsightly.  They last 10-20 years, and most of them can be recycled into new windmills when the time comes.

Hydroelectric power has a very high upfront cost, and a theoretical lifetime of about a century, but the production costs are extremely low.  They turn out to have had a much more profound impact than was at first realized, especially on wildlife, such as fisheries.  Fish ladders had seemed to work, but a lot of them were defective and only a tiny fraction of the fish got through, for example.

 Natural Gas, mostly produced in the US as a byproduct of oil drilling, or hydraulic fracturing, has come on line quite strongly in the last couple of decades.  The fuel is comparatively cheap, and produces less CO2 and other pollution per kWh than burning coal or liquid fuel, such as gasoline.  But still quite a lot.  Another useful technology has been Combined Cycle turbines, which pair an internal combustion gas turbine with a steam turbine that exploits the waste heat.  They're able to get 60% or more efficiency, while conventional turbines struggle to get 40%.  But perhaps the biggest advantage is that an individual turbine can be throttled, to match demand, without much efficiency loss.  This makes it incredibly useful as a variable power source that's easily adapted to rapidly changing situations.

Coal is a historic fuel.   Because it was easily available in most of the places to industrialize early (e.g. Scotland, northern Germany, much of the US, China), it was a convenient energy source, and in a very real way, powered the industrial revolution.   It is difficult to throttle power generation to match demand, and all of the ways cost a lot of efficiency.  Moreover, it's dirty, inefficient and has lot of nasty externalities, such as open pit mining and mountaintop removal.  It's been the most expensive of the leading ways to produce energy for about 3 decades and has been rapidly declining in popularity.

Geothermal vs Ground Source Heat Pumps

This entry is about an important point of terminology--and a short discussion of each.

Geothermal energy is a way of extracting heat from the earth.  Usually, this is heat from a volcano or geyser or other feature of the earth's surface that brings very hot material close enough to the surface that it's practical to extract energy from it, and either use it directly or turn it to electricity with a steam turbine or other mechanism.  The number of spots where mother nature has been cooperative enough to make this practical are relatively few--consequently, geothermal energy isn't viable in a majority of places.  There is a vast amount of heat down there, far more than we could ever use.  But there have been many attempts to drill down to it, and so far, all have failed.  It's only practical where geology has made an opening for us.

There is a significantly different technology which is often misleadingly called geothermal.  A more accurate name would be a Ground Source Heat Pump.  This technology takes advantage of the fact that large masses of soil, stone or especially water tend collect and store heat, and remain stable year round, at whatever the average temperature is.   This tends to be about 50F or 10C, although many factors affect this.   By passing liquid through it, a heat pump can extract the energy from this and use it to provide heating in the winter and cooling in the summer.   This has enormous efficiency advantages: to cool a space to 70F when it's 90 outside and the ground is 50 takes no energy at all, except for whatever is needed to power the pumps and heat exchanger, while heating above 50 when it's cold outside takes a lot less energy than heating above (say) 10F, a temperature at which heat pumps are very inefficient.   All it takes is some plumbing, buried deep enough and spread out enough to take advantage.

 

 

Al Gore was an early adopter of a Ground Source Heat Pump system in his Tennessee home, and spoke of it often during his presidential campaign, usually calling it Geothermal. 

 

Geothermal energy is one of the very few sources of energy that is not really just a way of extracting solar energy that's been stored up over time.  It's actually heat that's either left over from the formation of earth, compression by gravity, or nuclear reactions (both fission and fusion) occurring deep within the earth's core.  Ground Source, as well as wind, photovoltaic, fossil fuels, etc., are ultimately ways of using the energy from the sun that's stored up by the earth somehow.   e.g.: chlorophyll turns  CO2 from the air into cellulose and other burnable materials, which can be burned to get energy--either by burning it directly, or waiting until heat and pressure from the earth has turned it into fossil fuels over millions of years.  E.g. wind is air movement caused by the differential heating of the surface of the earth by the sun.

Are Electric Trucks Practical?

 I just watched a video where the speaker is arguing that the amount of trucking we have in the US will be impossible to achieve if we electrify it all.  This sounded wrong to me, so here are a few numbers.

 There are about 13M trucks over 10,000 lbs in the US, about 3M of them doing long haul.  The vast majority are doing short haul delivery.  Tesla claims it's semi can do 1.7 miles per KWh.  My model S gets about 2.8 miles per KWh and it weighs less than 5000 lbs, so I suspect that 1.7 is optimistic.    Let's guess a full size long haul truck can actually get 1.5 miles per KWh.  Long haul trucks are limited to about 500 miles a day by driver's hours regulations, which works out to about 333 KWh per day.   So to power all 3M of those trucks would take 1 million megawatt hours or 1 terrawatt hour.

That sounds like a big number and it is.  But here's another big number.   In 2022, the US produced about 434 Terrawatt hours of electricity from wind turbines.  That's about 1.2 TWh per day--about 20% more than would be needed to power all those trucks.  Could we add enough additional wind turbines to achieve this?   Yes, definitely.  Present deployment rates have us doubling every 6 or 7 years.

The other 10 million trucks do not consume nearly that much power.  The vast majority of trucks average much less than 100 miles per day, doing deliveries, moving containers around in seaports, moving concrete, moving garbage, etc. The postal service, Amazon and others have been finding that their fleets of electric vehicles are vastly cheaper to operate than fossil fuel counterparts.   I haven't got enough data to make better than a handwave estimate, but it seems unlikely to be more than 1TWh per day.

Bottom line:  While there are definitely some technical issues needing to be solved, none of them are particularly difficult.  The two biggest are adjusting the power grid to make charging available where it is needed, and dealing with the range issues for long haul trucking.   Neither of these requires a breakthrough in order to make near-universal electric trucks practical, just straightforward engineering.

------

 I am of the opinion that most long-haul trucking should be replaced by trains.  Where long haul trucks operate at a fuel efficiency of 10-20 ton/miles per gallon, trains operate at over 400 ton/miles per gallon, have far fewer accidents per ton/mile, and are easily electrified.   Trucks receive huge subsidies as compared to railroads (mainly in the form of taxpayer-provided roads) yet they are about the same cost per ton/mile.  Trains have a few disadvantages relative to trucks.  All are relatively easily fixed:

* Most railroads are full of bottlenecks that cause random time delays.  They used to have fewer--many routes were single-tracked or eliminated when the interstate highways took away a lot of their markets.  Computerized dispatching recovered some of this lost capacity, but not enough to accomodate growing demand.

* A high percentage of industries no longer use railroad sidings, and a lot of them have been torn out.  Thus railroads need to trans-ship a lot of traffic to trucks.  Switching entirely to long-haul trucks eliminated this transfer.  (an entertaining example:  Tesla's Fremont, CA plant is about 30 miles from where I grew up.  It was originally built as factory for GM, and placed along a major rail route.  Supplies are still shipped into the factory by rail, but the cars themselves are never shipped out by train--they are loaded onto trailers and shipped to their destination.  My Tesla Model S was delivered to a service center that is right next to a railroad track and could easily have been shipped by rail, but in fact, it came on a truck.  Because /some/ SCs are not on the rail.

* A high percentage of rail traffic is in what is called a "unit train" where every car is going from and to the same place.  Containers have made it possible to turn mixed trains into unit trains for a big part of their trip--ship from a big terminal on the east coast to a different big terminal on the west coast, and trans-ship or re-classify from there.  Because they can, the railroads have chosen to make the time this takes burdonsome:  typically well over half of the time a shipment takes to arrive is spent waiting in one of these terminals.

Converting Heat to Rotary Motion

This is the flow of all energy used by the US in 2015, as calculated by Lawrence Livermore Labs.

There's a tremendous amount of useful information here.   For example, wind and solar are now a noticeable part of the grid, and electricity is contributing a useful amount to transportation.

But the biggest thing is the amount of rejected energy.  That's 60.6% of all energy that's being consumed by the US that's being lost due to inefficiency of some sort.  80% of this and maybe more is coming from two uses: electricity generation and transportation.   What they have in common is that they're using the burning of something to convert heat into rotary motion.

Here's an example: a conventional car contains an internal combustion engine, which works by spraying a mix of gasoline and air into a cylinder and setting it alight.  The resulting explosion pushes the cylinder against a crankshaft which connects to the wheels of the car to make it move.  This conversion is at best about 33% efficient--most of the energy--and with it the pollution, CO2 and other stuff, is lost to heat and noise, without producing forward motion.    The same thing is happening when burning coal or gas to produce electricity: it heats something (either boiling water or the natural gas exhaust itself) so that it can be used to power a turbine, which is connected to a generator.  The theoretical peak efficiency of such a process is about 60% and the reality is a lot worse: the very best electric plants operate at about 40% efficiency.  The rest is waste heat.

If, on the other hand, the heat is the ultimate product, for example burning stuff to make heat in a furnace, it is much more efficient.  Modern furnaces are over 90% efficient.  The blast furnaces used in steelmaking are similarly efficient.  

Wind and Solar suffer different inefficiencies, but the waste doesn't end up in the atmosphere as CO2 or smog, and it doesn't consume a resource that takes a hundred million years to replenish.  Electricity can be used to make rotary motion much more efficiently than the reverse, so it's useful for a large fraction of the needed activity in industry and transportation.   according to the chart, .03% of electricity is used for transportation.  This includes electric cars like a Tesla or Leaf, and Trolley Buses and Electric Trains like on the northeast corridor, but does not include hybrids like the Prius or a diesel-electric locomotive.

It seems to me this is fairly low hanging fruit.  60% of our energy production is lost to the inefficiency of using heat to make rotary motion.   The second law of thermodynamics says we're doomed to lose some, but to get half of this back seems relatively straightforward.  Converting half the cars on the road to electric will save around 10 Quads of rejected energy.  Converting half the gas and coal power plants to wind and solar will save  around the same.

Sustainable Rates

Conservatives are fond of pointing out the hypocrisy of protesting oil drilling in plastic (made from oil) kayaks or protesting logging while living in wooden houses.   For purposes of argument, I'll ignore the nonsensical disconnect and pretend that it's a credible point.

Wood is potentially a sustainable resource.  If you harvest it at a rate lower than can restore itself, there is no direct risk from logging.  For much of the 20th century, this was indeed what was practiced, but by the 1960s, demand had exceeded capacity.  Especially in the 1980s under the Reagan/Watt regime, a high fraction of the old growth lumber in America was harvested and sold off, a huge part of it to the far east.

Total worldwide consumption of wood today is about 4B cubic meters.   There are about 4B hectares of forest in the world (down from 5.9B before industrialization.  We've been losing about 13M hectares/yr).    Canadian forests contain about 432 cubic meters per hectare.   Extrapolating that to the whole world, if we were to harvest 1 cubic meter per year per hectare, the forests would thus have 432 years to recover.   My forester friend tells me that if we harvest 10% or less every 15 or 20 years, that can be carried on indefinitely without harming the forest ecosystem--that's more than double that 1 m^3 per hectare rate.  But that's not what we're doing.  We mostly clear cut, in patches of a few hectares at a time, and during the '80s, we basically clear cut alternate quads, leaving a patchwork of tiny forests. This will recover eventually, but it'll take hundreds of years, and in the meantime, we've wrecked the local ecosystem, which will make it take longer to recover than necessary.  Worse yet, we're permanently clearing areas for agriculture, industrial and residential use.

Lots of privately owned forests are harvested sustainably.  We know it can be done.  We need for the government to protect our public forests this same way.



Total proven oil reserves are about 1.3T bbl.  Consumption has roughly tripled in the last 50 years, from about 11B bbl/yr in 1965 to about 33B bbl/yr today.  This suggests that the total amount of oil that was theoretically available at the start of the 20th century was about 2T bbl.    That oil took about 200M years for the earth to make.  2T/200M = 10,000.    So a sustainable rate for oil is about 10,000 bbl/year.    We are consuming it at 3 million times that rate.  


World coal production is about 8B tonnes.  The best guesses I can find say we can keep this up for about another 100 years, which suggests that there are just about 1T tonnes left and there were probably about 1.5T to start with.   Like oil, hard coal takes around 200-300M years for the earth to create.   That means we can sustainably harvest about 5000 Tonnes per year.  But there's a difference.  Soft coals such as lignite can form much more quickly, in the order of 5M years.  That might give us a sustainable rate for those coals of 250,000 Tonnes per year.  We use between 32,000 and 1.3M times too much to be sustainable.

Birds Killed by.....

During the campaign, Trump railed about windmills, emphasizing the birds killed by them and the supposed hypocrisy of liberals for wanting windmills despite this problem.,   At his rallies, he often expressed it as "killing all the birds".  So I did a little research.  All these numbers are bird mortality for the US

0.2-.35M     Windmills
1-2M         Oil and wastewater pits
15M          Hunters
60M          Collisions with cars
72M          Pesticides
174M         Collisions with power lines
400-900M  Collisions with windows
1400-3700M     Cats
So windows are at least 1000 times more dangerous to birds than are windmills.  And cats.   We need to ban cats and windows!!!! 

Sheesh!

Of course we should do everything reasonable to minimize all of these, but the idea that it would be a reason to not build windmills is foolish.  The oil business kills far more.

The  real reason he doesn't like windmills is because some would be built in the view of one of his resorts and he thinks that would reduce the price he could charge (maybe, maybe not),   but more importantly, he is beholden to fossil fuel interests and wants to drag his feet as much as possible on the inevitable switch to renewables.

He also claimed that all of the wind turbines come from Germany and Japan.  Uh, no, there are several wind turbine manufacturers in the US, and GE owns wind companies in several other countries.
China  10
South Korea 6
Germany 4
US  3
Japan  3
Spain  2
Denmark 1.  but it's by far the biggest

There were 88,000 wind power workers in the US at the start of 2016, growing at about 20% per year.  The American Wind Energy Association estimates that there will be 380,000 wind power jobs in the US by 2030.  The vast majority of these are in rural areas or semi-rural areas.  There were 63,000 US jobs in the coal business in 2015 and this number can only decline.  Virtually all coal mining is done in strip mines and mountaintop removal, which substitutes big machines for workers, and there is nothing Trump or anyone else can do to change this.

addenda 10 Mar 2017
Predatory birds like eagles are particularly susceptible to poisoning because many of them are also scavengers.  A particularly serious problem is birds and animals killed by hunters and abandoned, either because the hunter only took enough to make a trophy, or because they couldn't find their victim. Eagles and other scavengers eat the decaying flesh, including the lead shot.  Lead is a slow neurotoxin, so they don't die immediately, but their thinking and coordination is damaged, and they often crash into things that they'd have otherwise been able to avoid.

Eagles and other raptors are also particularly susceptible to being hit by windmills, because their hunting methods may put them at the same height as the moving blades.

September 11th

15 years ago today, 19 Arab guys, 15 from Saudi Arabia, 2 from UAE, one each from Egypt and Lebanon, funded by a wealthy scion of the Saudi regime and partly inspired by the same Wahhabi fundamentalist crackpotism that inspires the Saudi regime, attacked America and killed about 3000 of us.

Our initial response was to invade Afghanistan, which was a failed state and not involved in the attack, but some of the warring parties allowed some of bin Laden's other projects to train there and it appears that bin Laden himself was there for a time.  It didn't really address what had caused the attacks, but it made a little progress against Afghanistan's real problems.  But like our efforts 20 years earlier, after the first superficial successes, we gave up on the real solutions (e.g. schools and a way for ordinary Afghans to make a decent living) and moved on to making yet another failed state in Iraq.

All together, our response to the September 11th attacks have killed more than 3 times as many US and allied soldiers, wounded far more than that, killed hundreds of thousands in the countries we've attacked, exacerbated the problems of Afghanistan, created newly failed states in Iraq and Syria where there had been stable states before, and created a far more deadly terrorist organization in ISIS than bin Laden's al Qaeda ever was.

By any standard I can think of, the "war" on terror has given a huge victory to the terrorists, and to the cause of nihilism in general.

The Saudi regime in Arabia is extremely repressive, and their huge wealth allows them to stay in power in spite of it.   The Arab guys who attacked us 15 years ago knew there was no point in attacking their oppressor.  As long as the money keeps rolling in, the family of ibn Saud will use it to retain power.  So they attacked the enabler of their oppression.


Meanwhile, we have been quietly increasing development of US petroleum reserves.  Corrected for inflation, the price of gasoline is about where it was in the mid '70s, down from all-time highs.   The Saudi are feeling the bite, but they're still making good money.   But more important than that, we've been installing renewable energy sources, mostly wind and solar, at a rapidly increasing rate.   In another decade or so, we may be able to stop buying buying from them entirely.  Once that happens, and the profligacy of the Saudi family burns through all that money, there will be regime change in Arabia.

That will be a victory for all of us. 

Renewable Energy

There are three ultimate sources for energy on the earth, none of them renewable.  One of them, solar nuclear fusion, deposits so much energy on the earth that its consequences are left behind for millions or even billions of years.  Whether a sources is renewable or not is really whether these consequences can provide meaningful amounts of energy without causing environmental consequences or using the resource faster than it's produced.  I'll get into the two ultimate sources first.

Nuclear fission:  All chemical elements have some degree of of nuclear instability and spontaneously decompose, emitting sub-atomic particles and energy.   Leo Szillard figured out how to use the sub-atomic particles to create a self sustaining chain reaction, generating a lot of heat.  At controlled densities, this heat can be used to run power generating equipment, such as steam turbines.  At uncontrolled densities, it's an atomic bomb.   The atoms involved are destroyed, but there are plenty of them around so renewability of the supply is not a serious problem. But the process generates a lot of dangerous radioactive waste.  Presently, there are no really good solutions for dealing with this waste, although there are some that are not extremely terrible.

Geothermal compression.  Gravity causes the materials in the interior of the earth to be compressed, which causes them to heat up.  Most of this stays deep inside the earth, but a little comes out in volcanoes, geothermal vents and a few other things.  This heat can be used to run power generating equipment, such as steam turbines.   The total supply of geothermal energy is finite, but far larger than present consumption rates could use for millions of years.   As far as I know, there's no real downside to using geothermal energy, but there just aren't very many places where it's near enough to the surface to be of much help to energy needs.  In order to drill our own geothermal wells into the earth's mantle would require some big breakthroughs, none of which show any signs of happening.  For the time being, we pretty much have to rely on earthquakes and volcanoes to do it for us.

The third, and far and away the most important, is Nuclear fusion in the sun, causing lighting and heating of the earth.  The sun is finite, and will eventually burn itself out, but this won't be for billions of years. Radiation from the sun strikes the earth at about 1 kilowatt per square meter.  average US energy consumption is about 3 billion kilowatts.  The US is about 10 trillion square meters, so theoretical capacity is about 3000 times what we need.  No means of collecting that energy is very efficient, but that's a big multiplier.  Ultimately, the question of renewability is whether we can extract energy in a way that doesn't wreck the environment.  Here are a few:

There are numerous ways of using solar energy fairly directly.

Photovoltaic cells (mostly silicon but there are others) can be exposed to the sun and turn about 15% of the energy received into electricity.  This are almost completely renewable and will work just fine for as long as they aren't damaged or obstructed, although some of the chemistry in making the cells is nasty.

Solar heating: There are several ways of doing this.  The sun can be used to heat blackened tubes to generate hot water, which is a useful commodity, and can be used for bathing, washing, or radiant heating.   The sun can be focused and used to generate something hot enough to run generating equipment, such as steam turbines.  These are completely renewable, but a little fussy to manage.

Hydroelectric power: the sun heats bodies of water, including the ocean, and evaporates it.  Some of this later precipitates and fills other bodies of water, at higher elevations, where it can be run through a turbine to generate electricity as it falls to lower elevations.  Modulo the environmental damage involved in redirecting the water, this is completely renewable.

Wind Energy: Differential heating of the earth causes air to move from warmer places to cooler ones.  We call this air movement "Wind".  Windmills can be put up to collect electrical (or mechanical) energy from this.  This too is completely renewable, modulo the environmental consequences of building the windmills.  There is also a tiny but real risk that we affect the movement of air.  Windmills have also been known to injure flying animals, such as birds and bats, that can't recognize the fast moving blades.  The harm from this is considerably less than from fossil fuel plants.

Biofuel.  Ultimately, biofuel uses energy from the sun to cause plants to grow.  The complex molecule chlorophyll uses the suns energy to convert carbon dioxide from the air, and minerals and water from the ground, into plant material. There are many biofuels.  The most familiar is wood.  Trees can be harvested and burned, generating heat directly, or used to run power generating equipment, such as steam turbines.  Whether this is renewable or not is whether the trees are harvested at a rate that damages the forest or allows it to continue to be as healthy and large.   There is a lot of wood that is cut down and thrown awaybecause it's the easiest thing to do with it.  A lot of this could be turned into energy.  There are many wood-products industries that generate all the energy they need from wood waste.

There are numerous other biofuels.  The one in largest use is Ethanol, which in the US is mostly from corn.  Because the government subsidizes corn farmers, some of the surplus is fermented and distilled into ethanol.  Most of this gets added to gasoline.    This technically renewable but it is a terrible waste of agricultural resources and taxpayer money.   Professor Patzek calculates that it takes between 4 and 7 times as much fossil energy to create the ethanol as it produces, which makes it a big new loser on the renewable scale.

Other biofuels include converting dirty cooking oil into biodiesel.  This is certainly renewable: the cooking oil is an agricultural product, which probably was sun and carbon dioxide less than a year prior to it being turned into biodiesel, but the total supply is very small.

Another which seems promising until you do the math is microbes which secrete, or can be turned into, some sort of petroleum-like product.  Again, this is completely renewable, on short cycles, but chlorophyll is a very inefficient converter of the sun's energy to usable energy:  around 1/10th of 1% at the very best.  If we were to use the entire area of the US (including alaska) to generate biofuel this way, it would not be sufficient for present energy needs despite that amazing 3000:1 multiplier.  Their advantage is once the process is started, it's very low maintenance, but Photovoltaic is way more efficient.

Fossil fuels are technically biofuels and renewable, but the time it takes to renew it is in the hundreds of millions of years.  Plant matter in large quantities lived, grew, and died, over thousands or millions of years creating big piles.  When they're in the process of being created, they tend to be called things like marshes or peat bogs.  Eventually they get covered up by hundreds or thousands of feet of earth and over millions of years they turn into fossil fuels.  Whether it's coal, petroleum or natural gas depends on the exact temperature, pressure and chemical composition.   Coal stays where it was formed, but petroleum and gas are lighter than earth and stone, and if there's a path to the surface, they float up and eventually are lost into the atmosphere where the sun causes them to break down into CO2 and water.  When there's a geologic formation on top that causes it to be trapped, we can drill into it and recover the petroleum or gas.  Most petroleum was created around 300 million years ago, but over hundreds of millions of years.  By burning fossil fuels, we're emitting hundreds of millions of years of CO2 storage in a few decades.  Not renewable.

Recharging While Driving

Electric cars are here, and are already having a significant impact on the roads.  But they still have a problem with range.  The small ones: the Leaf, the BMW i3, etc., have a range which is fine for the typical daily commute, and even for a lot of delivery services.  But recharging in most cases can only be done at 20 miles of range per hour.  But for occasional longer trips, they remain a problem, and you have to spend several hours on the charger for every hour you drive.  Not adequate if you're taking a trip that's substantially longer than the range of the battery.   So far, Tesla has the most complete solution: 1) a much bigger battery than the others, giving 250+ miles of range.  2) faster "destination" charging, as much as 60 miles of range per hour, so you can put those 250 miles of range on in only 4 or so hours, allowing two or three cars to use the same charger over the course of a night, and finally, the Tesla "supercharger", which can put 150  miles of range on 20 minutes.  Presently, these are located about 150 miles apart on most of the big cross country routes in the US, but there are still a lot of gaps.   Tesla has promised to get around to them, but it'll be a while.

The second best solution is to make some sort of fuel-electric hybrid.  there are a number of plug-in hybrids available, but all have a very short range.  If your commute is very short and you're careful to plug in every night, it may work for you, but for long range travel, you're just a relatively efficient internal combustion engine car.

My idea is to put power rails along the roadway.   Whether these have inductive pickup or are actual physical rollers that run on rails is an engineering decision.  I'd put these on the interstates--the same routes where Tesla is putting its superchargers--and reserve a lane for them.   The new Teslas come with the ability hold lane, speed and following distance on their own (although lane following is still in alpha), and this is all it would take to have the car drive itself on an accurately delineated charger lane.   Many other carmakers have similar features available.  There is no particular reason that following distances couldn't be short and speeds high, as long as the computer is doing the driving.   The car will arrive at its destination city fully charged, and the driver rested.

The Tesla Model S charging at 19.2KW, charges at 60 miles of rated range per hour, so this is what it would need to maintain steady charge state at 60 mph at the rated characteristics.  This is probably all that's necessary.  The car draws less than that while going straight and level, but more accelerating or going uphill.  The battery can handle surges--the power rail is just for topping up.  It needs to be arranged so that all the cars needing to be charged can have their own charger.  Simplest would be for each car to tap into the same (say) 600VDC rail pair simultaneously and regulate itself so it never draws enough power to cause the voltage to drop more than a small amount.  If more power is needed, it reduces its charge rate and draws from its own battery.

Issues:
The power rails would probably need their own substation few miles or so.  Highly loaded, this would be 50 cars per mile or so. 19.2*50=960... a megawatt.   Those big, 200+ foot wind turbines put out 1.5 to 3 MW at their peak, so having one every  mile or so would work nicely when the wind is blowing.   A strip of solar panels along the road 6 meters wide (20 ft) would do the same thing when the sun is shining.  (there's a company that's trying to sell solar panels that work AS road surface...I think you could make this work by paving the special lane with such things)

If the power available is exceeded, the cars can run on their batteries, but they should probably be aware of what's happening.  If it's a traffic jam, they'd probably need to slow down and this will automatically reduce the power drawn.

600VDC with the availability to deliver 1600 amps would cause major badness to occur to any conductor that happened to fall across it, including a person or animal.  the conductors either need to be protected somehow, or buried with inductive pickup.

there should probably be some mechanism to keep non-robodriven cars out of the special lane.   I think the double white lines and reflective bumps as used on toll lanes, are not quite enough.  There is no real safety problem with illegal use of a toll lane.  There is with power rails and high-speed robocars,

billing should be straightforward.  the car would have a tamper-resistant meter which monitors the number, location and timing of watt/hr picked up and communicate it with substation, which would bill appropriately, and prevent incorrectly equipped cars from driving on the road.  If it becomes a problem, substation can monitor load and correlate it with the traffic data to detect tampering.  I suspect it won't be...once the roadside infrastructure is built, it'll last for decades and cost very little to maintain.   It'll pay itself off in a few years.

heavy trucks should be able to use this.  they'll obviously draw more power.  the road design can handle this straightforwardly.  more megawatts per mile.

because all the cars are robodriven, including the heavy trucks, all the vehicles will run down exactly the same part of the road, and rutting will be a worse problem than it will be on roads with human drivers.   Since there's a parallel road and all the vehicles have batteries, repairing it will be pretty much the same as with present roads.

Electric Car Chargers

Electric cars have finally crossed the line from being futuristic possibilities, through hybrids that exploit some of the benefits of electric power but still burn petroleum, to fully electric vehicles.  Tesla and Nissan are the big trendsetters, but now pretty much all the carmakers are on board.

Charging an electric car presents some special problems, mainly stemming from the fact that moving the mass of a car (and passengers) requires quite a bit of energy. Gasoline (and its relatives) have tremendously high energy density.  The breakthrough that's made electric cars finally feasible is the lithium ion battery, which is both smaller and lighter than it's predecessors.  But it still takes much longer to charge a battery than it does to fill a tank with gasoline.

Broadly speaking, there are three types of charger that have evolved, which go by the name "Level 1", "Level 2" and "Level 3".   The first two are not really chargers, but EVSEs "Electric Vehicle Service Equipment".  They just pass the power unchanged, with the addition of some safety equipment.  The charger is on the car.  A level 1 EVSE uses one phase, line voltage power.   In the US, this means 120V, 15 or 12A.   The SAE definition is that a level 1 EVSE is single phase and not capable of charging at faster than 1920W.  A level 2 EVSE is 2 phase, 208 to 240V and may go at up to 80 amps, but very few of them actually do that.  This is a theoretical peak of 19200W, but very few go higher than 7800.   "Level 3" is not defined, but in practice, it's anything that goes higher than 19200W.

The standard EVSE system is SAE J1772, which defines a 5 conductor connector, which has two power conductors, a ground, a "pilot" and a "proximity" conductor.  The latter are used for safety signaling.  The "pilot" carries a 12V 1KHz square wave which is used to detect continuity, and the "proximity" signals to the car that it's plugged in, and blocks the vehicle from being driven until it's disconnected.   Power is just passed through on the three power and ground conductors, but a relay opens the circuit unless pilot and proximity are deemed valid.  The idea is to make it impossible to charge a car with a kludged up power cord, just the sort of thing that might start fires, and to reinforce it, they tell you not to use extension cords when you buy the car. Virtually all electric and plug-in hybrid cars support J1772, although a few, such as Tesla S, require an adapter.  (Tesla exploits their incompatibility to use the same conductors for their "supercharger" level 3 scheme.)

The electric code also requires that lines being heavily loaded for more than half an hour only carry 80% of their design capacity.  This is also to prevent fire--heat builds up gradually, and your charger is likely running while you're sleeping.  This is the reason that a charger wired for 15 amps is only allowed to draw 12, 50 amps only allowed 40, and so forth. 

Level 3 chargers typically use special power arrangements to draw more than 80 amps.  Generally they are called DC chargers but that oversimplifies what they're doing.  They require smart electronics and high power connectors--generally beyond the capacity of home chargers The three main protocols are the still evolving J1772 level 3, the CHAdeMO, and the Tesla Supercharger.   The first two require a very special connector, while the Tesla uses the same connector as level 1 and 2.  (this is the main reason Tesla went its own way rather than allowing J1772 to plug in directly).  BMW and Nissan tried to negotiate a license to use the system during summer 2014, but BMW recently announced that it's building its own network of level 3 chargers using the J1772 "combo" connector, which is a kludge involving a J1772 and a separate, higher capacity connector in the same plug.

Types of Electric and Hybrid Vehicle

Ultimately, there are two ways to power a car:

Fuel, usually derived from Petroleum but sometimes agricultural sources or coal (namely Fischer-Tropsch), burned in an internal combustion engine, or Electricity, either generated in the car with an internal combustion energy or fuel cell, or stored in a battery--or both.  Since current batteries have low energy density compared with fossil fuel, all modern electric cars use various means to minimize or recapture electricity, most importantly, regenerative braking.

There are relatively few Pure Electric vehicles on the market although this is changing rapidly.  These must be charged by plugging them into a source of electricity.  The people with electrics I know seem to fall into two classes: about half have installed high power--7KW or more--chargers.  Tesla has a supercharger: 120KW, but as far as I know, nobody has installed one in their home.  The other half have not bothered.  They use ordinary 120V power to charge their cars and do not find it to be a problem.  In all cases, their daily commute is less than 15 miles.  (Most 120V chargers limit power draw to 1200W.  I suspect this is due to concern over sketchy wiring and long extension cords.)

Pure Electric cars include the Tesla (all models), and Nissan Leaf, the Smart ForTwo EV, quite a few others.  The runaway success of the Tesla has provoked nearly all the big carmakers to follow suit.   Most have range well under 100 miles, although the Tesla can go over 200.

The most common type of EV by far right now is the Parallel Hybrid.  All of these can be powered either directly by the internal combustion engine or directly by the electric motors, running off the battery.  This involves a complicated transmission.  The IC engine can power both the transmission, and a generator.  Most are have fairly limited range and top speed under electricity.  the battery and generator are usually quite small.

Nearly all of the parallel hybrid makers are making or soon will make a Plug-in version, which allows you to top up the small battery.  They generally haven't added enough battery to give much all-electric range.  There is a substantial aftermarket for larger batteries and plug-in circuitry.

Parallel Hybrids include the Prius and related Toyota products, such as the Camry, the Ford Focus Hybrid, Most of the Honda hybrids, and quite a few others.  The Chevy Volt and the new Cadillac ELR are technically Parallel Hybrids although they are mostly Series Hybrids.  I'll get into this later.

A Series Hybrid is mostly a pure electric vehicle, but it carries an on-board engine and generator.  This is much simpler and more efficient than a parallel hybrid, but for some reason, most of the carmakers didn't go that way.  I think they were concerned about the size and reliability of batteries when the first hybrids were being introduced.  The power of a series is constrained mainly by its battery: a bigger battery means you don't need as big an engine. 

The most important series hybrid on the market, the Chevy Volt, was apparently felt to not have quite enough oomph going up long hills, but instead of improving the battery, they put in a bigger engine and made a clever clutch device which engages the transmission with the engine when the electrics weren't enough.  This technically makes it a parallel, but as long as the clutch isn't engaged, it's a series.  I believe most buyers of Volts would have preferred to have less power on hills and longer electric range.  Most of my friends who own them rarely if ever drive fast or far enough to engage the engine.  The new Cadillac ELR uses the same mechanism.

GM has another variation, which they've dubbed a Mild Hybrid.  This is an entirely normal internal combustion car, with an oversized battery and starter motor, and a small amount of extra electronics.  When you stop the car, at a traffic light for example, the engine stops.  When you press on the gas pedal, the starter starts the car moving while it's restarting the engine.   I don't think any of them bother with regenerative braking or pure electric while driving at low speed, but if you spend a lot of time idling, this could be a fuel saving.  But it's mostly marketing.

A solar powered car that resembles the performance and comfort of today's cars is impossible.  Noonday equatorial sun on a clear day is about 1000W/square meter.   The total surface area available for exposure to the sun is about 7 square meters on the typical car.  Due to inefficiency, and the fact that most of the time it isn't noon on a clear day, and most of us don't live near the equator, that means we have an average of about 500 watts if the sun is up.  That's about 2/3rds of a horsepower.  even if somebody came up with magic 75% efficient solar panels (today's best are under 40% and most are about 15%), that would only bring it up to about 5hp...still not enough.   But that doesn't mean you can't use solar panels to help recharge the battery of your plug-in.

US Energy Flow

Two charts, produced by Lawrence Livermore Labs: https://flowcharts.llnl.gov/.
This is an attempt to characterize all the energy consumed by the United States.  I find this a very helpful way to look at energy use and production, and endlessly fascinating.

Some definitions and observations:

A "Quad" is a Quadrillion BTU.  1 Quad is about 293 Trillion KiloWatt/Hours.

"Rejected Energy" is that lost to inefficiency, including heat and entropy.

Total energy demand has shrunk by more than 4% in those 4 years.   For the preceding decades it had only risen, flattening in the mid 2000s.
 
Solar and wind have substantially more than doubled in 4 years.  Those numbers are on track to double again by 2016, but 2012 is the most recent year for which LLL has produced these charts.

Coal as an energy source has gone down by about 5 Quads, over 22%.  Natural Gas has only increased by 2.2 Quads.  Almost all of this is in the generation of electricity. 

Electric transportation is only .02 Quad, rising slowly.  I'm guessing the largest share of this in these reports is stuff like subways and trolleys.  We have seen a big increase in electric cars since 2012...I suspect this will see a more visible increase in the 2013 report once it's available.

I'm surprised at how much Biomass is used in industry.  I suspect it's substantially burning of agricultural waste.  The use of ethanol in transportation is visible.

I'm not sure why Rejected Energy has increased while total demand has dropped. The largest piece of this (over half) is electrical transmission losses and generation efficiency, but that's improved.  Most of the gain "Industrial". 

Hydrogen Cars

Hydrogen fuel cells sound to some like a promising technology: it allows an electric car which can be refueled quickly and its only emissions are water.  But it's up against some laws of physics that make it tough to be practical.  The main one is that the energy density of hydrogen is very low.  Liquid hydrogen has an energy density 1/4th that of gasoline.  This means to get a car with the same range, you'd need a gas tank 4 times as large.  But liquid hydrogen is incredibly expensive and the tank would be impractical.  Carrying it at much higher than about 3000 PSI (roughly what is used in a SCUBA airtank) is probably impractical, but at that pressure, the energy density is about 1/20th that of gasoline.  To replace the the 11.9 gallon fuel tank and internal combustion engine in a Prius with fuel cell and a 3000 PSI hydrogen tank and get the same range would take a tank of around 250 gallons.  This is a volume of about 33 cubic feet.  This is roughly the entire interior volume of the Prius....if you take out all the seats!

The bottom line here is that without some significant technological breakthrough (for example extremely high pressure tanks, or some scheme that where the hydrogen is dissolved in some other material at much higher than 3000 PSI equivalent), a hydrogen car isn't likely to have much range.   A bus or truck is a different story: it has a lot more places to put the tank(s).   Hydrogen does have one big advantage over gasoline or battery electric: the mechanism, tank, fuel (even 250 gallons), fuel cell and electric motor, is a lot lighter.  This may help with mileage to some extent but nothing like enough to give it the range we've come to expect.

Nevertheless, Toyota has just announced the Toyota FCV, a Camry-sized Hydrogen fuel cell car to be put on the market in 2015, which they claim has a 300 mile range and can be refueled in 3 minutes.   I'm skeptical.   None of the pictures show the interior or even say how many passengers it carries.  It pretty much has to be a bunch of hydrogen tanks on wheels, with very little room for the occupants.


In most places the cheapest way to make hydrogen involves heating natural gas with steam to split the hydrogen from the CO2.  This produces as much CO2 as burning it (even more if you burned something to produce the steam) and is at least as bad for the CO2 levels in the air as other fossil fuels.  It is NOT a green fuel, it is a fossil fuel, with all the problems that come from it.  The CO2 isn't coming out of the tailpipe of the car, but just as much (more actually) is coming out of the plant.

In some places, where you have extremely cheap electricity, you can make hydrogen with electrolysis of water. It takes more energy to do this than you will get out of it, so it makes the hydrogen a sort of storage battery, but because of the ease of refueling and the light weight may give it an advantage.  Iceland, with abundant geothermal electricity, is such a place, and as the whole island is only 250 miles across, a low range car is pretty practical.

Hydrogen is no more flammable than gasoline, although if punctured, a high pressure tank would spray a stream of hydrogen until it's exhausted.  If it ignites, that's potentially a very hot, very focused flame. 

Ethanol

For a little while there, ethanol was being pushed as the replacement fuel of the future.  The corn lobby was the biggest advocate, but other environmental-seeming groups jumped in.  The future of the automobile seemed bright, and even though corn was quickly recognized as not optimal, switchgrass and other approaches seemed popular for a while.  It looks like the hype has now died off, which is a good thing.  Much of the support was coming not from people concerned with improving the environment, but from front groups sponsored by fossil fuel companies.

Professor Patzek puts the energy inputs to produce corn ethanol at about 7 times the energy produced.  Today, nearly all of this is from fossil fuels, mostly petroleum and natural gas.  These include agricultural inputs--fertilizer and pesticides, (which are substantially petroleum products) fuel to operate machinery, harvesting, processing (distillation being the big one), and transport costs.  Ethanol is too volatile to be shipped in pipes--it has to be shipped in tanks, on trucks or trains.  This is also why gasoline with some ethanol in it--typically E10--goes bad after a few months.   All of these inputs are ignored and ethanol is treated as a zero emissions biofuel, for computing subsidies and tax breaks, over and above the crop subsidies for growing corn.  This clearly not true, and is bad policy, at many levels.  The process could be improved: no-till agriculture eliminates most fertilizer and pesticides and some of the other fuel consumption. Agricultural waste could be burned to power some of the distillation process (for corn, there's not enough of it to do it all).  Converting to switchgrass would improve much of this.  But even making the most optimistic assumptions, the energy required to produce ethanol is about four times the energy produced.

The bottom line is that ethanol, especially corn ethanol, is a boon to the fossil fuel and corn companies.  The subsidies have distorted the market, and the corn and fossil companies have put up a big environmental-appearing front to try to improve their own profits.  But it's a fraud.

Steam Engines in Space

Much current power generation is done with steam.  Water is confined in a "boiler" and exposed to some heat source (burning coal, nuclear fission, concentrated solar energy, etc.) and boiled.  This produces steam, which is passed through a pipe to a turbine or reciprocating piston(s) to produce rotary motion.  Usually this is used to power a generator, but sometimes the rotary motion is used directly--a steam locomotive for example.

An important but subtle component of a steam locomotive is what's called the "steam dome".  This is a dome on top of the boiler, where the steam collects.  It relies on gravity--water stays in the boiler, while the "dry" steam floats on top and in the dome.  It's important to keep water out of the steam tubes and pistons: it cools the steam, reducing pressure, and if there's enough of it, can prevent the piston from moving and break something.

As far as I can tell, conventional steam locomotives need gravity or something like it (e.g. centrifugal force) to operate.  There are heat engines which do not require this separation of dry steam, such as the Stirling Cycle, so it's still possible to produce rotary power in space.  It's also possible to make a direct ejection steam rocket which could operate in zero-g.

Are Electric Cars Really Green?

I've been seeing several commentators point out that to deliver power for an electric car, it has to go through several conversion phases, each of which add substantially to the inefficiency of the process.  This is absolutely true.  For example: burning natural gas to produce electricity produces at best about 60% of the theoretical conversion efficiency.  Realistically it's closer to 40%.  Line drops add 5-20% further loss, and charging/discharging the battery adds another 5-20%.     Burning coal or petroleum is even worse--30%.  Meanwhile, a Prius is able to turn gasoline into motion at about 30%.    So: picking relatively favorable numbers for natural gas, we get .4 * .9 * .85 = 30.6% conversion efficiency from the carbon in the natural gas to the motion.  The commentators are right: there's no efficiency advantage in a pure electric.

But this omits the really big point, which is that it's possible to power a pure electric car from completely carbon free resources: Hydroelectric, Wind, Solar, Geothermal, etc.  This is not true for a Hybrid, like the Prius.

Another complaint: There are a lot more batteries and they're harder to recycle.  This is true too.  Lithium and NiMH batteries are harder to recycle than lead acid batteries.  (Lead car batteries are over 99% recycled).   But several companies (including Toyota) are doing it fairly successfully.   There's another win too: even if a lithium battery does get into the landfill, it's far less toxic than a lead battery, and nowhere near as bad as the Cadmium from a NiCd, which can poison a compost heap at extremely low concentrations.

Another: the glass, brass and tungsten from incandescent light bulbs is worth recycling, but the largest component, the glass, is close to chemically inert.   CFCs (and other fluorescent bulbs) have about the same amount of glass, but they have a bunch of other stuff, all of which is easily recycled.   The most important, by far, is the small amount of Mercury, which is nearly as harmful as Cadmium. (The amount of Mercury in a Fluorescent bulb is tiny though, compared to the amount in a NiCd battery).  Home Depot and many others will recycle these.  I keep a box for this purpose in my garage.

Incandescent bulbs and heating

One of the popular complaints about the push to ban incandescent light bulbs is that a lot of people have been relying on them for heating.  Pretending that all the heat emitted is captured for heating (probably not a good assumption--entropy can never be zero), let's do an analysis.  (Wikipedia  does a simpler version of this)

Here are the cases:

* If you're cooling the building, the more heat the bulb emits, the worse.  Period.  Better to keep the heat out of the building.  Switch to lower heat bulbs, like LED or CFC.

* If you're heating the building with energy that costs the same as electricity, i.e. a resistance heater (e.g. electric baseboard or space heater), the thermal efficiency of incandescent lightbulbs and the resistance heater are very similar and the energy use/cost are thus identical.  Light bulbs, however, burn out fairly quickly while electric heaters last a long time.

(One factor that may be important:  The room may not have a separate heater, and the light bulbs were emitting enough heat to keep it comfortable.  Switching to low-heat bulbs may undermine this, and replacing them with an appropriate heater may be more trouble than the energy saving)

* If you're heating the building with something that costs more than electricity (e.g. gasoline), the light bulb is cheaper.  You should probably think about getting a resistance heater.  (this may not be practical in all cases, e.g. on a boat at sea)

* If you're heating the building with something that costs less than electricity (basically everything except gasoline...or burning money...), then you can lower your heating bill (and environmental impact) by switching to lower heat bulbs.   Here's a nice calculator that enumerates quite a few of the cases.  In case it goes away, here are the numbers at at the time of this writing:

Energy Source	Cost/MBTU
Electricity@ $0.09/kwhr	$26.37
Electricity@$0.12/kwhr	$35.16
Electricity@$0.046/kwhr	$13.48
Natural Gas	$17.38
#2 Fuel Oil	$33.25
Gasoline@$3.75	$41.25
Hardwood	$16.66

The prices of these commodities varies a lot.  12 cents is the national average for electricity.  Puget Power charges 9 cents here in suburban Seattle.  Seattle City Light charges 4.59 cents.  (30 years ago, PP charged 3/4 cent, about what SCL charged then.  Thanks, WPPS.)

Heater efficiency is complicated.  These days, natural gas furnaces over 90% are common.  Electric heaters and light bulbs are a little higher, but not much.   Since these are the major cases we're comparing with light bulbs, we can ignore the inefficiency differences for the most part although  the calculator (and the numbers above) factor these numbers in.

Heat pumps are an interesting special case.  Because their calculated efficiency is usually higher than 100% (COP>1...it's sounds like a second law of thermodynamics violation but it's not: they're stealing the heat from elsewhere) it's a significantly better use of electricity than a light bulb or resistance heater.

Bottom line: there's really only one case where it makes sense to stick with incandescent bulbs:  where other types of heater are impractical for some reason, and incandescent bulbs provide sufficient heat to keep the room comfortable.  In all other cases, lower heat bulbs use less energy

(There may be other reasons to stick with incandescents--for example, the aesthetics may be preferred, but that's not my topic for today)

Recharging while driving.

Electric cars are here!  They have many advantages over fossil-fuel powered cars, but they don't have much range.  If I want to go to my father's house in California in my gasoline car, it's about a 14 hour trip and it takes 3 fill ups--and a night in a motel somewhere Oregon.  That's a lot of gas and a lot of driving.  The electric cars we have today just can't do it.  Even the best batteries would require about 10 rechargings for that trip, and each recharge requires several hours of waiting.  Hydrogen has similar range, although with slightly faster rechargings.  It's up against some laws of physics that probably prevent it from getting much better.

Plug-in Hybrids like the Chevy Volt are a step in the right direction:  it carries a small engine that runs in a narrow, hence efficient, power range to recharge the battery while driving.  It gets much higher mileage for a trip like mine.  But for daily driving, it can run almost entirely from the electric grid.

Battery replacement is also appealing, but batteries are very expensive and in order to make it practical, BetterPlace has to subsidize replacement with very much higher prices on recharging.

There's another option that I haven't seen discussed much.  That is to recharge the battery while you're driving.  The way this would work is a power rail, something like the third rail we're familiar with for subways, would be built into or alongside the freeway.  The car would have an attachment to pickup power from the rail.   If you're doing a long drive, you'd tie into the power rail and your electric car would be recharged.

I think the way to do this is to build the rails into a lane of the interstate highways.  This has almost exactly the right usage characteristics:  they connect all our urban areas and many suburban and rural areas between. The electric car has plenty of range within the urban area, and if not, it will probably pass through one of our special roads en route.

I have several ideas for how this would work.  For example, a gizmo could be added to the bottom of an electric car such that when it detects the rail beneath, it drops a thing that would make contact with the rail.  when you leave the special lane, it disconnects and withdraws.  It would be entirely automatic.  Billing could easily be handled through the same sort of automated system in use on many toll roads today.

Another idea would be to have the vehicle go into auto-drive mode. The driver would surrender control of the vehicle and the vehicle would engage the rail and drive itself.  This has some appealing advantages:  The driver could sleep or read during the trip.  Because the vehicle can know exactly what the vehicle just ahead is doing, it could safely tailgate, which would significantly reduce energy use.  This would also allow much higher speeds.  My 14 hour trip could probably be done in 8.   But there are some problems too.  Most people are still pretty uncomfortable with robocars.  Teaching people how to engage and disengage safely could be tricky, especially if the pack is moving at over 100mph.  Because all of the vehicles will be following exactly the same track, wear on the pavement will be concentrated along that track rather than spread out as it is today.

This second version is essentially the same as an idea called Dual-Mode .  One big difference is that existing dual mode designs require a special vehicle.  I'm just proposing a special gizmo on an otherwise normal electric car.

There are lots of issues to be worked out: the design of the rail so that they're safe (imagine a car dragging a muffler causing a short between the power rails).  The design of the interface.  Convincing the highway departments to do it.  Etc.