A solar panel on a robocar

I often see people say they would like to see solar panels on electric cars, inspired by the solar-electric cars in the challenge races, and by the idea that the solar panel will provide some recharging for the car while it is running and without need to plug it in.

It turns out this isn't a tremendously good idea for a variety of reasons:

  • You're probably not going to get more than a couple hundred watts of PV peak power on a car with typical cells. Even properly mounted on a roof in a sunny place like California, each peak watt delivers an average of about 5 watt-hours in a day, so 200 watts gives you 1kw-h. That's good for around 4 to 6 miles on today's electric cars. Not a huge range boost.
  • While thin film panels don't weigh a lot the power they provide during actual driving would normally be only a minor boost. My math suggests they weigh more than the battery for the power they will provide while operating.
  • Panels on a car will instead be mounted flat, cutting about 30% of their output. Normally you want to tilt to the angle of the sun.
  • Cars are often in the shade, even parked indoors. Unless you work to pick your parking to have sun all day, you'll only get a fraction of the power.
  • If you do leave your car in the sun, in many places that means it will get quite hot, you'll burn up some of the solar energy cooling it down. (Indeed, the solar panels sometimes found on today's hybrids and EVs don't charge the battery, they just run a cooling fan.)
  • The worst one: If your battery is not somewhat discharged, it doesn't have any place to put the solar energy, and so it is just thrown away. But due to range anxiety, people prefer their electric cars be kept full. It takes careful planning to use that energy.
  • A car is a very bumpy place, so you need more rugged panels than what you might put on a roof.

It is possible to get more than 200w on a car -- some of the solar challenge cars that exist to be nothing but panels have gotten around a kw by using high price, high-efficiency panels. But it's still generally much better to just put the panels on a roof where they will realize their full potential, and feed the grid, and charge from the grid.

However, on Friday I was teaching a class on the future of Robocars to my students at Singularity University and in the exercises some students wondered if they might do something for solar powered cars. (I was impressed since the students, having had only a short time to think about the issue, have to work to bring up something new.)

Robocars might solve some of the problems above, and thus possibly make sense as a place to put panels.

  • A robocar parks itself and can move. So one with a solar panel can move around to make sure it's always in the sun, and that the sun is striking it from the right angle. It can't move too far or too often without wasting some of the power, but it can do something.
  • When the batteries get so full that they are not making proper use of the solar energy, a robocar can find a charging station, not to charge but rather to sell excess power back to the grid and other cars. (This presumes charging stations are set up this way.)
  • Robocars could dock with other robocars that are more discharged and offer them the extra solar power, no charging stations involved -- though fancy robotics are needed on the charging interface, or human beings who can do the connections.
  • If a robocar has an actuator that can tilt the panels, it can do even better. While an ordinary car could have this, an ordinary car would not have the ability to rotate in the plane of the ground to track the sun without another actuator.

It's still not great, but it might improve things. Generally it still may be better to have the panels on rooftops and get the most from them. However, when we start thinking about super lightweight cars, cars that travel for under 100 watt-hours/mile, as well as higher efficiency panels, we might get some value if the panels are light.

It's also expensive to install panels on top of existing facilities. Turns out that while panels are dropping below 1$/watt next year thanks to cheap Chinese capital and manufacturing, the cost of install is still over $2/watt. Cost of install on newly manufactured buildings -- or cars -- can be cheaper because it's designed in from the start. The car already has the complex electrical system, while houses need to add them if they go solar.

People really are in love with the idea of a solar powered car. It's not really possible to go green this way right now, but the future might bring something interesting.


I was thinking about many of your robocar ideas while vacationing in Walt Disney World this past week. To move guests they generally use standard city buses running at 20 minute intervals between the parks and resort hotels. This is fine when moving people at peak times of the day, but was wasteful during non-peak times. When we traveled mid-morning through dinner time, there was usually less than 15 people on the bus. Their quality of service demands this, but it is rather inefficient. Their hub-and-spoke system also made moving between resorts difficult. It took us up to an hour to go from our hotel to another for dinner and shopping. If there ever was a demand for a robocar implementation, this place is it. How great would it be to walk to the hotel terminal, and get in a small-ish 6-8 passenger car, and be shuttled to your desired location. At the end of the day, when thousands are leaving the parks, wouldn't it be great to just start putting guests into cars, rather then cuing them in long lines, half a mile from the park exit, waiting for the next bus?

I'm getting a moderately hilarious vision of lots of selfish, self-actuating robocars in a partially-sunlit parking lot, all frantically cruising around in search of the optimally sunlit space and trying to beat out all the other robocars to it: BUMPERCARS!!! REALLY EXPENSIVE ONES!!!

However, when we start thinking about super lightweight cars ...

At that point we have stopped thinking about the driving conditions and what is prudent safety wise. A superlight car is going to be more prone to hydroplaning not to mention problems on snow or ice covered roads.

There are a variety of solutions to this, such as adding more weight (like batteries) in any conditions where a light vehicle would not be safe. And that the light cars are not meant for high speeds, but rather urban streets, and in a world where (this is some time in the future) we've gone to mostly robocars so it's all safer.

But in general, I was not under the impression that traction was much a function of total weight, because while the weight creates more friction, it also has similarly more inertia. However, I have not studied how hydroplaning varies with weight, but my understanding is that semi-trucks can hydroplane as much, if not more, than small cars due to weight distribution issues.

One KW/m^2 is a hard limit, being all the energy that is available from the sun. Multiply that by the panel efficiency and the angle or incidence and you get "not a whole lot" -- the best experimental panels get less than 50%, and none of them overcome trigonometry.

A practical vehicle consumes at least 10KW, so it takes you at least 2 days of charging to drive one hour, and probably 3 or 4. And this number will never improve much. Which is why this gets such justifiable derision.

Put a bunch of panels on your house (10s of sq meters), charge batteries during the day while out. Swap the car's batteries when you get home.

But is it worth it? The only real thing it gets you is the ability to drive again quickly since the swap is faster than the recharge. But you need twice as many batteries, and some sort of swap mechanism in your house. Unless you need that fast swap, it's a lot better to just have your panels feed the grid by day, and you charge the car from the grid by night.

There really is no solar car, at least not without deliberately doing something silly and even anti-green to make yourself believe you have a solar car. There are just two things -- an electric car, and a grid. You can make the grid greener by connecting panels to it, and that is fine, but it doesn't make you the driver of a solar car.

thinking in terms of today's electric cars is the wrong model.
the solar challenge cars go 500 miles a day on solar alone.
they are solar-only powered because those are the rules of the contest.
take that as the model, start out fully charged from the grid,
add human power (the driver and a hitchhiker, both pedaling. solves the mass transit and obesity problems.)
some form of backup power such as gas engine, compressed air, booze, fuel cell, whatever works.
your robocar ideas are a useful part of the mix.
you might not want to drive it on snow, but where i live 97+% of the time there's no snow on the roads.
a herd of solarrobobikes could hang out in the pasture,and only come back to the barn when fully charged and ready to go,
so there would always be one ready. or they could seek out grid connections on a road trip while the driver is getting coffee or showering.
the typical driver is going no more than 500 miles a day anyway. a robovehicle with a range of at least a 100 miles would be a useful gadget to have around; most trips are less than 100 miles.
a lightweight robovehicle that was solar- and human- and grid- and alternative-fuel- powered could be a viable alternative to the current imported gasoline at 25 miles/gallon $2.50-$5.00/gallon model.
I dont know what the per unit price would be if it were mass-produced in india, but it could be less than the typical detroit-built car.

Yes, as I said, it becomes different with a super light vehicle. The solar challenge cars also deliberately have all these wide surfaces for solar, which you would not actually want for your own vehicle, but panels can still make some difference on a much lighter vehicle.

My friend tynan (tynan.com) has 1000 watts of solar on his RV. He doesn't use it as motive force, but for his power when he lives in the RV (which is usually.) A typical trailer of a tractor-trailer unit is 8 feet wide, and 48 or 54 feet long.
If the top and maybe the sides were fitted with solar panels, how many watts would that produce, and would it be a non-negligble benefit if the trailer were a deisel-electric hybrid? Let's assume the trailer isn't fully loaded at 80 tons of cargo, but is either deadheading (no cargo) or hauling a light load (no more than 20 tons.)

What about using the solar panels to keep charge when parked outside. One of the major complaints about electric car companies is that they can brick or completely discharge, damaging the battery. What if you have to go away for a month and the battery dies. Here, a solar panel could force enough charge into the battery to prevent it from completely dying. Not sure how much charge a Tesla or any other vehicle would need to prevent bricking. Obviously, this doesn't work if you are parked in a barn or garage.

Yes, if a car were to be stored for a long time you might want to have a small panel to maintain charge. It's a pretty small panel though. Some cars also use a panel to power a fan to keep the interior from heating up. That can make sense because it heats up when left in the sun, so you use the power from the panel when it is there, and have no need of it at night etc. Wasted in winter though.

Most cars are not left alone for that long, but if they might be, the small panel could make some sense. It's a panel for driving that doesn't make too much sense on a conventional car. A robocar with a data radio and computer actually could want a bit more constant power if left alone a long time. Of course, if left alone too long, it could know to drive somewhere to get a charge.

Add new comment