Replacing Gasoline with Solar Power
If you don’t want to run through the calculations, here is the summary. I attempted a thought experiment in which I calculated whether it would be feasible to use solar power to generate enough energy to offset all U.S. gasoline consumption. My conclusion is that it will take about 444,000 megawatts of electrical generating capacity. Current U.S. generating capacity is over 900,000 megawatts, but there isn’t a whole lot of spare capacity in that number.
To generate 444,000 megawatts with solar PV would require just under 1,300 square miles (a 36 mile by 36 mile square) of just PV surface area. To generate that much power with solar thermal – including supporting infrastructure – would require 4,719 square miles (a 69 mile by 69 mile square). A large area, but not impossible to envision us eventually achieving this.
Having made an attempt to calculate the number of square miles to replace current U.S. electricity consumption via solar PV or solar thermal, I have been challenged to do the same exercise for replacing our gasoline usage. (In fact, I was told by someone that they had never seen this kind of calculation done, so I told them I would do it). I have no idea how this calculation is going to turn out, but I suspect it is going to be similar to the previous calculation for replacing electrical consumption. My guess is less than 100 miles by 100 miles. Note that this is a thought experiment, in which I try to get an idea of what it would take to achieve this.
First, some caveats. There are still technical obstacles that prevent this scenario from being realized. Those are, 1). Battery range is still too low (The plug-in Prius is only going to be able to go 7 miles on battery power).; and 2). Solar power can’t be adequately stored. However, that’s not the purpose of the exercise. The purpose is to satisfy my curiosity: If we were going to try to replace gasoline with solar power, are the land requirements prohibitive?
I am only going to do this calculation for gasoline, as I think it is unlikely that electricity will ever power long-haul trucks or airplanes.
How Much Do We Need?
The U.S. currently consumes 389 million gallons of gasoline per day. (Source: EIA). A gallon of gasoline contains about 115,000 BTUs. (Source: EPA). The energy content of this is equivalent to 45 trillion BTUs per day. The average efficiency of an internal combustion engine (ICE) – that is the percentage of those BTUs that actually go into moving the vehicle down the road – is about 15%. (Source: DOE). Therefore, the energy that goes toward actually moving the vehicles is 6.7 trillion BTUs per day.
The efficiency of electric infrastructure can be broken down into several steps. According to this source, the respective efficiencies for the transmission lines, charging, and the vehicle efficiency are 95%, 88%, and 88%, for an overall efficiency (after the electricity is produced) of 74%. To replace the gasoline BTUs that go toward moving the vehicle with electricity is going to require 6.7 trillion/0.74, or 9.1 trillion BTUs. To convert to electricity, we use 3,413 BTUs/kilowatt-hour (kWh). Thus, 9.1 trillion BTUs/day is equal to 2.7 billion kWh/day. That’s how much energy we need. To convert this to power, we need to multiply by 1 day/24 hours, and that gives us 111 million kilowatts, or 111,000 megawatts (MW) of power generation required.
Looking back at my Solar Thought Experiment, I calculated 2,531 square miles to replace our peak electrical demand of 746,470 MW (746 GW). However, the current calculation is a different sort of calculation than what I did previously. The previous calculation attempted to have enough installed solar PV to meet peak demand. In the case of replacing our transportation fuel, I need enough panels to produce the required transportation energy in 8 hours or so while the sun is shining. To be conservative, we can assume 6 hours, which means we will actually need four times the 111,000 MW, or 444,000 MW.
Using Solar PV
From the previous essay, I used a conservative value of 12.5 watts per square foot as the generating capacity of an actual GE PV panel. To get 444,000 MW is going to take an area of 35.5 billion square feet, which is 1274 square miles. This is an area of just under 36 miles by 36 miles. However, this is just the surface area required to generate the electricity. It does not include area required for supporting infrastructure.
Using Solar Thermal
Doing the same calculation based on the solar thermal output from Running the U.S. on Solar Power, the expectation was that 0.147 megawatts could be produced per acre. This did include all of the land associated with infrastructure. If we use that number, we find that to generate 444,000 MW is going to take a little over 3 million acres, or 4,719 square miles. This is a square of just under 69 miles by 69 miles.
The reality is that we would use a combination of the solar PV and solar thermal. We have a lot of available rooftops that can create electricity with solar PV, and there are large tracts of land in sunny Arizona and Nevada that can create electricity with solar thermal.
Clearly, a lot of area is required, but it isn’t impossibly large. Of course to achieve this, a couple of big problems need to be resolved. First, battery life needs to improve somewhat before people are going to embrace electric transport. According to this ABC News story, the average commute is 16 miles one-way, but the range of the plug-in Prius is only expected to be 7 miles. The Aptera, on the other hand, claims a range of 120 miles. Maybe we just need to change the way we think about what we drive. (On the other hand, not a lot of commuters are going to climb into an Aptera if they have to share the road with large SUVs).
Second, and the bigger issue, is that we still don’t have a good way to store excess solar power. We need to have a good storage mechanism so electric cars can be charged at night from solar electricity produced during the day. One idea for this that I have seen floated is to use peak solar energy to electrolyze water, and then store the hydrogen in centralized locations. The hydrogen would then be burned at night to run centralized electrical generators. Not the most efficient method for storing solar energy, but technically workable.
Finally, the current electrical grid couldn’t handle such a large increase, but the model I envision would generate and consume the electricity locally.
I had delayed posting this for almost a week, because I was sure there was an error in the calculations. I finally found one (I had turned a kilowatt into a watt), but let me know if you find other errors or incorrect assumptions.