Is it practical to generate all US power by solar PV?

Gold Member
2019 Award
Dearly Missed

Main Question or Discussion Point

In another PF thread it was proposed to build a centralized PV farm of 1000 gigawatts , which is the order of magnitude of US installed generating capacity. It'd cover 1/10 the area of New Mexico, Arizona and Nevada.
Mostly desert and dry US states:

Arizona: 295254 km^2
New Mexico: 315194 km^2

Sum: 896815 km^2

If we would tile only 10% of this land with PV panels we'd generate
897 GW (on average). And then there are dry, inhospitable areas in
Utah, Colorado and Texas if we would ever need more.

√ (10%of 896815 km^2) = 299.5 km per side, 186 miles per side, not far from the 150 stated earlier in the same thread.
Close enough for thought experiments.

You can't drive maintenance trucks over solar panels so the dimensions will expand to accommodate roadways.
Unless they're elevated to serve as rooftops with access from below.
Stormwater runoff from a 150 mile square rooftop will be a challenge, Phoenix area has been known to get 6 inches in a storm.
http://www.fcd.maricopa.gov/Weather/Rainfall/raininfo.aspx [Broken]

It'd be interesting that's for sure.
Myself, i am far more afraid of huge storage batteries than of reactors. I wouldn't be go anywhere near them.

Maybe @anorlunda will assess the practicality of moving so much power over so much distance.

Last edited by a moderator:

Related Electrical Engineering News on Phys.org
anorlunda
Mentor
That's quite an ambitious thought. I'll give it some thought, but here are a couple of preliminary points.

• Northeast utilities have determined that they could take up to 20% of their power from the Midwest. That does not say that they could not take more, but that it has not been studied. This idea is an order of magnitude beyond the boundaries of what has been studied. On an engineering scale, compare it with a proposal to build a bridge from Galveston to Tampa over the Gulf of Mexico, or bridges across the English Channel and the Bering Straight.
• Having said that, a lot of power can be transmitted a long way using HVDC But it would be very expensive, maybe trillions. Also it needs hundreds of thousands of miles of lines. Expect one landowner lawsuit per mile, so maybe more trillions for lawyers.
• It makes a great rhetorical point to locate all that PV power in one state, but it makes no engineering sense. The sun also shines on the other 47 of the lower 48, and distributed facilities would be far more reliable. Why not distribute the locations?

Gold Member
2019 Award
Dearly Missed
It makes a great rhetorical point to locate all that PV power in one state, but it makes no engineering sense. The sun also shines on the other 47 of the lower 48, and distributed facilities would be far more reliable. Why not distribute the locations?
Distributed rooftop solar.. has utility industry is really worried .
http://www.eei.org/ourissues/finance/Documents/disruptivechallenges.pdf

mheslep
Gold Member
Total installed capacity of US today is 1 TW, not 2. Average consumption is, naturally, much less that 1 TW.
As I said, with *all* energy use converted to electricity for a future scenario, including transportation, space heat, industrial, etc, the total US power load is 2 TW. Just backing up the existing US electric load at .5 TW for 7 days still requires some billion tons of battery. Currently, there is not a single utility scale battery project with even one day of run time, not anywhere in the world. The idea of backing up a large country with batteries, powered by solar, is nonsense. Power to gas or power to liquids has some small pilot plant traction, but no more.

mheslep
Gold Member
...
If you want to go all solar, you have to consider total energy consumption, which is about 3 TW...
Right, current primary energy use. Converted to electric transportation, heat pumps, etc, 2 TW is about right.

1/10 works for the US I guess, in Europe you still need the best places to get 1/10 of the installed capacity as average power.
Capacity factor in Arizona is 25%. Today. Not theoretical, it's what realized now.

Gold Member
2019 Award
Dearly Missed
Mentors - Request title be changed to

"is it practical to generate all US power by solar PV? "

jim mcnamara
Mentor
Expect one landowner lawsuit per mile, so maybe more trillions for lawyers.
Hooray for NIMBY's. @anorlunda you really nailed it.

The whole concept has problems with practicality. Especially the "political" aspect. One example --
1. All of the existing scada systems would need some sort of realtime control, centralized.
2. It would take trillions to upgrade the electric grid. My 2005 WECC map show Texas with 2 tie lines, good luck pushing 80000MW through them or getting TX to change it's 'yee-hah' policy on electric system management. Which will not fit #1 at all.

The point is really that every state can dictate how they want to manage/not manage utilities. And if cooperating is perceived as unwanted control (or whatever), the state has the right to balk. So having states that are proudly unique and somewhat defiant you cannot have a truly functional national e- grid. And yes there are FERC and NERC guidelines about lots of things. And TX and NM (where I am) are still able to do wacky things.

CalcNerd
Gold Member
I believe the best benefit of PV cells on rooftops is to help offset the huge AC usage of electricity during the hot summer months. This is when the solar cells contribute the most to the grid. If the solar cells are NOT installed, a larger baseline power reserve is needed/generated. If this power isn't generated locally (ie PV cells), more power is needed on already overburdened electrical lines and wasting even more energy as I^2R losses are also a positive loop problem ie more current demand = hotter wires = higher resistance and hotter medium and high voltage transformers at distribution centers.
.
The local PV cells don't help much, but the power company has ample reserves for 95% of the year, and if the PV cells can contribute that last CRITICAL 5% (numbers pulled from my posterior end) the power company will not have to build another dirty filthy coal powered energy plant. A win-win if the PV panels work at their most needed time of when the sun is driving up the temps in the middle of the summer.

Gold Member
2019 Award
Dearly Missed
Wow i never heard of WECC before. Search takes one to this interesting tutorial (256 pages but quite readable with copious illustrations)

It does a good job of explaining a lot of power system basics.

"The grid" is a huge machine that we cannot see in action by looking at it .
We can see trains moving goods along train tracks but not megawatt-hours moving down power lines.
So it is not obvious what a distributed, moving, and fragile system is "the grid".

to Jim McNamara's point #2:
2. It would take trillions to upgrade the electric grid. My 2005 WECC map show Texas with 2 tie lines, good luck pushing 80000MW through them
(With today's security environment good luck finding detailed transmission line maps....)
Generation is presently scattered and near point of use so most transmission lines aren't very long.
Take this map and redraw it for all generation emanating from the region of green rectangle.

It goes from a structure supported by a foundation underlying pretty much all of itself to one that's suspended by long spindly arms.
And for right now that's more than enough for me to contemplate. Old guys just don't handle fundamental change all that well.

old jim

credit for that map goes to
http://docplayer.net/docview/25/5923397/#file=/storage/25/5923397/5923397.pdf
with a big "Thank You ! ".

Last edited:
Perhaps batteries wouldn't be the best energy storage device. I visited a power plant south of Luddington Michigan which pumped water from Lake Michigan to a large reservoir at night when electric rates were low and let it flow back through a generator during the day when the rates were high. Some of the energy was stored in a large flywheel. One advantage of a flywheel is that there are lower losses converting between DC and AC. Energy from DC solar panels can be added to the flywheel with a DC motor and extracted using an alternator to feed the grid.

Attachments

• 37 KB Views: 366
mheslep
Gold Member
The most recently completed HVDC transmission project in North America was the W. Alberta Transmission line: 500 kv, 1 GW, $1.7 B, 350 km, or about$7 million per GW mile.

Using the cost of that project, replacing one of the US nuclear 1 GW reactor projects with solar a 1000 km away has transmission cost alone of $7 B, assuming some kind of yet-to-be-inventec storage is colocated with the solar. If storage is remote, then 4 or 5 GW of intermittent transmission is required for up to$35 B for transmission alone.

Last edited:
mheslep
Gold Member
....visited a power plant south of Luddington Michigan which pumped water from Lake Michigan to a large reservoir at night when electric rates were low and let it flow back through a generator during the day when the rates were high. ....
While the pumped storage hydro capacity in the US is useful for leveling a bit of peak demand, it's a trivial amount compared to the hundreds of TWH required to backup up the entire US for several days, the topic of the OP. The typical run time for a PHS plant is a few hours.

I was thinking about something like this earlier, but perhaps storing the surplus electricity as natural gas,
which could be moved through the natural gas grid.
https://www.fraunhofer.de/en/press/research-news/2010/04/green-electricity-storage-gas.html
The Gas could be used to fire existing local power plants, heat homes, or reformed into liquid fuels.
If the carbon came from atmospheric CO2, all the gas would be carbon neutral.
If the electrical utilities are pushing back from home solar, it is likely because of the highly imbalanced
price structure from net metering. As popular as net metering is, it needs to go away before home solar can
progress very far.
(Forcing a retailer to buy a product at his own retail price will never be popular with the retailer.)

First, you wouldn't want all of the solar energy in one spot because of cloud coverage and weather. What happens if we get a big storm? Is everyone screwed until it passes? By spreading out the panels over a larger area, you essentially minimize the "risk" or probability of a total outage. It isn't unlike diversifying your bonds (props if you got the reference).

Second, all your solar energy is one area is like having a big bulls-eye for our enemies. If some country wants to bomb us, just hit the solar arrays and we are screwed.

To tell you the truth, I think something similar to solar roads have the most promise. Good luck getting anything done with the energy lobbyists in charge. Too bad this is also one of those venture capital crackpot/BS start ups. I do think the concept in principle has potential.

CalcNerd
Gold Member
If someone wants to harvest energy from roadways, and pie in the sky ideas are looked at, consider a piezoelectric roadway surface. The bigger the vehicle, the more juice you get. I suspect that this technology would cost a fortune to implement, but you would only install it on the busiest of roads and it might even be made durable enough to avoid potholes!!!

anorlunda
Mentor
To make this thread interesting, let's dump the suggestion to place all that capacity only in Arizona.

The sun shines in all places in the lower 48 states, but because of weather, and because of latitude, PV panels are most effective per $m^2$ in some regions. However as Jim pointed out in #3, the cost of panels is only $0.86 out of a total installed cost of$5/kw. If we had to add 50% or even 100% more $m^2/watt$ to compensate for location, it wouldn't have a big effect on costs. Therefore, the benefits in locating all the generation in a sunny region are tiny compared to the massive costs of transmitting that power to the entire 48. Please let's scratch that part of the idea and focus on the potential of PV distributed over the entire country, close to the loads.

The interesting part of the question is scalability. Could the use of PV solar plus storage be scaled all the way up to the ultimate extreme of 100% of our electric capcity? The question must be asked on three levels.
1. Science: Are there any unbreakable laws of physics that we would have to break thus making the proposition impossible? (i.e. things like conservation of energy or the speed of light). I can't think of any, so my answer to that is no. If anyone disagrees, please raise those objections early.

2. Engineering: This is where I would like to focus this thread. Is it practical? What are the engineering, economic, or legal obstacles to PV scalability?

3. Social: Just because we can do something doesn't mean we should. I would like to leave social questions, including renewability or the cleanliness of PV, out of this discussion if possible. I also note that it would not make sense to abandon wind or any of our existing investment in electric generation even if 100% PV is possible. I mostly fear that the discussion in this thread may become too unfocused if we include social issues. PF already has a thread YOU!: Fix the US Energy Crisis for unfocused energy discussion.
42% of the USA population lives in multifamily housing with 5 or more units with very few $m^2$ of rooftop per tenant. Also as a guess, half of the single family homes have shade trees or orientation, or blocking hills or mountains that make PV unsuitable for them. Therefore, a discussion of PV solar scalability necessarily includes both consumer-owned rooftop PV and utility-owned central PV generation.

Given all the above, the existing bulk power transmission grid would remain largely unchanged in form and in public need by conversion to PV. We don't need to include that in the discussion at all. Local power distribution at the neighborhood level however, faces some major challenges.

mheslep
Gold Member
I was thinking about something like this earlier, but perhaps storing the surplus electricity as natural gas,
which could be moved through the natural gas grid.
https://www.fraunhofer.de/en/press/research-news/2010/04/green-electricity-storage-gas.html
The Gas could be used to fire existing local power plants, heat homes, or reformed into liquid fuels.
If the carbon came from atmospheric CO2, all the gas would be carbon neutral.
If the electrical utilities are pushing back from home solar, it is likely because of the highly imbalanced
price structure from net metering. As popular as net metering is, it needs to go away before home solar can
progress very far.
(Forcing a retailer to buy a product at his own retail price will never be popular with the retailer.)
A very good and more recent summary of the economics involved for P2G (and then G2P again) is found here. One problem with this approach, that is, an all solar power grid backed up by power-to-gas, is that in addition to the cost of the 2.5 TW of solar power required a full load capable gas power fleet is required. With a reliability margin, one needs about what is in place now (minus the hydro), or about 1 TW of gas fired electric power plant fleet, that runs only at night. The cost would be the solar, the gas fleet, and the P2G fleet: enormous.

Last edited:
However as Jim pointed out in #3, the cost of panels is only $0.86 out of a total installed cost of$5/kw.
Obviously, you meant per watt.

anorlunda
Mentor
Obviously, you meant per watt.
Yes, sorry. $0.86/watt for the panel, and$5/watt "all in ncost" In addition, I should have said that most of the $5.00-$0.86=$4.14 difference is not much dependent on the $m^2$ of the panels. Next compare the costs of locating a watt in Maine instead of Arizona. Because of the latitude and weather, we may need twice as many panels in Maine, thus$5.86/watt. Say \$6/watt round numbers. Compare that with the cost of hundreds or thousands of dollars per watt in capital costs to transmit the power from AZ to ME (I didn't really calculate those numbers, just eyeballed them). The point is that the economics of locating the panels near the load far outweigh those of putting the panels in a sunny state far away.

jim mcnamara
Mentor
@anorlunda - consider basing your starting point (as above) on this kind of data:
http://petedanko.net/state-per-capita-solar-pv-generation-2015/

There are some few places where solar is has a big existing footprint compared with most of the US.
So the total cost there per capita will be lower. Maine notwithstanding. FWIW fogbound San Francisco has a large footprint of residential PV generation. So attitude counts for a lot, IMO.

anorlunda
Mentor
Energy Storage:

Of course, a necessary part of large scale PV is battery storage. Other types of storage exist, but batteries are fully dispatchable and thus very friendly to the power grid.

Both home and utility scale battery storage are showing rapid development recently. The measures are energy capacity in watt-hours, power delivery capacity in watts, and cost. They are improving in all three measures. However, I think it's fair to say that they are now adequate yet, so we are speculating on the future.

Computerworld www.computerworld.com/article/3060896/sustainable-it/teslas-battery-sales-this-year-to-dwarf-entire-industrys-sales-in-15.html said:
Tesla is expected to sell 168.5 MWh of energy storage systems to the nation's leading residential solar system installer, Solar City.
If PV solar is the dominant form of generation, then other types of backup would dissapear. Storage capacity for a single day would not be enough, the capacity would have to cover the worst case weather event.

Regional reserve capacity also serve to offset local storage. For example, if one region's panels are covered by snow (or by dust in case of a volcanic eruption), then power could be supplied from neighboring regions until the panels are cleaned. That requires transmission capacity, but that level of transmission capacity already exists in most cases.

Weather is also a big factor. I remember one winter I spent in Vermont with more than 100 consecutive cloudy days. Even places like Arizona have a monsoon season. But I also know that my solar panels make as much as 50% as much energy on a cloudy day, so the solar capacity calculations are difficult.

Accurately calculating the actual reserve and energy storage requirements is exceedingly difficult. For purposes of this discussion, I'll hazard a guess of 100% overcapacity for reserve purposes, plus storage of 7x24 hours of energy would be the minimum requirements for a national PV system. But those guesses could be very wrong.

mheslep
Gold Member
Energy Storage:

Of course, a necessary part of large scale PV is battery storage. Other types of storage exist, but batteries are fully dispatchable and thus very friendly to the power grid.

Both home and utility scale battery storage are showing rapid development recently. ...
To date there are no utility scale backups of size and depth to wholly back up a single large power plant taken offline, nowhere in the world. The battery installations used by utilities, such as they are, are typically for minutes long carry over for the like of accommodating spinning standby. Similarly the home battery packs such as the ones from Tesla are also insufficient to allow anyone to cut loose a common home from the grid.

anorlunda
Mentor
@anorlunda - consider basing your starting point (as above) on this kind of data:
http://petedanko.net/state-per-capita-solar-pv-generation-2015/

There are some few places where solar is has a big existing footprint compared with most of the US.
So the total cost there per capita will be lower. Maine notwithstanding. FWIW fogbound San Francisco has a large footprint of residential PV generation. So attitude counts for a lot, IMO.
Nice data. Thanks.

The highest state was Nevada with 615 kwh/person*year. According to the Energy Information Administration, electric sales (all sectors) in 2015 were
3,724,000,000,000 kwh. The US population was 326 million. Thus consuption was 11423 kwh/person*year. That make the PV footprint in Nevada about 5%, if I did the numbers right. We are still very far indeed from 100%. I think the national average was 0.2% of US electric generation was PV solar, so the premise of this thread is a x500 expansion in solar PV capacity.

anorlunda
Mentor
To date there are no utility scale backups of size and depth to wholly back up a single large power plant taken offline, nowhere in the world. The battery installations used by utilities, such as they are, are typically for minutes long carry over for the like of accommodating spinning standby. Similarly the home battery packs such as the ones from Tesla are also insufficient to allow anyone to cut loose a common home from the grid.
That's very true.

We are beginning to expose some of the magnitudes of the problem. IMO, laymen most need education of the magnitudes of energy problems. That is part of the benefits from conducting a thread such as this. If we talk about a national 100% PV system, then the magnitudes can not be hidden or obscured.