4 ms·

It'd take over 1 BILLION Powerwalls to store half the daily usage in the US (using old usage numbers off Wikipedia). I wonder if there's enough raw materials t

by ssmoot 10y ago

It'd take over 1 BILLION Powerwalls to store half the daily usage in the US (using old usage numbers off Wikipedia). I wonder if there's enough raw materials to make that realistic. And how to you manage recycling, repair, replacement, end-of-life concerns for all those batteries on that scale? Not asking a sarcastic hypothetical. Genuinely curious. It's hard to wrap my head around numbers that big. For reference I used this: https://en.wikipedia.org/wiki/List_of_countries_by_electricity_consumption https://en.wikipedia.org/wiki/List_of_countries_by_electrici..., divided it by days, then the Powerwall's 6.4kwh capacity and then halved that number. 4,686,400,000,000 / 365 / 6.4 / 2 == 1,003,082,191.8 Powerwalls.

epistasis 10y ago
Just throwing around "BILLION" doesn't make it seem big, you need to compare it the unit size. A single gigafactory is going to produce 35 GWh in a year, which is about 5.3 million power walls a year. If their lifespan is 10 years, than we need to produce 1e9 power walls / 10 years = 100 million powerwalls/year to maintain that capacity. So we'd only need 20 gigafactories worth to do this. And with wind and proper grid interconnects we won't even really need much storage at all to keep the electricity going; the national weather is very very consistent. We will need enough lithium for our ~300 million cars and their batteries, which are probably going to be 50-100kWh each, which is far far more than storing half a days worth of electricity.
- toomuchtodo 10y ago
  What if we used tens of millions of electric vehicles on the road as grid storage? Tesla has already planned on it: https://i.imgur.com/oLBNeW4.png https://i.imgur.com/oLBNeW4.png | https://i.imgur.com/pectzXn.png https://i.imgur.com/pectzXn.png
- ssmoot 10y ago
  That's over 3 Powerwalls every second. And I've got to think you're not going to see more than 75% original capacity in ten years as a best case? I'm not as concerned about the factory as the supply chain, can we actually supply that much raw material, and what will mining it do to the environment? I assume lithium is much rarer than iron or copper. But I don't really know. For reference 16 million vehicles were sold in the US in 2013, but many (most?) of those were trucks and large SUVs, which AFAIK aren't on the roadmap for any EV manufacturer right now. And then there's fleet/work vehicles, single car families in states without long distance transportation or quick chargers (like TX). It makes me wonder if 20 years from now fuel cells won't completely obsolete batteries?
gpm 10y ago
The number actually needed is a bit lower than that (I estimate half a billion), but the general question still stands. Variable electricity needs already mean we use less electricity at night time [0]. Shift electricity expensive processes preferentially run at night to the day and using 80%/120% seems like a reasonable rough approximation. So we need 4,686,400,000,000 / 365 / 2 * 0.8 ~= 513,5780,822 kwh in the '12 dark hours' Between current supplies nuclear hydro and wind we account for approximately 30% of electricity generation [1], on the assumption that electricity generation from these is constant, but can't be ramped up during the night, we get about 4,686,400,000,000 / 365 / 2 * 0.3 = 1,925,917,808 kwh. Leaving 3,209,863,014 kwh needed from storage, or 501,541,095 powerwalls. [0] http://www.eia.gov/todayinenergy/detail.cfm?id=13131 http://www.eia.gov/todayinenergy/detail.cfm?id=13131 [1] http://www.eia.gov/tools/faqs/faq.cfm?id=427&t=3 http://www.eia.gov/tools/faqs/faq.cfm?id=427&t=3
toomuchtodo 10y ago
Think about if Tesla vehicles were used instead of Powerwalls: https://i.imgur.com/oLBNeW4.png https://i.imgur.com/oLBNeW4.png | https://i.imgur.com/pectzXn.png https://i.imgur.com/pectzXn.png
- ssmoot 10y ago
  Your car isn't at the house while solar is being produced though for many/most people right? This seems unrealistic.
  - eru 10y ago
    A lot of cars sit in carparks during the day.
    - ssmoot 10y ago
      I guess the implication being there'd be a massive number of charge spots and you could swipe your card to pull off the grid what your home is putting on. But this would mean turning a car-park with just 40 plugs into needing 1,600 amps at 240v. With those grid connected homes having modest 2kWh solar installations you're talking 80kw/h. If you only spent 3kWh getting to work on average (double that for the round-trip) you'd fully charge the cars in a couple hours. So I guess the goal would be to run your home off the unused capacity. But then you're putting a far heavier duty cycle on those batteries. Whose going to pay for them when that $10,000 pack needs to be refurbished in five years? Or do we just accept EVs depreciating down to zero when the pack replacement cost exceeds their value? There may well be good answers to all this. Just some things that popped into my head. It would seem at first glance to be very costly for the car park, the grid and the consumer. I'm not sure any technology with the depreciation schedule of lithium batteries makes sense financially or environmentally?
      - eru 10y ago
        Thanks for working through some numbers. My modest intended implication was that if we only care about total sums, where the car is doesn't matter too much, as long as it has access to the grid. The economics could work out, if the car park charges you more for electricity than they pay the grid; and the grid your home less that it charges the car park. The car park and the grid both take a cut. For the home- and car owner, this works out to a slightly less efficient battery cycle. (If you convert the money charged back into the equivalent amount of energy.) Of course, if you are going to use your electric vehicle like as a battery like this, it's going to affect the battery's useful life. The same would happen with the single purpose battery you have sitting in your home. Now, the performance characteristics one looks for in a car battery and a grid storage battery are different. So I don't know whether using the car battery as a home battery even makes sense. (Ie home batteries can be bulky, but better be cheap. Car batteries have to be light and charge fast.) On a more refined note, one can play the same game, but only do it during spikes. Ie the car can stop charging during a spike in demand, and even uncharge if the spike gets big enough or the battery is already full. This won't do much for the diurnal cycle (that we discussed above), but a contribution during the spikiest peaks might already be worthwhile. My gut feeling is that we will see the demand shaping for charging, but I am more doubtful about whether we'll see cars actively feeding back into the grid.
maxerickson 10y ago
Residential is only about 1/3 of consumption. I guess at least industrial won't be using Powerwalls to store backup power. https://www.eia.gov/electricity/data/browser/#/topic/5?agg=0,1&geo=vvvvvvvvvvvvo&endsec=vg&linechart=ELEC.SALES.TX-ALL.A~ELEC.SALES.TX-RES.A~ELEC.SALES.TX-COM.A~ELEC.SALES.TX-IND.A&columnchart=ELEC.SALES.TX-ALL.A~ELEC.SALES.TX-RES.A~ELEC.SALES.TX-COM.A~ELEC.SALES.TX-IND.A&map=ELEC.SALES.US-ALL.A&freq=A&ctype=linechart&ltype=pin&rtype=s&pin=&rse=0&maptype=0 https://www.eia.gov/electricity/data/browser/#/topic/5?agg=0... It might not be possible to make all those batteries using lithium, but that doesn't really matter for stationary batteries.
- ssmoot 10y ago
  That's a huge factor I hadn't considered. Thanks.
DennisP 10y ago
Raw material supply is a big issue; see Berkeley physicist Tom Murphy's calculations: http://physics.ucsd.edu/do-the-math/2011/08/nation-sized-battery/ http://physics.ucsd.edu/do-the-math/2011/08/nation-sized-bat... He's mainly talking about lead but for lithium it's even worse. A lot of things look great until you look at the scaling limits. Hydro storage is another example; Murphy's got a post on that too.
- ssmoot 10y ago
  If these numbers are even close to accurate, this should be the top post. It's no use for everyone to argue over wether a completely implausible idea is financially competitive on a small scale when the implication is we're talking about scaling on a national level.

by ssmoot 10y ago

epistasis 10y ago

Just throwing around "BILLION" doesn't make it seem big, you need to compare it the unit size. A single gigafactory is going to produce 35 GWh in a year, which is about 5.3 million power walls a year. If their lifespan is 10 years, than we need to produce 1e9 power walls / 10 years = 100 million powerwalls/year to maintain that capacity. So we'd only need 20 gigafactories worth to do this. And with wind and proper grid interconnects we won't even really need much storage at all to keep the electricity going; the national weather is very very consistent. We will need enough lithium for our ~300 million cars and their batteries, which are probably going to be 50-100kWh each, which is far far more than storing half a days worth of electricity.

toomuchtodo 10y ago
What if we used tens of millions of electric vehicles on the road as grid storage? Tesla has already planned on it: https://i.imgur.com/oLBNeW4.png https://i.imgur.com/oLBNeW4.png | https://i.imgur.com/pectzXn.png https://i.imgur.com/pectzXn.png
ssmoot 10y ago
That's over 3 Powerwalls every second. And I've got to think you're not going to see more than 75% original capacity in ten years as a best case? I'm not as concerned about the factory as the supply chain, can we actually supply that much raw material, and what will mining it do to the environment? I assume lithium is much rarer than iron or copper. But I don't really know. For reference 16 million vehicles were sold in the US in 2013, but many (most?) of those were trucks and large SUVs, which AFAIK aren't on the roadmap for any EV manufacturer right now. And then there's fleet/work vehicles, single car families in states without long distance transportation or quick chargers (like TX). It makes me wonder if 20 years from now fuel cells won't completely obsolete batteries?

gpm 10y ago

The number actually needed is a bit lower than that (I estimate half a billion), but the general question still stands. Variable electricity needs already mean we use less electricity at night time [0]. Shift electricity expensive processes preferentially run at night to the day and using 80%/120% seems like a reasonable rough approximation. So we need 4,686,400,000,000 / 365 / 2 * 0.8 ~= 513,5780,822 kwh in the '12 dark hours' Between current supplies nuclear hydro and wind we account for approximately 30% of electricity generation [1], on the assumption that electricity generation from these is constant, but can't be ramped up during the night, we get about 4,686,400,000,000 / 365 / 2 * 0.3 = 1,925,917,808 kwh. Leaving 3,209,863,014 kwh needed from storage, or 501,541,095 powerwalls. [0] http://www.eia.gov/todayinenergy/detail.cfm?id=13131 http://www.eia.gov/todayinenergy/detail.cfm?id=13131 [1] http://www.eia.gov/tools/faqs/faq.cfm?id=427&t=3 http://www.eia.gov/tools/faqs/faq.cfm?id=427&t=3

toomuchtodo 10y ago

Think about if Tesla vehicles were used instead of Powerwalls: https://i.imgur.com/oLBNeW4.png https://i.imgur.com/oLBNeW4.png | https://i.imgur.com/pectzXn.png https://i.imgur.com/pectzXn.png

ssmoot 10y ago
Your car isn't at the house while solar is being produced though for many/most people right? This seems unrealistic.
- eru 10y ago
  A lot of cars sit in carparks during the day.
  - ssmoot 10y ago
    I guess the implication being there'd be a massive number of charge spots and you could swipe your card to pull off the grid what your home is putting on. But this would mean turning a car-park with just 40 plugs into needing 1,600 amps at 240v. With those grid connected homes having modest 2kWh solar installations you're talking 80kw/h. If you only spent 3kWh getting to work on average (double that for the round-trip) you'd fully charge the cars in a couple hours. So I guess the goal would be to run your home off the unused capacity. But then you're putting a far heavier duty cycle on those batteries. Whose going to pay for them when that $10,000 pack needs to be refurbished in five years? Or do we just accept EVs depreciating down to zero when the pack replacement cost exceeds their value? There may well be good answers to all this. Just some things that popped into my head. It would seem at first glance to be very costly for the car park, the grid and the consumer. I'm not sure any technology with the depreciation schedule of lithium batteries makes sense financially or environmentally?
    - eru 10y ago
      Thanks for working through some numbers. My modest intended implication was that if we only care about total sums, where the car is doesn't matter too much, as long as it has access to the grid. The economics could work out, if the car park charges you more for electricity than they pay the grid; and the grid your home less that it charges the car park. The car park and the grid both take a cut. For the home- and car owner, this works out to a slightly less efficient battery cycle. (If you convert the money charged back into the equivalent amount of energy.) Of course, if you are going to use your electric vehicle like as a battery like this, it's going to affect the battery's useful life. The same would happen with the single purpose battery you have sitting in your home. Now, the performance characteristics one looks for in a car battery and a grid storage battery are different. So I don't know whether using the car battery as a home battery even makes sense. (Ie home batteries can be bulky, but better be cheap. Car batteries have to be light and charge fast.) On a more refined note, one can play the same game, but only do it during spikes. Ie the car can stop charging during a spike in demand, and even uncharge if the spike gets big enough or the battery is already full. This won't do much for the diurnal cycle (that we discussed above), but a contribution during the spikiest peaks might already be worthwhile. My gut feeling is that we will see the demand shaping for charging, but I am more doubtful about whether we'll see cars actively feeding back into the grid.

maxerickson 10y ago

Residential is only about 1/3 of consumption. I guess at least industrial won't be using Powerwalls to store backup power. https://www.eia.gov/electricity/data/browser/#/topic/5?agg=0,1&geo=vvvvvvvvvvvvo&endsec=vg&linechart=ELEC.SALES.TX-ALL.A~ELEC.SALES.TX-RES.A~ELEC.SALES.TX-COM.A~ELEC.SALES.TX-IND.A&columnchart=ELEC.SALES.TX-ALL.A~ELEC.SALES.TX-RES.A~ELEC.SALES.TX-COM.A~ELEC.SALES.TX-IND.A&map=ELEC.SALES.US-ALL.A&freq=A&ctype=linechart&ltype=pin&rtype=s&pin=&rse=0&maptype=0 https://www.eia.gov/electricity/data/browser/#/topic/5?agg=0... It might not be possible to make all those batteries using lithium, but that doesn't really matter for stationary batteries.

ssmoot 10y ago
That's a huge factor I hadn't considered. Thanks.

DennisP 10y ago

Raw material supply is a big issue; see Berkeley physicist Tom Murphy's calculations: http://physics.ucsd.edu/do-the-math/2011/08/nation-sized-battery/ http://physics.ucsd.edu/do-the-math/2011/08/nation-sized-bat... He's mainly talking about lead but for lithium it's even worse. A lot of things look great until you look at the scaling limits. Hydro storage is another example; Murphy's got a post on that too.

ssmoot 10y ago
If these numbers are even close to accurate, this should be the top post. It's no use for everyone to argue over wether a completely implausible idea is financially competitive on a small scale when the implication is we're talking about scaling on a national level.