9 ms·

> Tesla's contribution is enough to power about 2,500 homes for a full day This is < 0.1% of the total number of homes in SoCal, just to put it in perspective.

elliotec 10y ago
Yeah... seems odd to describe it as "massive"
nl 10y ago
It's pretty important to realize that this isn't the problem that needs solving. Batteries aren't great for large amounts of power for long periods of time (yet). What they are good at is smoothing out the supply curve from potentially intermittent suppliers like solar and wind, and that is one of the things that causes price spikes.
- WalterBright 10y ago
  Having real-time pricing of electricity coupled with users who can adapt their usage automatically to real-time prices will also considerably moderate power demand.
  - nl 10y ago
    Yes. Also, trading off the reliability of power. This seems outrageous to many of us in the West, but here in Australia our power utility already cuts power to neighborhoods in half hour blocks during peak power consumption during summer (I believe they exclude hospitals and some other sites like that from this policy). The infrastructure savings made by this are huge (think of 99% reliability vs 99.99% in software), and the few that really need it can invest in backup solutions for those times.
    - andruby 10y ago
      How huge are the savings? Do you have any numbers or sources? Do those power utility savings translate into savings to the people who are cut off? I for one would not want to trade electricity price for reliability.
      - jakub_h 10y ago
        Isn't it somewhat hypocritical to not want to pay for the expensive marginal case but to ask for someone else to do it for you?
        zaroth 10y ago
        It's surely less expensive per capita to have reliable electricity for everyone than to have everyone who wants reliable electricity to have to install a generator. 99% uptime means 7 hours of downtime every month. And it's not random downtime, it's downtime under peak demand. If the choice was free electricity at 99% uptime and you have to take care of the rest yourself, or $0.13 / kWh for 99.999% uptime, I would go for the 0.13 / kWh. Of course in AUS they probably charge more than 0.13 / kWh and make you deal with the brownouts anyway. Actually some quick searches reveal AUS pays about the highest rates in the world. Reliably not being able to handle peak demand isn't a cost saving measure, it's an excuse for a major failure of the infrastructure.
        nl 10y ago
        To be clear, in Aus we don't have 99% reliability. It's about 45 min every 2 years (whatever that works out to). That number was to explain the cost savings. In much of the world those cost saving are directly applicable. Most of the world's population doesn't have 99% reliable electricity, and so they build reliability at the edges with everything from batteries in phone towers, to batteries in lighting. It isn't at all clear to me if this is more or less expensive for new infrastructure. Given the increasing popularity of roof-top solar, the pricing model for 24/7 reliable wiring doesn't work out now in places where it is already built out. Network operators are trying various legislative measures to get subsidies for the networks, because no one wants to pay the rates they cost to maintain. If I had free (or very cheap) electricity 99% of the time, spending $5K to get 99.99% reliability via a battery system is very tempting (and we are getting close to that point now). $1K for 99.9% - maybe. One interesting thing is that the distribution of the downtime matters.
  - wlesieutre 10y ago
    Commercial users do see this, with higher costs for energy in peak hours as well as "demand charges" based on their highest rate of consumption (typically measured in 15 minute increments). Charging more for those gives an economic incentive to trim down the energy consumption bursts that require peaker plants to run. Buildings can design load leveling into their systems with things like thermal energy storage that shifts air conditioning consumption off-peak by freezing ice overnight and using it to cool the building during the day. One example with a Calmac IceBank system: http://www.greenbuildingadvisor.com/blogs/dept/energy-solutions/air-conditioning-ice http://www.greenbuildingadvisor.com/blogs/dept/energy-soluti...
    - WalterBright 10y ago
      People have always tried to come up with a reason to connect the refrigerator (and other appliances) to the internet. Finally, we have a reason - to moderate the power consumption of the refrigerator based on the spot price.
      - afarrell 10y ago
        Do you trust the software industry to keep others' from turning off your refrigerator for the lulz?
        toast0 10y ago
        Sensible price conscious refrigeration control would just be different temperature targets based on the price, not turning off the system. If l33t d00dz wanna make my milk one degree warmer, it's not a big deal.
        Ar-Curunir 10y ago
        You overestimate programming ability; people will find a way to fuck up.
        deleted 10y ago
  - flukus 10y ago
    Most peoples electricity usage is relatively inelastic and no one wants the cognitive overload of checking electricity prices before they put the TV on.
    - ptaipale 10y ago
      However, wouldn't it make sense to e.g. adjust heating/air conditioning depending on electricity spot price? That's then just a gadget that gets the information and guides the home system. It'll probably be more meaningful in some industrial plants though, if there's any such energy-intensive industry.
      - rdl 10y ago
        I'd be pretty happy to pause an EV charge many times. Since Teslas are data-connected, they could possibly do this with no grid interface beyond the power connection.
        greglindahl 10y ago
        This is one very nice aspect of a grid when there are a lot of electric cars which charge at night, and have big enough batteries that they don't need to fully charge: you can absorb a lot of variability of overnight wind energy. And if electric cars also plug in at work... jackpot.
        WalterBright 10y ago
        Even better than that, it could be set up so that your house drew power from the car battery when rates were high! You could set it to, for example, draw from the car battery until the charge drops to 80% or some settable limit. Then all the rechargeable cars become part of the battery moderation of the whole grid!
        ZeroGravitas 10y ago
        The Tesla engineering head said giving power back isn't likely, as it would degrade the battery too much, but he stressed how much benefit to grid stability could be delivered just by altering when the batteries drew power to charge.
  - Scarblac 10y ago
    In the Netherlands there is a lot of reclaimed land that needs a lot of water pumped out regularly (though not 24/7, during normal weather). I know there have been pilots with some of the larger pumping stations to incorporate the current real time price of electricity, as well as the forecast rain within the next few hours to decide when to turn on. Not sure if it's actually used in production now. That sort of thing is a great fit for that problem.
  - JustUhThought 10y ago
    I'd prefer batteries to having to regulate my electricity usage in real-time.
TheSpiceIsLife 10y ago
Or 60,000 for 1 hr. Or 120,000 homes for half an hour. Which goes a long way to cover wind / solar intermittency
derefr 10y ago
Most load is predictable; only a little bit is spiky. You don't really need to power 2,500 homes for a full day; you need to power five or ten huge factories at temporary overcapacity for a few minutes at a time, as reserve plants spin up/down. Consider EC2: you have reserved instances, and then you have autoscaling. If you know the total amount of work is going to predictably rise, you just buy more reserved instances (i.e. build more real power plants.) The non-reserved "elastic" VMs (like the batteries) are just to soak up your highly unpredictable excess load. Or, think less of a power plant, and more of a capacitor: if a capacitor is way more than is strictly needed to soak up any possible load spike in a system, then we'd say that the system is overengineered, even if a later version of the same system might have enough load to blow that same capacitor. It's "massive" in terms of the predicted load spikes it's facing, not massive in terms of the long-term consumption trend of the system.
api 10y ago
Once upon a time Facebook had less than 100 users.
mikeash 10y ago
A more useful perspective would be how much of the gap between peak generating capacity and peak load it covers. It's not intended to run SoCal for a day. Unfortunately, the "2,500 homes for a full day" statistic is really misleading, since it seems to imply that's what it would actually do.
kbenson 10y ago
But put in another perspective, it's enough to power ~3,6000,000 homes for a minute,or 720,000 homes for 5 minutes. If that can be used to prevent a brownout or blackout, that's a net win in my book. This of course assumes average use, which isn't really likely in the case you are protecting against the system being unable to cope, but at the same time it's supplemental to the current system, just as an extra few percent of capacity for an area, not meant to power homes directly or without other sources.

elliotec 10y ago

Yeah... seems odd to describe it as "massive"

nl 10y ago

It's pretty important to realize that this isn't the problem that needs solving. Batteries aren't great for large amounts of power for long periods of time (yet). What they are good at is smoothing out the supply curve from potentially intermittent suppliers like solar and wind, and that is one of the things that causes price spikes.

WalterBright 10y ago
Having real-time pricing of electricity coupled with users who can adapt their usage automatically to real-time prices will also considerably moderate power demand.
- nl 10y ago
  Yes. Also, trading off the reliability of power. This seems outrageous to many of us in the West, but here in Australia our power utility already cuts power to neighborhoods in half hour blocks during peak power consumption during summer (I believe they exclude hospitals and some other sites like that from this policy). The infrastructure savings made by this are huge (think of 99% reliability vs 99.99% in software), and the few that really need it can invest in backup solutions for those times.
  - andruby 10y ago
    How huge are the savings? Do you have any numbers or sources? Do those power utility savings translate into savings to the people who are cut off? I for one would not want to trade electricity price for reliability.
    - jakub_h 10y ago
      Isn't it somewhat hypocritical to not want to pay for the expensive marginal case but to ask for someone else to do it for you?
      - zaroth 10y ago
        It's surely less expensive per capita to have reliable electricity for everyone than to have everyone who wants reliable electricity to have to install a generator. 99% uptime means 7 hours of downtime every month. And it's not random downtime, it's downtime under peak demand. If the choice was free electricity at 99% uptime and you have to take care of the rest yourself, or $0.13 / kWh for 99.999% uptime, I would go for the 0.13 / kWh. Of course in AUS they probably charge more than 0.13 / kWh and make you deal with the brownouts anyway. Actually some quick searches reveal AUS pays about the highest rates in the world. Reliably not being able to handle peak demand isn't a cost saving measure, it's an excuse for a major failure of the infrastructure.
        nl 10y ago
        To be clear, in Aus we don't have 99% reliability. It's about 45 min every 2 years (whatever that works out to). That number was to explain the cost savings. In much of the world those cost saving are directly applicable. Most of the world's population doesn't have 99% reliable electricity, and so they build reliability at the edges with everything from batteries in phone towers, to batteries in lighting. It isn't at all clear to me if this is more or less expensive for new infrastructure. Given the increasing popularity of roof-top solar, the pricing model for 24/7 reliable wiring doesn't work out now in places where it is already built out. Network operators are trying various legislative measures to get subsidies for the networks, because no one wants to pay the rates they cost to maintain. If I had free (or very cheap) electricity 99% of the time, spending $5K to get 99.99% reliability via a battery system is very tempting (and we are getting close to that point now). $1K for 99.9% - maybe. One interesting thing is that the distribution of the downtime matters.
- wlesieutre 10y ago
  Commercial users do see this, with higher costs for energy in peak hours as well as "demand charges" based on their highest rate of consumption (typically measured in 15 minute increments). Charging more for those gives an economic incentive to trim down the energy consumption bursts that require peaker plants to run. Buildings can design load leveling into their systems with things like thermal energy storage that shifts air conditioning consumption off-peak by freezing ice overnight and using it to cool the building during the day. One example with a Calmac IceBank system: http://www.greenbuildingadvisor.com/blogs/dept/energy-solutions/air-conditioning-ice http://www.greenbuildingadvisor.com/blogs/dept/energy-soluti...
  - WalterBright 10y ago
    People have always tried to come up with a reason to connect the refrigerator (and other appliances) to the internet. Finally, we have a reason - to moderate the power consumption of the refrigerator based on the spot price.
    - afarrell 10y ago
      Do you trust the software industry to keep others' from turning off your refrigerator for the lulz?
      - toast0 10y ago
        Sensible price conscious refrigeration control would just be different temperature targets based on the price, not turning off the system. If l33t d00dz wanna make my milk one degree warmer, it's not a big deal.
        Ar-Curunir 10y ago
        You overestimate programming ability; people will find a way to fuck up.
        deleted 10y ago
- flukus 10y ago
  Most peoples electricity usage is relatively inelastic and no one wants the cognitive overload of checking electricity prices before they put the TV on.
  - ptaipale 10y ago
    However, wouldn't it make sense to e.g. adjust heating/air conditioning depending on electricity spot price? That's then just a gadget that gets the information and guides the home system. It'll probably be more meaningful in some industrial plants though, if there's any such energy-intensive industry.
    - rdl 10y ago
      I'd be pretty happy to pause an EV charge many times. Since Teslas are data-connected, they could possibly do this with no grid interface beyond the power connection.
      - greglindahl 10y ago
        This is one very nice aspect of a grid when there are a lot of electric cars which charge at night, and have big enough batteries that they don't need to fully charge: you can absorb a lot of variability of overnight wind energy. And if electric cars also plug in at work... jackpot.
        WalterBright 10y ago
        Even better than that, it could be set up so that your house drew power from the car battery when rates were high! You could set it to, for example, draw from the car battery until the charge drops to 80% or some settable limit. Then all the rechargeable cars become part of the battery moderation of the whole grid!
        ZeroGravitas 10y ago
        The Tesla engineering head said giving power back isn't likely, as it would degrade the battery too much, but he stressed how much benefit to grid stability could be delivered just by altering when the batteries drew power to charge.
- Scarblac 10y ago
  In the Netherlands there is a lot of reclaimed land that needs a lot of water pumped out regularly (though not 24/7, during normal weather). I know there have been pilots with some of the larger pumping stations to incorporate the current real time price of electricity, as well as the forecast rain within the next few hours to decide when to turn on. Not sure if it's actually used in production now. That sort of thing is a great fit for that problem.
- JustUhThought 10y ago
  I'd prefer batteries to having to regulate my electricity usage in real-time.

TheSpiceIsLife 10y ago

Or 60,000 for 1 hr. Or 120,000 homes for half an hour. Which goes a long way to cover wind / solar intermittency

derefr 10y ago

Most load is predictable; only a little bit is spiky. You don't really need to power 2,500 homes for a full day; you need to power five or ten huge factories at temporary overcapacity for a few minutes at a time, as reserve plants spin up/down. Consider EC2: you have reserved instances, and then you have autoscaling. If you know the total amount of work is going to predictably rise, you just buy more reserved instances (i.e. build more real power plants.) The non-reserved "elastic" VMs (like the batteries) are just to soak up your highly unpredictable excess load. Or, think less of a power plant, and more of a capacitor: if a capacitor is way more than is strictly needed to soak up any possible load spike in a system, then we'd say that the system is overengineered, even if a later version of the same system might have enough load to blow that same capacitor. It's "massive" in terms of the predicted load spikes it's facing, not massive in terms of the long-term consumption trend of the system.

api 10y ago

Once upon a time Facebook had less than 100 users.

mikeash 10y ago

A more useful perspective would be how much of the gap between peak generating capacity and peak load it covers. It's not intended to run SoCal for a day. Unfortunately, the "2,500 homes for a full day" statistic is really misleading, since it seems to imply that's what it would actually do.

kbenson 10y ago

But put in another perspective, it's enough to power ~3,6000,000 homes for a minute,or 720,000 homes for 5 minutes. If that can be used to prevent a brownout or blackout, that's a net win in my book. This of course assumes average use, which isn't really likely in the case you are protecting against the system being unable to cope, but at the same time it's supplemental to the current system, just as an extra few percent of capacity for an area, not meant to power homes directly or without other sources.