36 ms·

Overprovisioning doesn't help unless you distribute it wide enough that they don't all have the same points of inactivity. Even then wide regional issues can st

Overprovisioning doesn't help unless you distribute it wide enough that they don't all have the same points of inactivity. Even then wide regional issues can still give temporally correlated downtime, eg. Texas's wind turbines all not working at the same time due to low temperatures and icing. It doesn't matter how much you overprovision, 10 x 0 = 0. Overprovisioning plus a wide regional distribution could work, but then you need lots of extra power transmission capacity, which is also expensive, and will be 80% idle 99% of the time. Remember, we're engineering for the worst case scenario where we still need to provide power here. If we can think of it, it will happen sooner or later. Nuclear power plants are designed to be safe even if an airplane is flown into them, it's not good enough for renewables to then turn around and say "you need to redesign your entire society around our power being intermittent". I'm hugely pro-renewables, but only for remote areas. For cities, wind/solar don't make sense due to reliability and energy density.

ben_w 4y ago
> It doesn't matter how much you overprovision, 10 x 0 = 0. Yeah, but x0 only happens at night. I'm not sure what the multiplier is in a really bad storm, nor how long that lasts (~= night would be fine, you're already putting in that kind of storage). > Overprovisioning plus a wide regional distribution could work, but then you need lots of extra power transmission capacity, which is also expensive, and will be 80% idle 99% of the time. IIRC a global grid is (naïvely) ~= the cost of 6 months crude oil. Expensive in absolute money terms, but not relative money terms; though political cost is something else entirely.
- snovv_crash 4y ago
  x0 can also happen due to equipment failure, eg. if there was extreme heat that put all of the inverters into safety shutdown. Again, we aren't dealing with day-to-day, we're dealing with the 0.1% chance scenarios. These are what make renewables expensive when we're trying to supply high availability power. The problem with the global grid is that it would need to be similarly overprovisioned so that during low-probability failure scenarios the few remaining power nodes could supply the entire thing. 10x overprovisioning here looks a lot more expensive.
  - ben_w 4y ago
    > x0 can also happen due to equipment failure Is trivially true, and applies to everything. It's not really worth bothering to mention because everything else also sufferers this. > The problem with the global grid is that it would need to be similarly overprovisioned so that during low-probability failure scenarios the few remaining power nodes could supply the entire thing. The need to over-provision a grid is obvious, but… > 10x overprovisioning here looks a lot more expensive. First: Why do that by a factor of x10? Best redundancy here is geographical diversity rather than a fatter… I was going to say "cable", but it is (or collectively, they are) the order of a few square meters cross section and that feels wrong as a name. But that thing is best spread out, not kept singular and made wider, whatever you call it. Second: Even x10, the main limit is "that's a lot of stuff to mine, how do we reorganise the miners from coal and oil to metals" rather than the $ cost — while "a trillion" of anything is a lot for one person to contemplate, compared to the cost of what is currently dug up and then set on fire to provide the same power, it's quite cheap.

ben_w 4y ago

> It doesn't matter how much you overprovision, 10 x 0 = 0. Yeah, but x0 only happens at night. I'm not sure what the multiplier is in a really bad storm, nor how long that lasts (~= night would be fine, you're already putting in that kind of storage). > Overprovisioning plus a wide regional distribution could work, but then you need lots of extra power transmission capacity, which is also expensive, and will be 80% idle 99% of the time. IIRC a global grid is (naïvely) ~= the cost of 6 months crude oil. Expensive in absolute money terms, but not relative money terms; though political cost is something else entirely.

snovv_crash 4y ago
x0 can also happen due to equipment failure, eg. if there was extreme heat that put all of the inverters into safety shutdown. Again, we aren't dealing with day-to-day, we're dealing with the 0.1% chance scenarios. These are what make renewables expensive when we're trying to supply high availability power. The problem with the global grid is that it would need to be similarly overprovisioned so that during low-probability failure scenarios the few remaining power nodes could supply the entire thing. 10x overprovisioning here looks a lot more expensive.
- ben_w 4y ago
  > x0 can also happen due to equipment failure Is trivially true, and applies to everything. It's not really worth bothering to mention because everything else also sufferers this. > The problem with the global grid is that it would need to be similarly overprovisioned so that during low-probability failure scenarios the few remaining power nodes could supply the entire thing. The need to over-provision a grid is obvious, but… > 10x overprovisioning here looks a lot more expensive. First: Why do that by a factor of x10? Best redundancy here is geographical diversity rather than a fatter… I was going to say "cable", but it is (or collectively, they are) the order of a few square meters cross section and that feels wrong as a name. But that thing is best spread out, not kept singular and made wider, whatever you call it. Second: Even x10, the main limit is "that's a lot of stuff to mine, how do we reorganise the miners from coal and oil to metals" rather than the $ cost — while "a trillion" of anything is a lot for one person to contemplate, compared to the cost of what is currently dug up and then set on fire to provide the same power, it's quite cheap.