3 ms·

> Electricity markets have been exhaustively researched for ~2 decades (the type we have in the US) and even longer for economic power systems in general. Do y

by rsj_hn 5y ago

> Electricity markets have been exhaustively researched for ~2 decades (the type we have in the US) and even longer for economic power systems in general. Do you think these cost functions properly price disruptions? I'm saying this because the US is an order of magnitude less reliable than in Europe, and most of these costs are pushed onto consumers. E.g. When the power goes out and I have to throw away food or lose a day of work, PG&E has never compensated me for that. It's externalized. Now imagine a scenario where lots of people rely on the grid to charge their cars. I'm really skeptical of the obsessive hyper-optimization that's going on. If you have a situation where there is an extremely cheap energy provider that is highly unreliable, it could drive out of business the more expensive, reliable provider, and that needs to be taken into account. The zero or negative spot prices that were common in Texas as a result of windpower look to be an example of that pathology at work.

7thaccount 5y ago
They do have "scarcity pricing" which comes up when energy is either deficit or the required reserves are. ERCOT in particular has something like $9000/MWh.
- rsj_hn 5y ago
  So let's go with this number. A price is the result of optimizing some cost function, so let's see if the underlying cost function correctly measures the cost of a disruption. 9K/MWh would attribute 1 day of electricity as being worth $108 (assuming standard per-capita usage of 12kW). So for a single living alone, that's a bad deal, as suffering a power outage for 1 day and then being given $108 will not make them whole -- it wont even pay for the spoiled food, let alone the inconvenience of being without AC or internet for a day. For a household of 2.5, they would get $270, which would pay for the spoiled food and inconvenience of buying it and maybe get $30 left over for the inconvenience of no A.C., cell phone running out, not being able to use your computer or internet, do your laundry, perhaps not being able to use your mobile phone, etc. And we aren't even assuming PEV or people working from home. Then they would also lose 1 day of wages. So it would seem to me that just to make someone whole you should probably double or triple that, say to 20K/MWh, but then this is supposed to be a penalty rate, so it should cost more than what it would take to make a household whole, because we don't want the utility to be indifferent between making the household whole and providing the electricity, we want the utility to be heavily incentivized to avoid the disruption.
  - 7thaccount 5y ago
    These are wholesale prices, not retail, but yes the two are linked. $9000/MWh isn't crazy for a few intervals once a year, but if a generator clears 200 MW for each hour of the day for a day in the day ahead market and then has issues and can't provide that power in the real time market, it has to buy back that power at the spot price. That can get insanely expensive fast. Now imagine the owner of that generator has other generators in the same boat and it just stacks. These aren't trivial prices. Similarly, the load serving entities have to buy power at that spot price if they don't have it locked in at a lower price in the forward market (hard to do when events occur over multiple days). A little googling will turn up some chatter about many of those companies nearly defaulting from the ERCOT scarcity pricing during the February event.
    - rsj_hn 5y ago
      >$9000/MWh isn't crazy for a few intervals once a year, but if a generator clears 200 MW for each hour of the day for a day in the day ahead market and then has issues and can't provide that power in the real time market, it has to buy back that power at the spot price. That can get insanely expensive fast. Exactly! And it should be 30K, not 9K, due to my arguments above. You are pointing out that renewables, while unreliable, are so cheap they are driving the reliable electricity out of business. That means something isn't being correctly priced, because we as a nation require reliability. So I propose to put a price on that reliability, and you point out it can drive the unreliable provider out of business, which is kinda the point -- that unreliable provider will need to purchase insurance from the reliable provider to cover them, or "side payments" as you call it, in case something goes wrong. This puts a price on having the nuclear provider available in case the wind doesn't blow, and creates payments to that provider, but via market mechanisms. If the wind provider is confident in their battery storage they don't need to purchase insurance from the nuclear provider. But the point is to stop externalizing the costs of unreliable power. This is to end up in a situation where we as a nation are able to fund reliable sources of electricity rather than everyone switching to wind and the nation going dark when the wind stops blowing, throwing up it's hands and saying "oops", or "wow, I can't afford this. Gonna declare bankruptcy now, see you on the other side with more cheap spot prices". That's how you do it with the market mechanisms - by putting much bigger penalties on unreliability than we have now and requiring proof of insurance for being able to handle those payments. > These are wholesale prices, not retail Of course, because we are talking about how much to penalize providers for disrupting the lives of households - remember we are trying to price reliability. The providers pay wholesale, so that is the relevant cost function to target, not retail.

by rsj_hn 5y ago

7thaccount 5y ago

They do have "scarcity pricing" which comes up when energy is either deficit or the required reserves are. ERCOT in particular has something like $9000/MWh.

rsj_hn 5y ago
So let's go with this number. A price is the result of optimizing some cost function, so let's see if the underlying cost function correctly measures the cost of a disruption. 9K/MWh would attribute 1 day of electricity as being worth $108 (assuming standard per-capita usage of 12kW). So for a single living alone, that's a bad deal, as suffering a power outage for 1 day and then being given $108 will not make them whole -- it wont even pay for the spoiled food, let alone the inconvenience of being without AC or internet for a day. For a household of 2.5, they would get $270, which would pay for the spoiled food and inconvenience of buying it and maybe get $30 left over for the inconvenience of no A.C., cell phone running out, not being able to use your computer or internet, do your laundry, perhaps not being able to use your mobile phone, etc. And we aren't even assuming PEV or people working from home. Then they would also lose 1 day of wages. So it would seem to me that just to make someone whole you should probably double or triple that, say to 20K/MWh, but then this is supposed to be a penalty rate, so it should cost more than what it would take to make a household whole, because we don't want the utility to be indifferent between making the household whole and providing the electricity, we want the utility to be heavily incentivized to avoid the disruption.
- 7thaccount 5y ago
  These are wholesale prices, not retail, but yes the two are linked. $9000/MWh isn't crazy for a few intervals once a year, but if a generator clears 200 MW for each hour of the day for a day in the day ahead market and then has issues and can't provide that power in the real time market, it has to buy back that power at the spot price. That can get insanely expensive fast. Now imagine the owner of that generator has other generators in the same boat and it just stacks. These aren't trivial prices. Similarly, the load serving entities have to buy power at that spot price if they don't have it locked in at a lower price in the forward market (hard to do when events occur over multiple days). A little googling will turn up some chatter about many of those companies nearly defaulting from the ERCOT scarcity pricing during the February event.
  - rsj_hn 5y ago
    >$9000/MWh isn't crazy for a few intervals once a year, but if a generator clears 200 MW for each hour of the day for a day in the day ahead market and then has issues and can't provide that power in the real time market, it has to buy back that power at the spot price. That can get insanely expensive fast. Exactly! And it should be 30K, not 9K, due to my arguments above. You are pointing out that renewables, while unreliable, are so cheap they are driving the reliable electricity out of business. That means something isn't being correctly priced, because we as a nation require reliability. So I propose to put a price on that reliability, and you point out it can drive the unreliable provider out of business, which is kinda the point -- that unreliable provider will need to purchase insurance from the reliable provider to cover them, or "side payments" as you call it, in case something goes wrong. This puts a price on having the nuclear provider available in case the wind doesn't blow, and creates payments to that provider, but via market mechanisms. If the wind provider is confident in their battery storage they don't need to purchase insurance from the nuclear provider. But the point is to stop externalizing the costs of unreliable power. This is to end up in a situation where we as a nation are able to fund reliable sources of electricity rather than everyone switching to wind and the nation going dark when the wind stops blowing, throwing up it's hands and saying "oops", or "wow, I can't afford this. Gonna declare bankruptcy now, see you on the other side with more cheap spot prices". That's how you do it with the market mechanisms - by putting much bigger penalties on unreliability than we have now and requiring proof of insurance for being able to handle those payments. > These are wholesale prices, not retail Of course, because we are talking about how much to penalize providers for disrupting the lives of households - remember we are trying to price reliability. The providers pay wholesale, so that is the relevant cost function to target, not retail.