4 ms·
I'm not talking about real-time demand matching/load shedding here as a consumer solution - as I mentioned, I think matching on a day-ahead (and perhaps also eg
by tomatocracy 7y ago
I'm not talking about real-time demand matching/load shedding here as a consumer solution - as I mentioned, I think matching on a day-ahead (and perhaps also eg 6-hour ahead) basis could well be a lot better than we have now because wind and solar generation is quite predictable on a day-ahead basis, as are outside temperatures (so eg for space heating energy consumption may be relatively well predictable too).
The beauty of this type of approach is that it can be done by incrementally building on things which already exist.
There's already an electricity market, including day-ahead pricing in most pools. Professional power traders already trade these markets. Smart meters already exist and are capable of dealing with pool prices not just flat rates. Smart thermostats already exist. EVs already have the capability to stop/start charging based on more than just whether they're plugged in or not. Even better, most of these already have standardised APIs.
What's missing is the software stack which ties all of these together - and that can be done incrementally, device by device (on the load side).
- generatorguy 7y agoThanks for the discussion. The information we are missing is what the percentage of total load can or would be willing to be shifted, and how strong the incentive (savings) is for the load to allow themselves to be shifted versus their inconvenience.
- fulafel 7y agoWe know that the answer is 100% as the price has no upper limit, no?