Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

>"dummy load"

You mean crypto miner.



Surprisingly, not always.

If I buy a device for $100 that, given free electricity, will mine $500 of cryptocurrency in its useful life - I can easily lose money if I run it less than 20% of the time.

And I doubt electricity is negative priced >20% of the time.


Yeah, there are a ton of plans in this thread for what to do with excess energy. The problem is, that’s the wrong question. The goal is to answer the question “what should we do with excess energy where we don’t mind building the capacity, but then only rarely running it.”

Rather than coming up with some grand scheme, maybe it would be good if our dishwashers and washing machines could listen to the grid and activate when power cost was negative. (We may need to coordinate a bit though, so we don’t all activate at once).


I have an electricity contract with dynamic pricing that changes every hour based on the day-ahead electricity market for Belgium. I know what the prices for the next day will be around 13h10. I charge the car whenever the prices are lowest: around noon in the sunny months, at night during winter, preferably weekends. I save around 25% of my electricity bill like this. (More in summer, less in winter.)

So it's already possible to incentivize people correctly with price signals, at least in some regions of the world. But people are not yet familiar with this. I guess that will change as the pricing between dynamic and traditional contracts keeps diverging. With a traditional contract, you are essentially paying the average evening peak price all the time. With a dynamic contract, you get access to the cheaper and even negative rates.


In some areas negative prices account for up to 25% of hours so it's a decent number but still a rough number of spins up and down and a lowish duty cycle. A solution might be to build battery capacity along side these loads to effectively buffer the negative cost power to be able to run continuously. That would skyrocket the initial capital investment though.


Yeah, batteries are just the sort of expensive/straightforward solution.

If you think of it, a dryer is sort of a combination of a flywheel and a heating element, so it should be the over-provisioner’s best friend. IMO a real failure has been not taking advantage of our appliances.


The issue there is connectivity and most residential customers don't pay spot prices so you need to upgrade their meters as well or build metering into the appliance so they can get credit for the energy they burn off. Plus you're looking at putting a lot of extra cycles on equipment not built as well as it used to be so you're burning the useful life of a hard to repair device and probably not getting paid enough to cover that, plus they more and more designed to burn as little energy as possible.

I know there are some places where this happens though but it's more along the lines of the devices delaying their start until energy is cheap rather than being used as loads to shed excess capacity afaik.


> I know there are some places where this happens though but it's more along the lines of the devices delaying their start until energy is cheap rather than being used as loads to shed excess capacity afaik.

This is what I meant, sorry for the ambiguity. Load the washer up and kick it off whenever energy is cheap. I don’t care when it happens other than, like, that it happens once a day, so why not defer this to the power company, right?


Like I said support isn't really there for a lot of electric customers. I pay a single flat rate for electricity so there's no point in time shifting consumption.

Also there are downsides to having clothes just sit there for hours potentially before you dry them. They can get pretty dank from the moisture and for dryers some clothes need to be removed immediately when the cycle finishes.



Not all crypto mining requires dedicated hardware. I'm sure plenty of datacenters would jump. I know at least one instance where certain internal use autoscaling groups increase in size with lowering electricity prices (AI training, batch Hadoop jobs, and the like) to reap the benefits. With negative prices it'd just be bounded by the total capacity.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: