Comment by ethagknight | Hacker News Reader

Comment by ethagknight | Hacker News Reader

This is not a battery, or a battery killer, and the author surely knows better. This is a distributed load bank, located in places the heat can be put to good use.

Interesting idea though. The paper seems like it’s pushing a product instead of pushing concepts, because the applicability of this concept goes much broader than the article goes into, for greater benefit to the consumer. A heat pump (a compressor and some fans) can be switched on or off just as fast as a heating element, at least as far as the article is concerned. The water heater switch is a dead simple solution though, credit due for that.

A bit of a rabbit trail here: A heat pump would not provide as much potential load, but it gives the flexibility of working for both heating and cooling, and also puts the electricity to better use. The key would be to put a cheap heat storage element in between the customer-premises heat pump and the final use, so that the utility can fire up the heat pump regardless of the customer’s need, and the customer has a huge bank of hot or chilled water at their disposal. Combined with a radiant floor heating system, and tied into a water-source HVAC system, this is a real energy efficiency boon by taking waste energy and getting it past the meter into the customer premises. Go ahead and throw a Steffes(tm) resistance heater in the water heater while we are at it. Have your cake and eat it too.

My father works in the heating world. He installed a large efficient storage water heater in the house which would basically warm the water in one go at 5am and provide it to the house all day.

What I don't understand from the article is why this is being controlled by the grid? The US is a market system, can't I buy a smart immersion heater which can then react to real time pricing information? So I warm my water up at times when the grid are paying people to use electricity?

That places the incentives correctly:

a) The entity making the investment in a new heater also gets the financial benefits thereof.

b) What's to stop your electricity supplier making your water 5% hotter every year to boost their revenues?

"What I don't understand from the article is why this is being controlled by the grid? "

Hypothetically if you bought your own you would need a smart meter so you're electricity company knew exactly when you had used power, there also needs to be a system to communicate to your smart appliance, to tell it to turn on. Neither of these really exist so I suppose its sensible for the utility to take ownership to iron out kinks, feel out how in works in practise and stuff.

To clarify, smart meters do exist, I don't believe they have the accuracy of metering needed for this application though.

The modern smart meters are quite accurate, much more accurate than the mechanical ones, usually to the ratepayer’s chagrin. The modern meters will pick up an induction spike from a noisy motor such as in an inefficient refrigerator or fan, where the older generation would not. How accurate does it need to be?

Its the time accuracy that matters, if its for frequency stabilisation then you need granularity in the order of seconds, rather than minutes as I believe smart meters are at the moment.

Oh no, they're way down in the milliseconds.

Electric vehicle chargers that can be controlled by the utility company exists in Norway. So in the event of electricity scarcity they can turn it off. I assume it's controlled by ripple or something, but the details are unknown to me.

They are the only ones that can know when it's best to turn it off an on.

The utility can signal this by varying the price from minute to minute. The heater can retrieve this signal from the internet.

Most utilities won't allow data like that over the internet.

If the electricity is down, the. Internet will be down, and without this feedback mechanism working they wouldn't be able to keep the grid stable.

No, but a growing number of utilities can inform customers that it is a 'peak usage' time and customer devices can adapt accordingly.

Nest (and I assume other internet-connected) thermostats are able to accept peak-usage notifications from the power company and dynamically reduce usage. In my case (Portland Electric), it's two-way communications, if Nest can prove to the power company that I let its device reduce usage during peak-load times, I get a credit at the end of the year. The Nest in my house has almost paid for itself.

> can't I buy a smart immersion heater which can then react to real time pricing information?

This is exactly how Nest makes money. About 40$ per thermostat per year from the utilities.

Wow - I had no idea thats who their real customer was. Brilliant.

I’m confused, the utility companies pay Nest?

I assume thats what was meant - but these two articles are interesting:

1) getting into big data to help utils

2) shutting down (as a google does) a data service for users ---

https://www.forbes.com/sites/parmyolson/2014/01/13/nest-give...

http://fortune.com/2016/03/01/nest-shuttering-data-service-m...

---

Disclaimer: I do not know if this means they are/was/will-be providing data to utils/ users or not...

I don't know, have you seen a price graph? Part of the point of financial markets is to predict the future. Since this isn't possible to do accurately, the result is a random walk of meaninglessly precise prices when really the best prediction you can do is a wide range of uncertainty.

The point of having smart water heaters is to stabilize the market, not make it more chaotic in pursuit of more efficient trades. So instead of going with something complicated like a financial market, a simple feedback system controlled by the power company will do.

Electricity prices fluctuate on a daily basis to an extremely predictable degree.

They do now. Add a bunch of noise traders, and maybe not.

The grid has topological constraints, because links in the network have finite capacity. Encoding that in a "price" would require pricing to vary not just based on total grid usage but also on the location of each load on the grid.

Tell that to grid controlled devices. You just encode a signal onto the carrier wave when you want to turn things off and on.

In most electric markets in the US, you can’t gain access as a consumer to electricity spot pricing. Consumers and regulators value stability and predictability in power pricing.

Additionally, the system is a lot more stable if a central utility can control these water heaters as a bank of shapeable demand, rather than having each individual device making independent decisions. It’s more efficient for them to give you a periodic rebate for participating in the program and have them manage the details.

> A heat pump (a compressor and some fans) can be switched on or off just as fast as a heating element,

This is not true, and the article specifically addresses this. A heating element can be switched on and off roughly five orders of magnitude faster (100 000× faster) than a heat pump.

You clipped out the qualifying phrase, “as far as the article is concerned”, dumping a few extra MW of electricity for 30 mins or a few hours. I don’t see a need presented in the article to rapidly switch a heating element on and off a few 1000 times a second.

Grid frequency stabilisation would require that.

I'm not actually sure that that would be the major problem. A heat pump is basically a motor, which can cause a lot of noise in electrical circuits. I'm not sure you would want 1000s of electric motors turning on in unison. I wouldn't like to say whether that is a solved/solvable problem though.

I don't think frequency stabilization would require that. Unless there is some complex approach I simply can't imagine.

You don't need it to switch on and off 1000s of times a second (is that what you're referring to?) you need it to switch fast enough that it could do that though. If it takes too long you would get brown outs and surges.

> Grid frequency stabilisation would require that.

The grid has a lot of inertia and a motor can go from full load to none, or from no load to extra-high startup load, in a fraction of a second.

So we're back to "just as fast" for this use case.

"The grid has a lot of inertia"

Where does the inertia come from? I understood inertia in the grid to come from turbines and motors spinning up and down, which seems to be different from what you're suggesting?

Capacitance in the wires?

The inertia largely comes from all the turbines and synchronous motors that stay connected.

With something like a heat pump compressor, you're connecting or disconnecting it completely with the flip of a switch. It's also an asynchronous motor and unable to feed power back into the grid at all, it's a pure load.

Ok I don't understand how that kind of inertia would help in this situation. If a turbine is running, and supply and demand are at equilibrium, if you then added another load, that inertia wouldn't help maintain equilibrium, it would slow down spinning up the turbine hampering that goal, wouldn't it?

To put it another way. Say if energy use jumps X%, inertia may make up Y% difference in the short term, you'd still have to put in more than X% in the slightly less short term, and it would take longer to make up the Y% shortfall. Is the Y% shortfall not big enough to be important here.

> Ok I don't understand how that kind of inertia would help in this situation. If a turbine is running, and supply and demand are at equilibrium, if you then added another load, that inertia wouldn't help maintain equilibrium, it would slow down spinning up the turbine hampering that goal, wouldn't it?

If you add more load, you start dragging down the frequency. The more inertia you have in the turbines, the less the frequency is impacted. I'm not sure what you mean by "spinning up a turbine" in this situation. Do you mean to produce more power? You don't spin turbines faster to do that.

> To put it another way. Say if energy use jumps X%, inertia may make up Y% difference in the short term, you'd still have to put in more than X% in the slightly less short term, and it would take longer to make up the Y% shortfall. Is the Y% shortfall not big enough to be important here.

Yes, you have to make up the deficit. But the alternative means having a brownout. In a choice between "drain some inertia to cover the increased load, then overproduce by .1% to refill the inertia" and "don't cover the increased load, voltage freefalls until load drops", I prefer the former.

Ok think I've got it, I thought the grid was more sensitive then it actually is.

Reading around, as the amount of renewables (solar) increases this might become an issue due to less turbines being in use.

"spinning up a turbine"

As in spinning up to speed. It can't go 0 to X000 rpm instantaneously. Inertia!

Didn't realise they were designed to spin at a constant speed though, thanks.

When you bring a new turbine online you get it up to speed before connecting it, and you can do this long before it's actually needed. So inertia hurts your response rate slightly but the main factor is how fast you can get your boiler or water pipe ramped up. That's why you see gas turbines as good peaker plants, they have very little to ramp up.

Solar is trickier but it's also more flexible in how it injects power. The real issue is not lack of inertia but the fact that it can lose production when a cloud comes by.

Maximum switching resolution is about 10ms in Europe (50Hz) and 8ms in the states (60Hz). As you usually switch at the zero crossing. In real life it is usually two-three orders of magnitude longer due to calculation and communicaton overhead.

This is not really an issue. Usually you have 4-5 seconds to react for the "hot reserve". Also for the known peaks you can preventively switch off the loads (it is amazing what happens at 7:02am when everybody switches on their kettle at the same time, even worse with water supply and flushing peaks).

In addition such systems sometimes measure frequency deviation and switch off/on the load when it goes out of bounds.

The result is that this works really well for fast balancing as soon as you have few hundred installed spots in one grid. And there is real money in it with ROI 1..3 years for the service provider.

Disclaimer: my electronics engineering company has worked with research, design and engineering of such load balancing hardware

I'm not an expert, but I'd think letting external factors control the switching on/off of your heat pump would have two downsides: (1) efficiency, and (2) wear and tear.

If you're talking about a heat pump used in air conditioning, at least, longer cycle times result in better dehumidification and higher efficiency. If you interrupt the cycle, you're effectively short cycling.

There are a couple downsides here. When sizing a climate control system of any sort, it’s nice to size it such that, at full output, it will be just barely sufficient for the worst case usage. If you use it for demand response and time-of-use optimization, you either need to size for a worst-case day such that it’s still powerful enough when it’s switched off at expensive times or you need to store heat or cold for multiple days. Second, while a simple-ish hydronic system is great for heat, it’s less ideal for cooling. You need to make sure that a cooling system can remove sufficient water vapor to keep a comfortable humidity or you need a separate dehumidifier. (Ideally every new building in a humid climate would have independent dehumidification anyway, so this is not a show stopper.). It’s also important to avoid condensation in bad places — if you cool your floor or ceiling much below the dew point, you have a problem.

I was trying to work out their attitude to heat pumps too. Its almost "they're too efficient, lets use something that uses more electricity" ?!

I don't share your concerns with calling it a battery though, it stores energy. Combined with a peltier it could generate electricity with that energy.

If it eases marketing, and gets this into peoples homes, I'm all for it.

I did the math on this once. For residential it is not realistic due to the amount of water you would need to store. For commercial I think you would need a small lake.

Correct, if you are trying to utilize it for a sole method of heat and cool. You will still need to dump excess / add additional through more conventional means, depending on local climate.

To your underlying point, none of the added complexity is worth the effort unless a) the utility is paying you to take the excess load and/or b) someone really innovates on a slick, simple package.

Reply on news.ycombinator.com