Hacker News Reader

Comment by digikata | Hacker News Reader

If Tesla can really put up grid storage modules as fast as they claim, shouldn't this just be a big opportunity to print money? Get paid to charge the batteries on surplus surges, get paid to supply power at night discharging them...

View on HN

Yes. The "obvious" question when reading an article like this is "Why pay someone to take the power, why not just turn off a gas or coal fired plant?" There are some obvious restrictions around that (like the plant would need to be in the general area of the grid where power was coming from of course) but the issue according to the PUC is that "cycle time" which is the amount of time it takes to turn off a power plant and then turn it on again, is longer than the "event time" where you get an over power event transient.

Grid storage nicely addresses both of these issues because it can absorb the extra energy (charging) and when the grid storage units are charged you can safely turn off a power plant because even if the power suddenly dips for what ever reason the grid storage can cover that loss while the power plant is restarting.

If you reach the point where you have enough grid storage to completely cover the variance between day time and night time power demand, then things get even more efficient as you switch to that storage a 'primary' and then cycle on plants as needed to top them off.

But so much of the demand is flexible. For example, I get zero benefit from heating my water tank at a particular time. Building grid storage so I can run my heat pump when it's dark and the air is still is inefficient and silly.

I'm not sure what you are trying to say here. At the individual user level energy usage may appear random and chaotic but at the community level that is not the case. To use your example, at any given instant in time someone in your city is heating water (whether intentionally by consuming water from a hot water storage facility, or passively as that facility activates automatically to keep the water at a given temperature). All of those usages level out to an aggregate demand on energy to heat water over unit time.

If you think about it as a capacitor that "filters" high frequency noise on the electrical grid it might be clearer.

I'm saying that with dynamic pricing grid loads no longer have to be noise. Everyone can charge, heat, clean when energy is cheap/available.

They level out to some extent, but they also enable far larger spikes to form when external events cause normally relatively "random" things to synchronise into narrower intervals across a large number of households.

E.g. the UK has what is called "TV pickup": Shortly after East Enders ends (daily evening TV soap opera), you'll get the sudden surge of 1.5m+ electric kettles turned on at once.

We see that type of situation where "random" events coalesce into very narrow time bands in all kinds of areas. I'd imagine we'll see this type of thing also happen for automated processes. E.g. air-con units ramping up based on thermostats in very lose proximity in the same areas etc.

I feel this is one of the areas where we need a smarter grid and smart home tech and home based batteries to work in conjunction to time-shift load and smooth out the demand more.

E.g. a hot water storage system can afford to turn off for 5 minutes (or whatever) and let the water cool slightly to take advantage of a discount 5 minutes down the line, or an air con unit can reduce speed for the same reason. Or a deep freezer that takes advantage of whatever the safe variation in freezer temperature would be.

Or, hey, I'd like a button on my washing machine and dishwashers that say "I want it to be done by time X. Find the lowest guaranteed rate (e.g. nightly rates) that will let you complete by X. If between now and then you get an offer for a lower rate, start then instead". Most of the time we just set "random" time delays to run them during the night, but most of the time I don't care if it actually runs then as long as it's done by the morning or by the time I'm done working etc.

You already have that happening to an extent at industrial scale. E.g. with industrial processes that require lots of energy that can be slowed or sped up according to prices, but there's a lot of discretionary demand in home as well that could be leveraged to smooth out the demand curve a lot more, and with a battery tied in it could smooth it even more.

You can build on this solution by using the money you save on selling at a loss to other states by using the funds to build storage. As solar capacity increases, more storage is built and in turn more coal plants are taken offline.

The only reason that this is not being implemented is corruption and campaign financing. If never fails to disappoint me the distance politicians will go to support big business as opposed to taking care of actual humans and their future.

Printing money is not a great business if the printers cost $10M and they print out $1,000 a day and lose the ability to print after 1,000 printings.

The economics are still not there yet for widespread battery energy storage. Here are some real numbers:

Wholesale prices fluctuate, but are often in the range of 0 - 10 cents per kWh. Lithium ion batteries cost something like $250/kWh. The inverting electronics cost like $100/kW. Round trip charging efficiencies are like 80%. Battery lifetimes are like 1,000 charge cycles. So assuming a system with 10 kWh at 5 kW, your costs will be $3,000. Or, on a per cycle basis, $3. If you charge 10 kWh for free and then sell 8 kWh for $0.10 each, you're making $0.80 on the transaction. Not enough to even cover your $3 in levelized costs.

(Obviously the economics improve with better assumptions. But there's still maintenance, interest rates, hookup costs, and destruction costs to take into account.)

Fortunately for Tesla, it also sells a few more products that make use of batteries.

I don't think the charging cycles is quite as open and closed as you represent. Its probably 1000 cycles to maintain 80% capacity (or some such threshhold), but you can keep running batteries in a storage-only application them far longer than one would find acceptable for keeping in a car.

But overall, yes, there are lot of technical considerations to actually doing it. And the article and discussions just touch the tip of the actual complexity involved - but it seems like there are power arbitrage opportunities of various corners of the grid in both long and short lasting sources.

I'm not an expert, but my understanding is that reduced capacity of old batteries is a symptom, not the problem. If reduced capacity was the only issue, then yes, it would make total sense to run old batteries into the ground.

But the issue is that the reduced capacity is a symptom that the battery is decaying internally, through a variety of mechanisms (cathode/anode micro-cracking, dendrite growth, physical leakage, who knows). Old batteries are dangerous fire risks, exactly what you don't want in a warehouse full of flammable batteries. Even a 1 in 100,000 risk of fire is unacceptably high if it causes a warehouse to explode. For this reason, it's risky to run old batteries into the ground. This is why there isn't a market to turn old electric car batteries into utility-scale storage.

Then again, this doesn't seem to be an issue for consumer electronics like laptops and phones and electric toothbrushes, so I wonder what's different, if anything.

If 65 days of the year you charge and 300 days you don't, it likely isn't cost effective.

Peak time is also during the daytime which means baseload power sources like Gas/Hydro/Nuclear/etc. are likely fulfilling demand at night just fine.

Most likely you are charging during winter and discharging during heat waves in the summer which may not be a viable business model despite the large delta in energy prices.

At least that is my understanding, I'm not an expert. xD

seems like a good opportunity, but how much would it cost for these storage units?

Reply on news.ycombinator.com