Hacker News Reader

Comment by Retric | Hacker News Reader

Using batteries for the grid requires roughly the same amount of batteries as EV’s do. A 100kWh battery pack for a car * 1,000 charge cycles / 25 years = 4,000 kWh of battery per year * ~290 million cars. That’s enough to store almost 1/3 of all current electricity used in the US while ignoring busses and farm equipment etc.

Of course actual cost depends on how much storage we need and future battery prices but adding something like 20-50$/month on future electricity bills gets offset by both cheaper electricity from solar and wind and less excess generating capacity. In that context large scale batteries seem like a perfectly reasonable solution.

Which is supported by many current grid scale PV instillations including them so they can avoid a AC>DC step when charging the batteries and then sell during peak demand at a premium.

View on HN

> Using batteries for the grid requires roughly the same amount of batteries as EV’s do

EVs have much harder constraints. The most obvious is that they need to move around (requiring high energy density, for both mass and volume); they also need to cope with sporadic charging times, and be reasonably fast to charge. It's very hard to compete with the leading Lithium-ion batteries in this space.

Grid storage isn't as constrained. Larger, heavier batteries are fine since they're just going to sit in one place. It's also easier to accomodate awkward/slow charging requirements, since they're always plugged in to the grid, and can be coordinated with other electricity sources/sinks if needed.

This allows different chemistries to compete, based on e.g. price, longevity, safety, etc.

Yep, it’s very possible that wildly different chemistry or even some other method wins. However, using the same battery chemistry in a cheaper form factor is the worst case. Aka if 2.3 Trillion on EV batteries works then the winner must cost less than 2.3 trillion.

Or have fewer externalities, since that’s never properly priced into anything.

This is never going to happen. As a car owner, you are in a market competing with large providers of grid scale storage, that have large purchasing power, control over the battery chemistry and finer specifications of their hardware which is made to order for their needs etc. Meanwhile, you have a mobility-optimized low weight battery that has a single, monopoly supplier that will very likely treat the spare parts market as a profit center 5 years down the road.

You will never be able to compete because you cannot differentiate, the product is fungible so mass always wins. It's essentially the bitcoin mining rig drama, the going was good until large scale mining operations were set up, with optimized ASICs etc. After that, good luck destroying your graphics card and car battery for pennies on the dollar.

Couldn't scheduling the EV charge when the demand is low while production is high help a lot balancing the grid, at basically no cost for the EV owner?

Let's say I have a EV with a smart charger that will keep the vehicle at least 60% charged, but charge up to 100% when the energy price is low (e.g. during the night).

> at basically no cost for the EV owner?

Only if you have a battery with unlimited charge cycles. This doesn't seem possible with current technology - and even if it were, manufacturers would still optimize for higher density and reasonable longevity after 100,000 miles. Most people would average a charging cycle a week, so they can't see the difference between a 2000 cycles battery and a 50,000 cycles one in the life of the car, but they can definitely feel the effects of range anxiety.

When you do daily or multiple times a day cycles, as is typical for grid applications, that's a completely different beast, for example a shallow charge cycle at 70% which increases the life 5x is much more profitable because it reduces overall battery replacement costs.

And when you factor in the much higher costs per Wh for car batteries, which are custom spare parts not mass produced commodity cells, you will find that the cost you incur in vehicle depreciation far exceeds the value you could earn from intra-day energy market speculation.

The idea is you wait to charge until prices drop, rather than charge as soon as possible which adds zero charge cycles or degradation. Discharging into the grid is unlikely to ever be profitable for the average consumer but it isn’t impossible for the economics to work out just look at how high Texes Grid prices have gotten during extreme events.

Delayed charging is already a common feature on many EV and could shift demand quite a bit in aggregate.

Also, battery degradation reduces range so there is an impetus to extend useful life well past expedited useable life. Aka if you want 95+% capacity at 100,000 miles that’s inherently increasing capacity at 1 million+ miles. Manufactures do care about resale value so useful capacity at 100k miles is likely to improve over time.

There is some talk about virtual energy providers that could agregate a large number of home users, receive an availability fee and only physically discharge during emergencies; that could work out economically and allow owners a positive revenue after depreciation, that could translate, for example, in lower prices for energy. That being said, I still think the whole fixed costs of the scheme (smart bidirectional meters, EV and charger support, coordination and administration costs etc.) would make it a money-losing proposal. A nice idea in theory, like say IP multicasting, but that cannot be made to work effectively in the real world.

The benefits of delayed charging are limited by the electric consumption of the transport sector. For an average family driving no more than 800 miles/month at 4 miles/kWh it works out less than 200kWh/month. If half of that charge is time flexible (with the rest being achieved at fastchargers or when the owner is in a hurry etc.), you get a 100 kWh/month dynamic load, a sizeable yet small fraction of the total household consumption. I would earmark that under smart-grid approaches, if correct pricing incentives are set at the meter it will happen automatically and not just for the EV charging loads.

A family only driving 800 miles a month is very low. The average driver is over 12k miles per year and 2 car families are extremely common.

Also, a relatively low percentage of charging is vis fast charging. Even a normal wall outlet can provide enough power do drive 15,000+ miles a year assuming normal habits, and level 2 home chargers are common.

I am not suggesting we use EV batteries for grid storage, just ~double the number of batteries created with half going to EV’s and half going to the grid.

Of course different constraints means different chemistry and form factors etc, but that’s only going to make grid batteries cheaper.

As someone that charges lithium batteries for both cars as well as for my house, I can promise you there is 0% of me that wants to cycle my vehicle battery from 100 to 0 every day to arbitrage $1.80 of electricity.

I am suggesting we build ~twice as many batteries as EV’s alone would need. With half being used for the grid not to directly use EV’s for grid storage.

Presumably if it gets anywhere close to similar scale grid batteries would end up with dramatically different form factors and chemistry as they don’t need to worry about collisions, charging time, etc.

But if you could cycle your battery from 50%-80% it’s possible you don’t get any meaningful wear on the battery.

You could also just charge in period of low usage of the grid (if there were special tariffs for that)

Depending on charging strategy, LFP batteries can do 4000 - 7000 cycles. 1000 is way too conservative, for any kind of EV battery.

Reply on news.ycombinator.com