Swapping is Coming [Discuss how it will be accomplished]

[ggies07]

how big would this thing be in the frunk though? That just takes up more space and creates less. Does this justify one of their marketing points as "no more engine = more storage for you"?

 

[Trnsl8r]

Why are you so sure that's coming?
There are other ways to accomplish the same goal.

Not to derail the thread, but yeah I have to say I disagree with the premise, especially if it's based on the 10-Q report (remember the "forward-looking statements" disclaimer) and Elon's tweet below (my emphasis added):

There is a way for the Tesla Model S to be recharged throughout the country faster than you could fill a gas tank.

 

[Babylonfive]

I disagree with the premise. There is some sort of main-pack swapping for service, but the 'swapping' warnings don't have to be for the main pack. They could be a secondary battery, especially a non-rechargable.

Forgive me, but I am dreaming of the ideal: a 500-1000 mile non-rechargeable battery you rent for the convenience plus value of the lost anode material, and it only must charge the car while it runs faster than average discharge rate at 60mph (DC charge at 20-25KW) to be a distance tamer.

<dream mode status: latched and locked>

 

[neroden]

Emphasis mine:

There is a way for the Tesla Model S to be recharged throughout the country faster than you could fill a gas tank.

Here is a completely goofy and unlikely interpretation of this. :wink:

I have it on good authority that the "answer is obvious in general concept, but unexpected in implementation."

Charge the car while moving. Obvious in general concept, and would *literally* allow the car to be recharged throughout the country. :wink: Obviously you can't do this under the car's own power, so do it while the car is loaded on a moving train or truck. Unexpected in implementation...

...I don't think this is it. :wink:

Regarding battery swapping, did anyone remember that the Superchargers consist of battery pack units very similar to the Model S battery packs, stacked one on top of another? This may give a hint.

 

[Johan]

Regarding battery swapping, did anyone remember that the Superchargers consist of battery pack units very similar to the Model S battery packs, stacked one on top of another? This may give a hint.

No, chargers. That is the cars have either one (10 kW) or two (20kW) chargers, the SC has the same one's but only 12 of them (120kW output). No batteries in the superchargers as far as anyone knows.

 

[neroden]

I can't see how it works with owner-owned batteries. With leasing, sure, but if you own the battery how can swapping work?

Sure, Tesla warranties the batteries for everything but 'normal' capacity loss.

For eight years only.

I've been on the receiving end of a battery swap from Tesla (for my roadster), and I can tell you that no matter what the Ranger tells you about the replacement battery being of similar capacity / quality, I still ended up with an uncomfortable feeling. The new battery becomes 'mine', and I have no idea how the previous owner treated it.

Yeah; it certainly won't be acceptable to anyone planning to keep the car beyond 8 years. Thanks to resale considerations, that'll make it unacceptable to anyone. People whose batteries are already dying will accept swaps, obviously, but if you know the condition of your battery, why would you ever accept a permanent swap for one of unknown condition?

The only way to do swapping is, as someone else said, to make sure that your owned batteries are never used by anyone else, simply stored while you use the 'loaner' batteries.

- - - Updated - - -

No, chargers. That is the cars have either one (10 kW) or two (20kW) chargers, the SC has the same one's but only 12 of them (120kW output). No batteries in the superchargers as far as anyone knows.

Oh right. Thanks for the correction.

 

[bollar]

Yeah; it certainly won't be acceptable to anyone planning to keep the car beyond 8 years. Thanks to resale considerations, that'll make it unacceptable to anyone. People whose batteries are already dying will accept swaps, obviously, but if you know the condition of your battery, why would you ever accept a permanent swap for one of unknown condition?

The only way to do swapping is, as someone else said, to make sure that your owned batteries are never used by anyone else, simply stored while you use the 'loaner' batteries.

I think we'll find that this is a non-issue in eight years for one of two reasons: Either the batteries are as reliable as Tesla asserts they are (you've no doubt seen them bragging about the pack with 500K miles on it), or the battery technology has advanced significantly enough to make the eight year old pack obsolete anyway.

As for resale, how would one go about documenting that the pack was given pristine care? I assume it's possible to document that it's the original pack, but how about the way it was charged?

As I said in another thread, I think this swapping concept won't be economically feasible unless Tesla can convince enough owners that the batteries are appliances that will just keep working, regardless of how they're treated. I'm sure there's no way to make early adopters and tech users to accept this view, but I do think that the majority of future purchasers will be okay with not keeping the same battery if they choose to participate in swapping.

Personally, I think swapping is kludgy, but maybe I will be swayed by a demo.

 

[stopcrazypp]

Of course it does, do some math with me. If you have a swap station that gets one visitor per hour on average. And, the average time it takes for a owner to come back to get their battery is 2 days (call it a weekend trip), then how many extra batteries do you need on hand to handle the average case?

There's no need to do any arbitrary assumptions. Just start with the assumption that you can't use any owner packs. That means you remove the same number of packs from the whole system as the number of owners needing packs. This might be a huge portion of the network if it's small, but with a large enough network it should be fine. Think of it like a rental car network.

And I think your math is way off for the "average" case. Better place uses 20 packs per station (1 minute swap, 20 minute to fully charge). 1 pack per station means you are not charging any faster than you are swapping.

 

[CapitalistOppressor]

I think your RedBox analogy is a good one, but you glossed over a lot of detail and thereby drew the wrong conclusion. Yes, you can reserve a movie, and that's exactly what I said you should be able to do with a battery. The thing is, RedBox allows you to then return that movie to any RedBox. You don't have to bring it back to the same one. Also, you only have a limited time to pick up your rented movie, before it goes back in the available pool and you get charged anyway. The whole system is built on the assumption that any copy of Iron Man 3 is as good as any other.

I don't see how this supports the idea of storing your battery for you to come back and pick up at all. It sounds a lot more like the analogy supports the description in my OP. We need to be able to make the same assumption with battery swapping to make it work.

You are mistaking the conclusion I was attempting to draw with the Redbox concept. I was only demonstrating the basic concept of reservation and inventory control, which some folks seemed to think was a difficult task. I was not attempting to extend the analogy to the entirety of the debate.

My fault really, because I was responding to you after responding to someone in another thread who was questioning how you could keep track of the batteries at all, or design a robotic storage device at a reasonable cost. The bulk of my response to your comment was the second sentence -

It just requires additional resources be devoted to the swap station infrastructure and a larger inventory for the batteries.

While I did not attempt to support that conclusion, I don't necessarily believe its the wrong one, and I dispute the notion that you have refuted it. If Tesla is going to force customers into a leasing program then it vastly simplifies the logistics of a swapping system. But I don't see an easy path to do that, while accommodating existing customers. If existing customers aren't able to access the system I see it causing all kinds of problems for Tesla. For that matter, Tesla could have adopted a battery leasing program without ever doing battery swap. They chose not to, and having made that choice it seems increasingly difficult to me to back out of it.

Of course it does, do some math with me. If you have a swap station that gets one visitor per hour on average. And, the average time it takes for a owner to come back to get their battery is 2 days (call it a weekend trip), then how many extra batteries do you need on hand to handle the average case?

2d * 24h = 48 packs

You'd need to have 48 batteries. Right? By the time you get your first "swapped" battery back and can give it to a second person, you've had 47 other owners stop by for a battery. Actually, you need 49 because you still need to supercharge that battery you just got back before you can give it to someone else and/or you need one to do the very first swap.

Now, what if you could swap a battery, charge the battery you just got, then give it to the next owner that shows up for a swap? How many batteries do you need stock the station with?

One battery. You need to give a battery to the first person that shows up, but then you can charge their battery in the hour before the next owner needs a swap.

Now, this is only the average case, and uses completely arbitrary parameters. Obviously, you need more than this in both cases to be able to deal with peak travel times. Varying the average time between swaps and average time before an owner returns to claim their battery changes the number of batteries you need on hand, but the relationship is really the same. Storing batteries for owners makes the whole system prohibitively expensive.

Think for a moment about your use case.

We know that Tesla has said that barebones coverage of the U.S. requires something like 100 SuperChargers. Whatever the number is, the requirements for swap stations is the same. So lets just assume 100 to start.

The OHPI estimates that trips exceeding 50 miles represent only ~3% of daily vehicle trips -

Office of Highway Policy Information (OHPI) Highway Finance Data Collection

If we further assume that all of those trips require a swap (unlikely, but lets go with it), then 100 swap stations doing 24 swaps per day implies a vehicle fleet of ~80,000 vehicles in the U.S.

100*24/0.03

If you are charging $20/swap that gives you $17,520,000 in swap revenue per year. (that is an absurdly simplistic methodology. You need to charge based on both miles and time used IMHO, but that is a different debate)

That is not enough to make a profit with a ~$200-$300m investment. But with 80,000 vehicles we are talking a 2017 timeframe before GenIII is released. Once GenIII is released, volumes (and inventory requirements) start to increase nicely.

Your objections seem to be based on the large numbers of batteries that need to be stored, both customer and rental.

Apparently this objection is based on space requirements and the cost of batteries.

As to the cost of batteries, there is a ton of terrible information floating around. The current cost for the 18650 cells that Tesla is using are ranging between $120/kWh and $200/kWh. That works out to between $10,200 and $17,000 for the 85kWh pack.

Sources -

Real wholesale price of kWh li-ion cells, at the end of 2012/start of 2013.
Global Li-ion Cell Shipments to Increase 9% in 2013: IEK-ITIS | CENS.com - Taiwan Industry Updates | Industry In-Focus | HTML |Ta1-CaE-Dy2012/12/03-Id42230

In addition Tesla has repeatedly promised that that their delivered packs would come in under $200/kWh -

Endless-sphere.com View topic - Tesla pack costs $200/kwh? Tipping point price?

My own estimate is that because Tesla appears set to account for 3.6% of global production of 18650 cells this year they are likely getting a discounted price closer to the lower limit than the upper limit.

Regardless, battery prices continue to fall at a rapid rate. Most inventory purchases need to occur in later years, and they are already quite cheap.

If you assume an 85kWh pack costs $12,000, that is an inventory cost of $576,000/station for 48 packs. If those are all built in 2013, it will cost somewhat more than that, but not drastically so, and there is simply no need for that many packs right now. By 2016 when inventory levels are that high pack costs are likely to be much less.

So that leaves space. Vast warehouses will not be required any time soon. Using the regulatory size of the Model S as a proxy (based on how the footprint is calculated, I believe it should approximate the size of the battery. If anyone has measured the battery, chime in), each battery is approximately ~5.5' wide by 9.75' long. You should be able to stack 300 in a space that is 10' x 60' x 10'. In terms of scale, this is a bit larger than a large shipping container.

For a swapping system you can't just stack them, and thus need a support structure, which inflates the volume required by an amount dependent on your engineering solution. But basically these just need to be shelves with rollers, and must be capable of supporting the weight.

Space is not particularly expensive. To the extent that millions of vehicles end up on the road, this can scale however much is needed.

How much space is taken up by the hundreds of thousands of large gasoline storage tanks at existing gas stations? The volume is seems likely to be far higher than will ever be required for a battery swap system, because while gas has a higher energy density, gas stations are required to store enough to provide fuel for almost all vehicle miles traveled, while these swap stations will need to store only enough energy to support a small fraction of vehicle trips. Regardless, if gas stations didn't already exist, the prospect of building enough to support 100m+ vehicles would seem pretty daunting.

Back to your use case, it seems reasonable to me that they could build 100 swap stations for ~$2m/station, including the cost of your 48 batteries. Certainly the specific numbers are debatable, and speculative at this point. But regardless, that $200m in investment represents maybe ~2.5% of Tesla's revenue from building 80,000 cars.

When GenIII comes out the new form factor will necessarily include smaller, cheaper batteries. You will be scaling into a situation where the primary barriers to it are falling. There is no reason to think that this would necessarily be unafordable for Tesla, even ignoring the potentially huge windfall from swap fees as volumes go up.

- - - Updated - - -

There's no need to do any arbitrary assumptions. Just start with the assumption that you can't use any owner packs. That means you remove the same number of packs from the whole system as the number of owners needing packs. This might be a huge portion of the network if it's small, but with a large enough network it should be fine. Think of it like a rental car network.

And I think your math is way off for the "average" case. Better place uses 20 packs per station (1 minute swap, 20 minute to fully charge). 1 pack per station means you are not charging any faster than you are swapping.

I believe he agrees with you and is arguing against a non-lease model. If these batteries are leased, then the logistics for storing batteries are simplified considerably, which I believe was his point.

 

[PRJIM]

Tesla can use the electric motor as a generator. On the Roadster there are thermal limits that only allow for 90A (2.0/2.5). With the liquid cooling it would make sense that these thermal limits may not be an issue.

 

[stopcrazypp]

I believe he agrees with you and is arguing against a non-lease model. If these batteries are leased, then the logistics for storing batteries are simplified considerably, which I believe was his point.

Not really. Our argument is over the number of batteries required for lease vs non-lease. The question is if the number for non-lease really "exponential" compared to lease.

There's two ways to do the math. One way is his way which is to make some assumptions about the average demand per station and then make the pack numbers work out to satisfy that demand. My way is looking at the whole system and figuring the total number of extra packs (this would correspond with number of owners using the network), to be more accurate you would have to look at average owner demand per station and then have the same number of battery packs available there. Then add in the packs necessary for non-lease operation (say for example 20 packs if assuming same operation as Better Place).

If we start with your assumption of 80k vehicles using the network, 3% utilization rate (3% using the network at any one time) that works out to 2400 extra packs for the whole network (which you have to distribute according to owner location). If the network is fairly large, say 100 stations, that's 24 extra packs per station for the loaning on top of the 20 packs that are necessary for normal operation. Obviously if you have very few stations (say 10), then the number needed to be available to loan is huge per station (240).

I think the subtlety here is that you have to account for the fact that in the non-lease case, not every car coming to the station will be carrying their original pack (some of them will be in the middle of a trip with a borrowed pack already attached to the car).

 

[jerry33]

What typical car gets 500 miles on a full tank?? More like 300

My Prius typically gets 650 miles.

 

[Ben W]

I'd say they should build a combination battery-swap-station / touchless car wash. New battery and clean car at the same time!

 

[Citizen-T]

I need some more time to sift through the longer posts and think about them. Good points though from what I've been able to digest so far.

I just wanted to write something down before I forgot about it. In another thread someone compared battery packs in the swapping station to dollar bills. I think this is interesting. What if Tesla set some floor on the quality of the batteries in the system. They would have to do that anyway, right? They don't want to put a battery that is only holding a 70% charge in your car if you need more than that to get to the next Supercharger and/or battery station. What if they made the statement that, any battery that falls below 90% capacity would be pulled out of circulation and refurbished? In the same way that beat up dollar bills are taken out of circulation. Would that be enough to alleviate some of the concerns about giving up your battery?

It might be. By participating in the swapping system, you'd be guarantee that you will forever have a pack with at least 90% of its original capacity in it. Maybe you'll lose the top few percentage points faster because you got a "bad" swap. But, in 10 years the guy who babied his battery is going to have some unknown charging capacity left, while you've still got 90% or more. Regardless of where your battery has been before being swapped into your car.

I'm pretty sure I've heard that refurbishing the batteries is pretty cheap. This cost would just be part of the operating cost of the swap network and be offset by the fee charged to participate.

What do you think?

 

[jeff_adams]

Here's my theory on what could happen, and there is some evidence for it too (patent applications, etc.)

A smaller range extending pack that fits in the frunk. The pack uses one of the new metal-air technologies, therefore it is significantly lighter (from an energy density point of view) and may extend the range by say, 100 miles.

This turns the model S into a hybrid-battery car. The range extending pack keeps the main pack charged longer and doesn't drive the car directly so to speak.

Now maybe the range extending pack can be swapped out at the supercharger station for those customers who are in a big hurry. But the ability to have 100 miles more range opens up a LOT of possibilities and lessens the burden on number of SC stations needed in the first place (spacing, distance, etc.) I know having 100 miles of extra range for me means the difference between a trip to the coast without a stop, or not.

This should have a seperate thread. I believe this is the most logical answer I've heard. Think about these applications of a "trickle charger" mounted in the frunk cubby.

*No more Brodering. You always have spare power with you if you run into bad weather or detours. Worse case, you have to pull over for a while to charge up enough.

*You get home late with little charge only to discover power outage at home. No problem, you let the Air Metal run while you sleep and you have a charged up car in the morning.

*You go to the lake to fish. No power outlet, no problem again. Camp all weekend and drive back on a full charge.

*Extended distance travel. Turn on the Air Metal once the battery is down 10% and trickle charge as you drive. Can't do this with an ICE!

What I like about this solution:

You don't need swapping stations. You would simply go to existing service centers or have a Ranger come install it. It would be rated on "hours of charge", not how many miles you could drive. Software would keep track and you would know when it's time to replace. Because it's not designed to put out enough power to "drive" the car, it can be light enough to not cause weight or storage issues.

 

[ModelS8794]

Cross posting from another thread, since this looks like the more appropriate spot for it. Perhaps "unexpected in implementation" refers not to the mechanics of physically swapping, but to the way the swapping network is made financially feasible. In that vein...

i wonder if any of the brains here can help me with this. I have a notion that battery swap can be incorporated into a system that's designed for 2 uses instead of one: Tesla long distance travel, and demand response/ utility rate arbitrage as well. My question is, is the capacity of the 85kwh battery sufficient that they could be utilized at some scale that makes it meaningful to a utility as a demand response system, but also built in a location/environment that is sensible for Tesla drivers looking for long distance travel swaps? I'm thinking something like enterprise-scale version of the Solar City/Tesla Home Energy Storage solution.

 

[Citizen-T]

You are mistaking the conclusion I was attempting to draw with the Redbox concept. I was only demonstrating the basic concept of reservation and inventory control, which some folks seemed to think was a difficult task. I was not attempting to extend the analogy to the entirety of the debate.

Ah, yes. Much wasted breath on my part. I see your intention now and you are right, this is a solvable engineering problem and RedBox is a good example. I agree.

If Tesla is going to force customers into a leasing program then it vastly simplifies the logistics of a swapping system. But I don't see an easy path to do that, while accommodating existing customers.

Agreed. I'm not arguing for leasing. I think Tesla would prefer that you not care about the battery. Their recent moves seem to support that. I think they want you to stop thinking that you car is your battery. If they can accomplish that in some way, then I don't think people will mind having their battery swapped any more than they mind having an air filter changed.

I don't think we are there yet, not by a long shot. But maybe the "dollar bill" theory I posted above gets them a step closer.

We know that Tesla has said that barebones coverage of the U.S. requires something like 100 SuperChargers. Whatever the number is, the requirements for swap stations is the same. So lets just assume 100 to start.

Not the same in my OP. I suggested that only a subset of Superchargers needed to be upgraded to battery swap. I don't know how many, maybe the top 25% most used? <== complete guess

Apparently this objection is based on space requirements and the cost of batteries.

Mostly true. Also there is the added cost of building the system that as a requirement needs to store customer batteries. There is going to be incrimental cost in the design, software, and hardware of such a system over a simpler one that treats all batteries as interchangable.

I'm a software guy, so just thinking through all the edge cases that the system would need to account for makes me think there is a lot of incremental cost in adding that requirement.

The current cost for the 18650 cells that Tesla is using are ranging between $120/kWh and $200/kWh. That works out to between $10,200 and $17,000 for the 85kWh pack.

There is a lot more to the pack than just the cells. Tesla is selling packs 8 years from now for $12k. If we work backwards from there assuming 8% price per kWh improvement each year (as per Elon's comments) that gives us a price of something like $23k today. So the cell cost might only be something like 65% the cost of a whole pack.

That doesn't ruin your argument. You're right, it is a price that can be stomached, especially once you take into account my point that a subset of Supercharging stations should be swap stations. I'm still not sure it scales though. I'll so some real math on that one.

The root of my original argument is really that no matter what assumptions you make, swapping with any battery is a lot cheaper than storing batteries for owners. If you can convince owners that any battery in the system is as good as any other, then the system is a lot cheaper.

 

[cwerdna]

What typical car gets 500 miles on a full tank?? More like 300

Depends on the tank size, your mileage and how low you want to go.

Take look at Compare Side-by-Side. Or, take the combined mileages and multiply the tank sizes.

 

[stopcrazypp]

It might be. By participating in the swapping system, you'd be guarantee that you will forever have a pack with at least 90% of its original capacity in it. Maybe you'll lose the top few percentage points faster because you got a "bad" swap. But, in 10 years the guy who babied his battery is going to have some unknown charging capacity left, while you've still got 90% or more. Regardless of where your battery has been before being swapped into your car.

I'm pretty sure I've heard that refurbishing the batteries is pretty cheap. This cost would just be part of the operating cost of the swap network and be offset by the fee charged to participate.

I'm not sure how much cheaper this would be compared to the battery loan case. While it would take about 870 full cycles (~180k miles, using 212 miles/cycle average) for a pack to reach 70% capacity, it will take only about 80 full cycles (~20k miles using 265 miles/cycle average) for a pack to reach 90% capacity, since the degradation is not linear (happens much faster in the start).
http://www.teslamotorsclub.com/show...evealed/page13?p=172284&viewfull=1#post172284

That implies an average 1.5 years of use before refurbishing or replacement.

And "refurbishing" a pack requires replacing the cells ones that are 90% capacity or higher. That means you can only salvage cells from packs with ~20k miles or less that failed for other reasons. I doubt there will be many failed packs that satisfy this criteria, so most of the packs will not be refurbish-able to 90% capacity without using brand new cells. You can not mix cells with lower capacity (say 70%) with ones with higher capacity (say 100%) and hope to reach 90% overall for the pack because the nature of cell balancing will limit you to the cell with the lowest capacity (you will end up with a pack with 70% capacity).

This is the reason why even battery lease schemes usually use a 75% (Renault) to 80% (Smart eD) limit for replacement.

 

[Citizen-T]

I'm not sure how much cheaper this would be compared to the battery loan case. While it would take about 870 full cycles (~180k miles, using 212 miles/cycle average) for a pack to reach 70% capacity, it will take only about 80 full cycles (~20k miles using 265 miles/cycle average) for a pack to reach 90% capacity, since the degradation is not linear (happens much faster in the start).
http://www.teslamotorsclub.com/show...evealed/page13?p=172284&viewfull=1#post172284

That implies an average 1.5 years of use before refurbishing or replacement.

And "refurbishing" a pack requires replacing the cells ones that are 90% capacity or higher. That means you can only salvage cells from packs with ~20k miles or less that failed for other reasons. I doubt there will be many failed packs that satisfy this criteria, so most of the packs will not be refurbish-able to 90% capacity without using brand new cells. You can not mix cells with lower capacity (say 70%) with ones with higher capacity (say 100%) and hope to reach 90% overall for the pack because the nature of cell balancing will limit you to the cell with the lowest capacity (you will end up with a pack with 70% capacity).

This is the reason why even battery lease schemes usually use a 75% (Renault) to 80% (Smart eD) limit for replacement.

Thanks for that. Good information. I'll ponder this some more.

In the mean time I tried to model a comparison between a scheme that stores an owner's battery versus one that lets anyone take any battery. The formulas come from my earlier (admittedly simple) use case, but the values have been abstracted so I can change any parameter to see the comparison.

The formulas are:

p = v * t + 2 (for the case where we store customer packs and return them)
p = v * c + 1 (for the case where any pack can be swapped for any other)

p = packs per station
v = visitors per min
t = average time before a customer returns to get their pack
c = average time it takes to charge a pack

Each formula adds one pack because you need to have a pack to give your first visitor. The first formula adds an additional pack, because when your first visitor returns, you can't immediate give that pack out, you still need to charge it.

The result looks like this:

When t=2 days and c=45 mins

I don't pretend that these formulas are perfect. If you have a suggestion on how to improve the model let me know maybe I can enhance it.

The bottom line is, if my model is even remotely correct, then one model is far more scaleable than the other. Again, this still doesn't take into account any of the added complexity inherent in the system that stores packs for customers.

Edit: Just noticed there are no values on the X axis. The chart starts at 1 visitor every 100 mins and goes to 1 visitor a min.