Potential Technical Challenges With Automated Battery Swaps

[CapitalistOppressor]

I don't have time to delve into this today, but that's no reason not to kick off the discussion. Tesla is announcing Battery Swap, and if its automated (as I expect) there are real potential issues. Above all else is reliability. Nobody wants to get stuck on a SuperSwapper with the baby half delivered.

So what is our model? Elevators? Car Washes? I hope it's not rollercoasters.

What data do we have? Did Better Place publish anything?

I'm not really interested in whether its possible. Tesla has claimed the capability from the beginning, and Better Place built a working system which did not fail (to my knowledge) due to technical problems.

 

[BenTesla]

Well I think until the event A LOT of questions will have to remain unanswered. But from my understanding of the way Better Place did their system in Israel, is that it's sort a car wash type thing.

My real question would be how they will make this system work in a country as big as the US. Better Placed failed/is failing, how does Tesla do anything different enough to be suggessfull but on a much larger scale?

The only legitimate reasoning i've heard behind all this is to use this tech in China. China wants EV's (badly) and they've going to need something like this so why shouldn't Tesla throw it's hat in the ring?

 

[JPP]

Having been to the factory and seen how the battery is installed on a Model S, I don't see an obvious way to make this battery system in this vehicle automatically swappable. There are a bunch of bolts (like 15-16) holding a heavy 1000 lb structure, plus there are electrical connections and coolant lines, all of which get the human touch on the assembly line.

 

[ItsNotAboutTheMoney]

Having been to the factory and seen how the battery is installed on a Model S, I don't see an obvious way to make this battery system in this vehicle automatically swappable. There are a bunch of bolts (like 15-16) holding a heavy 1000 lb structure, plus there are electrical connections and coolant lines, all of which get the human touch on the assembly line.

Having seen my 3, 000lb Prius on a car lift, I don't think the battery weight is a problem. Automating connections could be, although I don't see why it should be. In an assembly line, humans not only do assembly but can act as QA. Do you know whether fhey are installing rather than checking?

 

[arg]

So what is our model? Elevators? Car Washes? I hope it's not rollercoasters.

If you envisage it running unattended, safety seems to me the biggest technical challenge. Rollercoasters are easier because they are always supervised. Elevators are a good example of the situation (because they are often used totally unsupervised in 'private'), but the solutions there are much easier. Car washes and driverless metros are closer to the degree of risk, but tend to be at least lightly supervised - you feel intimidated from poking around to see how it works, even if there's not a mechanism actively preventing you from doing so. Trains also have the advantage of having been around a long time and we are supposed to know that they are dangerous.

My initial thought was risk to kids etc. passing by, but you probably make the driver responsible for checking the station is clear of other people before pressing the 'go' button; practical risk is probably more the driver having a look underneath to see what's going on, or worse still trying to fix problems when it has (or appears to have) gone wrong.


There's other challenges more in the category of hard to do for a respectable cost, rather than hard to do in principle. Reliability issues caused by weather etc. are a case in point - if you put the thing under cover, you have a high-pressure hot-water wash that cleans the underside of the car before starting, and a vision system to check the cleaning has been effective before starting swap operations, then it seems not too challenging to design. If it's sitting naked in a parking space in winter, then that's a tougher problem.


But I think you were much more on the right track with one of your earlier posts where you said this announcement would be more about business innovation than engineering. I'm much more expecting to see some innovative business plan that finds a way to exploit what is easily possible in engineering terms, rather than a big engineering breakthrough. So, automated at least to some degree but probably still supervised and maybe some minor manual handling steps. Innovation in somehow arranging that you can afford technicians hanging around at the locations where people actually want battery swaps. Maybe with ranger service you can make sense of having service centres at highway rest stops rather than in cities?

 

[Kaivball]

Is the battery and all related parts installed by robot in the factory today?

Me thinks, if they don't even use automated installation in the factory, how can we reliably expect it in the field?

 

[Rla]

Better Placed failed/is failing, how does Tesla do anything different enough to be successful but on a much larger scale?

Imho the reason BP failed is because there were no customers; not because of the tech. Also the cars did not have a fast-charge capability.
TM has the customers, they already have home-charging, supercharging. So there is 'no' risk of Tesla failing because of swapping. There is only opportunity to add value (huge value) for achieving faster refuelling than ICE's.

 

[NigelM]

Tesla is announcing Battery Swap.....

Source? That's a statement.

 

[JPP]

Kaivball--NOT robotically installed. As I have said before, for all of you postulating and speculating, go on a factory tour to see how it is done. The only robot involved is a self driving cart that brings the battery pack to the assembly line and lifts it into position under the vehicle. Then a team of techs actually bolts the pack into place and makes the electrical and coolant connections. As I said, IMHO in the Model S as currently designed, there is essentially no likelihood of designing a fully automated battery swap/install system. Now with proper engineering and the integration of quick connect/disconnect electrical and coolant fittings, sure it would be possible.

 

[Discoducky]

Having been to the factory and seen how the battery is installed on a Model S, I don't see an obvious way to make this battery system in this vehicle automatically swappable. There are a bunch of bolts (like 15-16) holding a heavy 1000 lb structure, plus there are electrical connections and coolant lines, all of which get the human touch on the assembly line.

I don't think the bolts are an issue for automation, but other plastic connections could be. Do you have a source for your knowledge that you can share on these lines? Would be interested to see where the connection lies and what the connector looks like. And I'm sure TM would automate the installation of the battery in the factory first prior to installing a robot to do it at a swap station.

MTTF of bolts could be mitigated if the female end is replaceable as I'm sure the male end (bolt) is replaceable. But MTTF (sorry, Mean Time To Failure) of the connectors and wires, if installed by automation, could potentially be much lower over the lifetime of swaps.

*IF* the 5th announcement is swapping I imagine that taxis, police, delivery vehicles will really take notice.

 

[jerry33]

Kaivball--NOT robotically installed. As I have said before, for all of you postulating and speculating, go on a factory tour to see how it is done. The only robot involved is a self driving cart that brings the battery pack to the assembly line and lifts it into position under the vehicle. Then a team of techs actually bolts the pack into place and makes the electrical and coolant connections. As I said, IMHO in the Model S as currently designed, there is essentially no likelihood of designing a fully automated battery swap/install system. Now with proper engineering and the integration of quick connect/disconnect electrical and coolant fittings, sure it would be possible.

The big problem I see with robots is that they generally like things pristine. The bolts on the underside of the car are unlikely to be clean and may be caked with mud. I have difficulty seeing how a fully automatic system could work reliably. Even 85% reliable wouldn't be good enough.

 

[brianman]

MTTF ... could be mitigated if the female end is replaceable as I'm sure the male end (bolt) is replaceable.

No wonder the divorce rate is so high.

Not as funny as my original joke, but less troublesome.

 

[Kaivball]

Kaivball--NOT robotically installed. As I have said before, for all of you postulating and speculating, go on a factory tour to see how it is done. The only robot involved is a self driving cart that brings the battery pack to the assembly line and lifts it into position under the vehicle. Then a team of techs actually bolts the pack into place and makes the electrical and coolant connections. As I said, IMHO in the Model S as currently designed, there is essentially no likelihood of designing a fully automated battery swap/install system. Now with proper engineering and the integration of quick connect/disconnect electrical and coolant fittings, sure it would be possible.

I know...



I have done the factory tour. I know the answer.

I was wondering if those at support battery swapping have seen the factory and know that the battery is not even installed by robots in the factory.

You'd think that they perfect this step in the factory first.

I have alway doubted battery swapping.

 

[brianoflondon]

Kinda funny reading all this stuff. I’ve sat through 90+ automated Better Place switches without a hitch. Over the network (while the receiver works out what to do following the bankruptcy) we’re still seeing 300+ switches per day and failure rates are very low. The question is not if it can be done, it’s how cheaply can it be done reliably.

 

[NigelM]

All the focus seems to be on how fast/easily can the battery be swapped. Another question (more relevant technically?) is how will the stored batteries be recharged? This will require coolant hook-ups, pumps, thermal monitoring and high power (super?) chargers in multiple quantities. Never mind storage and portability issues.

 

[scaesare]

Several concerns about the logistics of battery swapping are mentioned in this thread. Given that Tesla and Elon have both stated that battery-swapping was a design consideration, I'm going to assume they've thought about this.

• BOLTS: Is there any reason to believe that bolts couldn't be swapped out for a quick-connect type fastener more apropos for automated handling? There are many designs out there...
• COOLANT CONNECTORS: The pictures of the pack I've seen, as well as descriptions written, describe them as quick-disconnects.
• ELECTRICAL CONNECTORS: Those also appear to be single plug... ostensibly easy and quick to disconnect.
• CONNECTOR LIFETIME: Why is there any reason to believe the pack connector would be designed with any less rated insertion cycles than the charge port, which would get far more use? Similarly, quick disconnects for fluid couplers can be rated in the 10's of thousands of cycles.

While there are certainly many logistical challenges in swapping batteries, it doesn't seem like many of the criticisms I've seen mentioned here are real show stoppers.

As a note, in reference to the "recharge faster than refueling" announcement that it's "obvious in general concept, but unexpected in implementation." So expect something novel...

 

[Discoducky]

Well, looks like we'll all get a look in a few days. Going to be interesting to see if it's done all by human, mostly by human, some by human or none by human. And hopefully they'll talk about logistics and economics for implementation. Could, however, just be a demonstration...

 

[brianman]

we’re still seeing 300+ switches per day and failure rates are very low.

Can you elaborate on the user experience for the failure cases?

 

[Palpatine]

I think that technically it is not an issue. It can certainly be done in the time frame that Elon has said. "Faster than it takes to fill a gas tank"

But then the variables start. Cars collect dirt, road grime, mud, ice/snow, water when it is raining, etc.
Will any of those issues make a battery swap impossible?

For example, let's say it is raining really hard outside and my tires and underbody of the car are dripping water. Will I still be allowed to roll over the top of the battery swap machinery?
Logically a lot of water would begin to collect down into that area where batteries are moving around.

A fully automated system with no humans around to keep it clean seems unlikely. So once we need humans to keep it clean or regularly maintain the swap station, that raises costs dramatically.

So while this works in testing in Southern California with very few rain days and 70 degree weather.... what about in the Pacific Northwest where it rains for 6 months of the year? Or in New York with snow and ice caked around the connection points of the battery pack?

 

[CapitalistOppressor]

This will all be done with a jig, so no need to swap bolts out. Bolts are handy and won't slow anything down.

I hadn't thought of doing a quick pressure wash to the undercarriage. That's a good idea, though possibly not necessary. In my experience, the worst caked on goo tends to be a result of leaky lubricants from the car itself collecting dust and grime. In areas not exposed to local sprays of lubricant, there often is not much grime. Of course, I live in Calofornia, which is a valid criticism. So a pressure wash might be useful, especially in northern states with salted, snowy, muddy roads.

The mechanism itself will be simple and robust. Dirt and grime should not affect the mechanics of it. A sump and proper drainage should be fine.

Speculation about details aside, I am interested most in reliability. I think it can be engineered up to a high degree of reliability, but there will be failures. If we can get data about Better Place, that would be ideal. But if this failed during winter in Montana, the customer will be stuck in a car that is unable to heat itself. So it's immediately a big problem.

I don't like the idea of attendants, but it seems to be necessary for at least many locations. Of course this also appears like it will roll out initially in California because of the CES incentives. It might not go nationwide at all if demand is low, but instead may just be added as a high throughput supplement to high traffic corridors. We'll have to wait and see.