Chevrolet Bolt

[wwhitney]

I considered these same estimates and decided they are not appropriate,
because you are starting from an assumption that the risk is evenly distributed. I would not take that leap of faith

I mentioned different models beyond the constant chance per unit time model, so I am not assuming anything. But that model gives you a rough estimate of the order of magnitude of the risk. I can imagine situations where the current risk is 10 times the average historical frequency, but it seems unlikely that is 100 times the average historical frequency.

For example, I think we can discount the model where all the batteries are programmed to burst into fire on 12/31/2020, and that 5 of them had a bug and executed self-conflagration prematurely. : - ) But that would certainly be a situation where the observed historical frequency would not be a good predictor of the future risk.

Cheers, Wayne

[cwerdna]

In one of the Bolt FB groups, there was an ~1.5 page FAQ floating around that did say that GM intends to find the root cause and remove the 90% charge limit once the investigation has been completed and defective battery packs have been identified and repaired or replaced.

It's at https://www.facebook.com/photo.php?fbid ... 094&type=3, but I think people not in that group won't be able to see it. There's also a question and answer about why some '19 and all '20 Bolts aren't involved in the recall. It claims select '19 vehicles were built using cells at LG Chem's Ochang, Korea facility and they believe that's where the issue developed. And, that the verified incidents investigated so far had batteries from there.

[SageBrush]

I considered these same estimates and decided they are not appropriate,
because you are starting from an assumption that the risk is evenly distributed. I would not take that leap of faith

I mentioned different models beyond the constant chance per unit time model, so I am not assuming anything. But that model gives you a rough estimate of the order of magnitude of the risk. I can imagine situations where the current risk is 10 times the average historical frequency, but it seems unlikely that is 100 times the average historical frequency.

For example, I think we can discount the model where all the batteries are programmed to burst into fire on 12/31/2020, and that 5 of them had a bug and executed self-conflagration prematurely. : - ) But that would certainly be a situation where the observed historical frequency would not be a good predictor of the future risk.

Cheers, Wayne

Cute humor aside, can you exclude environmental variables ? Usage variables ? Production variables ? Supply variables ? Each would be a subset of the entire group. It is just not reasonable to homogenize the fleet

Moreover, there are no good reasons to think that reported fire numbers will not increase, and at a different rate than seen thus far.

[wwhitney]

It is just not reasonable to homogenize the fleet

I'm going to stand by my statement that it is a reasonable starting point for risk estimates, with the caveat there could be some emergent phenomenon that will cause the future frequency to exceed the observed historical frequency by up to a factor of ten. I would be amazed if the frequency suddenly jumped by a factor of 100, I don't find that a possibility worth worrying about.

Cheers, Wayne

P.S. My initial comment was in response to the claim that we have "no idea" what the risk is, and I'm just pointing out we can easily get some idea, to within an order of magnitude at least.

[cwerdna]

https://www.nhtsa.gov/vehicle/2019/CHEV ... WD#recalls has a lot more associated documents now (10, at the moment). I'm trying to work my way thru them now.

I saw a FAQ floating around before (screenshot and the posted in a private Bolt FB group). I see a a copy at https://static.nhtsa.gov/odi/rcl/2020/R ... 1-2689.pdf but it looks like a slightly older copy with only 14 Q&A instead of 16.

It's at https://www.facebook.com/photo.php?fbid ... 094&type=3, but I think people not in that group won't be able to see it. There's also a question and answer about why some '19 and all '20 Bolts aren't involved in the recall. It claims select '19 vehicles were built using cells at LG Chem's Ochang, Korea facility and they believe that's where the issue developed. And, that the verified incidents investigated so far had batteries from there.

From the above doc, there is:

Q12) Why aren’t some 2019 and all 2020 BOLT EVs involved in this recall? Is there a different battery?
A12) Select vehicles in the 2019 model year were built using battery cells produced at LG Chem’s Ochang, Korea, facility, which is where we believe the issue developed. The verified incidents investigated thus far had batteries produced at this facility. Additionally, the 2020 Bolt EV uses a different cell design than the vehicles affected by this recall, which enabled us to increase range to 259 miles.

The My Chevrolet app still shows no recalls for my specific VIN of my '19 Bolt. https://www.nhtsa.gov/recalls still shows 0 unrepaired recalls for my VIN.

[cwerdna]

To quote from boltage at https://www.chevybolt.org/threads/ev-re ... eys.36819/, he pointed to https://www.consumerreports.org/car-rel ... able-cars/ and

85 Chevrolet Bolt
54 Hyundai Kona Electric
54 Nissan Leaf
53 Tesla Model 3
31 Tesla Model X
26 Tesla Model S
18 Kia Niro EV
12 Audi E-Tron
5 Tesla Model Y

It does seem odd that the EV versions of the Kona and Niro do much worse than the non-EV versions. In contrast, the Bolt is the most reliable Chevrolet model in the surveys.

I didn't realize the ratings at https://www.consumerreports.org/car-rel ... able-cars/ are actually visible to those w/o a CR account until I tried (while not signed in). I can confirm his numbers are right and you can too.

From going thru the GM nameplates there, seems like the Bolt is the 2nd most reliable GM vehicle listed there.

[SageBrush]

This may be old news ..
GM has issued a recall for 68k Bolts due to fire risk, and I think is now recommending that all Bolts not exceed 90% SoC

This link
https://www.theverge.com/2020/12/2/2213 ... on-lawsuit
has some details although the reason for the "news" was to report the filing of a class action lawsuit.

So far as I know, GM has identified a *correlation* between high SoC, a specific factory, and fires. The cause(s) remain unknown, so other Bolts may turn out to be fire risks also and the 90% SoC advice may turn out to be an inadequate or ineffective measure.

[WetEV]

This may be old news ..
GM has issued a recall for 68k Bolts due to fire risk, and I think is now recommending that all Bolts not exceed 90% SoC

This link
https://www.theverge.com/2020/12/2/2213 ... on-lawsuit
has some details although the reason for the "news" was to report the filing of a class action lawsuit.

So far as I know, GM has identified a *correlation* between high SoC, a specific factory, and fires. The cause(s) remain unknown, so other Bolts may turn out to be fire risks also and the 90% SoC advice may turn out to be an inadequate or ineffective measure.

In other news, FUD spreading has helped Elon become the second richest person in the USA. Keep it up and perhaps he might be the first Trillionaire.

[cwerdna]

To quote from boltage at https://www.chevybolt.org/threads/ev-re ... eys.36819/, he pointed to https://www.consumerreports.org/car-rel ... able-cars/ and

85 Chevrolet Bolt
54 Hyundai Kona Electric
54 Nissan Leaf
53 Tesla Model 3
31 Tesla Model X
26 Tesla Model S
18 Kia Niro EV
12 Audi E-Tron
5 Tesla Model Y

It does seem odd that the EV versions of the Kona and Niro do much worse than the non-EV versions. In contrast, the Bolt is the most reliable Chevrolet model in the surveys.

I didn't realize the ratings at https://www.consumerreports.org/car-rel ... able-cars/ are actually visible to those w/o a CR account until I tried (while not signed in). I can confirm his numbers are right and you can too.

From going thru the GM nameplates there, seems like the Bolt is the 2nd most reliable GM vehicle listed there.

Seems like the CR page's ratings are behind a paywall now. I'm guessing that was the original intent.

https://web.archive.org/web/20201123161 ... able-cars/ has a copy not blocked by paywall.

[jlv]

Seems like the CR page's ratings are behind a paywall now. I'm guessing that was the original intent.

I have access to CR's online subscription content via my public library; you might be able to get it that way, too.

15-30 years ago when CR was my go to resource for researching what to buy. Sadly, I find the current reviews and ratings to be far from useful. Having access to the on-line content only helps to show me I wouldn't be missing much if my library dropped it, because I wouldn't pay for what they now offer. My current print subscription runs through the end of 2021, and I'll likely not renew.