This could be for any number of reasons, and also a big coincidence, and I don't have time or the interest to read more on it, but the thing that I kept thinking while I read the story was "why did you have to do that thing with Boeing... you don't look too good now".
Lets see the amount of fires of, lets say, the Ford Fusion in the past 6 weeks. Or maybe the amount the fires for all of Ford vehicles. It's only because Tesla is so focused on.
This kind of reminds me how the media reports every single airplane crash. How about report every car crash with fatalities. According to Wikipedia over 2,500 per month.
Gasoline gars don't catch on fire in non-fatal accident with any frequency
Now let's talk numbers.
Ford Fusion fleet, in the US: 1,520,000 vehicles.
Tesla Model S fleet, world: 18,300. If 300 Ford Fusions caught on fire in a 6 week span I guarantee NHTSA would be investigating.
A: selective editing
B: There are a ~11 million auto accidents (that's accidents, not vehicles involved in accidents) in the US each year. ~10.9 million of them don't catch on fire.
Unfair comparison still. How many of those cars were old and in disrepair? How many of them were non-fatal accidents? How many of them were due to running over debris?
You have a point with the old and disrepair but fatal accidents is a bad classifier because what proof do you have that fatal accidents of other cars are fatal accidents of a model S and by fatal I am assuming you mean fatal not due the fire.
As for running over debris, again it is not important because who cares if gasoline cars have different causes for their fires vs the causes of the model S. What matters is the percentage that actually do catch on fire with a large enough population.
After that, the next question that should be asked is, what is the difference between the danger of these fires. Are gasoline fires more or less dangerous.
Are there stats around how many traditional cars exploded because of road debris hitting the gasoline tank? That's the key point here - everything can explode but what is the relative likelihood of a traditional car exploding because of road debris (as seems to be the case with the Tesla)? May be the gasoline tanks on cars are better shielded than the batteries on Tesla are.
It's _extremely_ rare for a gas tank to explode. The only way that could happen is if it pressurizes to the point that the tank itself explodes from the pressure.
There are pressure relief valves on the tank that prevent that from happening in the vast majority of cases.
Exactly right. It might very well be that the way traditional cars are designed the gasolene tank is well shielded from debris than the batteries in the Tesla are. May be Tesla needs to find better, stronger compartment for the battery.
It's not clear whether the Tesla catches fire more often than other cars from road debris, 2 instances could be coincidence (I've never run over something that caused actual damage to my car after years of driving, crosses-finger).
However, that is the government's job, to examine the fire and determine whether they should have been prevented.
Well for comparison, the Chevy Volt and Nissan Leaf have never had a fire. This is despite the two combined having sold over 100,000 cars. Meanwhile the Tesla S has only sold about 19,000 cars.
> Well for comparison, the Chevy Volt and Nissan Leaf have never had a fire.
The volt had one, rather famously when it was tested by the NHTSA. In that test (which was pretty ludicrous if you dig into the details) as in the recent Tesla crashes, I don't think they're really reflective of the engineering quality of the car.
The comparisons to the Leaf and Volt are meaningless if none of those vehicles have been in similar accidents. You can't just point to numbers and imply that a lack of reported fires means they're safer.
> The comparisons to the Leaf and Volt are meaningless if none of those vehicles have been in similar accidents. You can't just point to numbers and imply that a lack of reported fires means they're safer.
So...the question is...well, why should we assume that the problem is underreporting...that it's just Tesla's that are getting all the attention? That's not even true:
> DETROIT — Nissan inadvertently gained some valuable insight into the durability of its electric car, the Leaf, when about two dozen of them were destroyed in the tsunami that ravaged Japan in March.
>
> None of the cars caught fire, and their batteries remained fully intact, shielded by an airtight steel exoskeleton and two other layers of protection that surround the 660-pound packs.
“Considering how they were tossed around and crushed, we think that is a very good indication of the safety performance of that vehicle,” said Bob Yakushi, the director of product safety for Nissan North America.
Nissan’s decision to encase the Leaf’s battery in steel may help explain why federal safety regulators investigating postcrash fire risks in the Chevrolet Volt do not have the same concerns about the Leaf. General Motors packages the Volt’s battery cells on a T-shaped steel tray with a plastic cover.
They might appear safer given the available evidence, but if it's a small enough difference, we will need more data to discern the difference with any reliability.
We should expect more fires given enough time. I wonder for how many they will continue to report on every single one?
two of the three fires were from objects hitting the underside of the car.
Nissan Leaf has sold 3x more vehicles than Tesla.
Are you saying that objects don't hit the underside of Leafs? (Perhaps they are repelled magically?) Or that objects are somehow more attracted to the underside of Teslas cars? Or the Leaf accidents don't count because the Leaf owners didn't file a report that nothing happened?
This might just be the case, simply due to the extended range of the Tesla. Since you can commute farther, you're more likely to be on a highway in a Tesla. I haven't seen many Leafs on 101/280, but I see Teslas all the time.
Since these are high-energy impacts, you'd have to be traveling at highway speed to have a damaging impact.
It might be more interesting to compare how a similar car style would behave (think high performance sedans).
Tesla is prevalent in CA, particularly in the SF Bay area. Nissan manufacturers the Leaf battery in TN, so it's more prevalent in the Southeast. I see them all the time in Atlanta. It helps that GA has a $5000 state tax credit on top of the $7500 federal tax credit.
I have seen a couple leafs on my drive to work. I know of 2 Teslas in my neighborhood and have commonly seen a third between my house and work. I know they are different ones because 2 of them have vanity plates.
I was surprised to see them because I don't consider where I live to be a big Tesla type area.
There may be a problem. We can consider the Tesla as a car body bolted on top of a huge battery. Collisions with road debris seem to be sufficient to pierce the battery's armour and start a fire.
This isn't disastrous. Problems like this routinely emerge with new designs. And at least the flammable materials stay in one place (unlike a gasoline engine). But still, a fire's a fire, and I'm sure Tesla's engineers are working overtime trying to figure out a solution.
> and I'm sure Tesla's engineers are working overtime trying to figure out a solution.
What could the solution possibly be? Puncture the battery and it catches fire, and the battery covers almost the entire bottom of the car. Practically speaking there's only so much steel you can put on the bottom of the car.
It seems that Model S spreading the battery across the bottom of the car may simply be a defective design.
The conditions that cause the fires seem to be narrow and well-defined enough that a clever workaround is possible. Maybe add an extra layer of sacrificial armour? Or move some cells away from the vulnerable region.
Where is this claim coming from? Wired? I can't find any source. To me it seems like some writer for Wired figured they could get some clicks by speculating.
The National Highway Traffic Safety Administration “absolutely has to investigate” the Nov. 6 Tennessee incident, Clarence Ditlow, executive director of the Center for Auto Safety, based in Washington, said in a phone interview.
Apparently the battery pack in a Tesla has liquid cooling so it probably gets fairly hot. In a gas car, the liquid cooling is for the engine and if it fails, the engine overheats and stops. But imagine if the gas tank itself got hot when running and needed to be actively cooled. That's basically where the Tesla is at. A failed cooling system means the energy source overheats (and not the motor/engine).
I'm not a car expert but I imagine it's hard to protect all the pipes that are carrying the liquid. And that may be why these fires are starting... some puncture causes the liquid to release. Just a theory.
I would assume a car that has been involved in a crash would stop moving, and not need that cooling any more. Even if there weren't failsafe systems to shut things off.
There's only so much you can do with a punctured lithium ion.
What do you mean failsafe? That the batteries would shut down? My understanding is that in two of the incidents the drivers were able to continue driving until they found a safe place to get off. I think there would be a safety issue if the car completely shut off power on a freeway due to the cooling system failing.
The battery cooling prolongs the life of the batteries but they won't go critical or something without it.
The presence or absence of that liquid won't affect the outcome of what's happening in these accidents, which involves driving steel into the battery packs, compromising the batteries and causing a short circuit.
150,000 car fires per year / 52 weeks is 2880 fires per week. There are 135900000[1] vehicles in the US, which gives 21.2 fires per million-vehicles week. With 19000 Teslas on the road and three fires in six weeks we have 0.5 (fires / week) / 0.019 (million vehicles) = 26.3 fires per million-vehicles week. So even if this is typical for the Model S (and it doesn't seem to be), the cars aren't much more likely to catch fire than ordinary cars. Also, I haven't accounted for the fact that the Model S's probably get driven more than your average passenger car, both due to low per-mile costs and the fact that they're more likely to be new and thus valued by their owners. Taking this into account would probably bring the Model S's fire-rate below gas cars again.
The trouble is that there are a lot of ways to slice the data, and just not a lot of data yet.
For example, I wonder what happens when you group the Tesla S in with cars that can do the quarter mile in under 13 seconds... http://www.wreckedexotics.com/
Now that they did catch on fire, and the fire was suppressed every time, and everything was just ok... Well, now I'm beginning to trust electrical cars.
This story is unfortunate spin for two reasons. The minor reason is Wired appears to be completely speculating when they state, "Could Spark Federal Inquiry".
The second and really much larger reason this article is spin is that they are painting this as a major issue without giving enough information to quantify it. I try to quantify it in this comment and the results are the opposite of how Wired is spinning this issue. Here's what they said:
While there are an average of 150,000 car fires in the U.S. each year, with only 19,000 Model S sedans on the road, having three vehicles catch fire in six weeks has garnered plenty of attention.
So this is comparing two data sets:
A: % of Tesla Vehicles that catch fire. Dividing 3 by 19,000 leads to a fire incidence rate of 0.0158%.
B: % of all vehicles that catch fire. For this data set, Wired helpfully tells you the numerator of 150,000 vehicles catch fire in the U.S. but completely leaves out the denominator.
Without giving the complete information for data set B, it's difficult to determine if Tesla's fire incidence rate is out of the ordinary and warrants the speculation of, "Could Spark Federal Inquiry". So in an attempt to find the denominator, I discovered:
Overall, there were an estimated 254.4 million registered passenger vehicles in the United States according to a 2007 DOT study
Seems reasonable to use this as the denominator. If anything, since not every registered car is regularly driven, it is likely a high estimate and therefore understates the percentage.
But calculating it, 150,000 divided by 254,400,000 equals a fire incidence rate of 0.059% for data set B. So the average passenger vehicle catches fire 3.73 times as often as a Tesla Model S. Now, caveats:
This is comparing an annual number (0.059%) for some other time period. Without thinking much about it, you might believe there is weight to Wired's assertion that this is "3 vehicle fires in 6 weeks". The Model S has been driving on the road for much longer than 6 weeks. In fact, retail deliveries started about a year and a half ago. While technically accurate, to frame the data as "3 fires in 6 weeks" is misleading due to cherry picking data.
Another datapoint from Tesla blog if you don't think they are making their numbers up is that they had a rate of 1 fire in 100 million miles driven at the time of the first fire caused by debris. Not adding to the total miles driven since then we are now at 1 fire in 33 million miles while the national average is 1 fire in 20 million miles.
Its possible this exists for all cars but I am not aware that the data has been given for Tesla but it almost assuredly exists with their cars reporting home all the time.
I think in order for a reasonable comparison to be made, it would have to be between the frequency of fires in Tesla vehicles vs the frequency of fires in vehicles less than a tear old.
In my experience (as a firefighter), most vehicle fires occur in older (>10 years) vehicles.
They are having more accidents than other electric cars (3 vs 0), and it certainly seems that having four times as much battery power, and lining the base with them, is the problem.
There's nice spin you could put on this, which is likely true: people who have Teslas drive them more, because they are fun, cool, excellent - and have 4 times the range and power.
Secondly, with more power, you will drive faster, hence greater likelihood of accident, and of greater severity.
But Tesla's gotta fix this. For PR reasons, and to be making a great car. There's always going to be unforeseen problems with new tech, the only issue is how you handle it.
It's a bit of a shame though: low slung batteries give low center of gravity and great handling and stability. Thicker armour adds unnecessary weight. Perhaps armour that deflects rather than stops? Perhaps a different failure mode for the batteries; built-in extinguishers? (i.e. the issue is fire, not car damage. Ideally, solve both).
> Perhaps a different failure mode for the batteries; built-in extinguishers?
I hesitate to say 'it's not possible', but:
-Until a new battery formula becomes viable, we're pretty much stuck with batteries that light up when abused. Hopefully new formulas come online soon, but who knows wether or not they'll actually be safer.
-Chemical fires are pretty tough to extinguish via conventional means.
A quick back of the envelope calculation shows that, at this rate, the Tesla Model S is roughly 25x more likely to suffer a fire after a collision than a non-Tesla.
While gas powered cars must, out of necessity, carry a tank of flammable liquid around, which is fairly unsafe, the gas tank is not in the front of the car. No other electric car manufacturer puts their batteries so far forward under the hood, either. The Tesla also has very low road clearance.
Such small numbers make terrible statistics. For instance, I could say that "the Tesla Model S is infinitely times less likely than a non-Tesla to suffer a fire which leads to a death." 0 for Tesla, X > 0 for others.
Face it, battery technology is not sufficiently advanced to deliver good range without some serious trade offs. The trade off of the Tesla may prove to be the wrong choice long term.
No one would ever suggest lining the bottom of a car with a gasoline tank, fortunately we don't have too because the power density is so high. Hence they can put tanks where a collision or road debris damage are least likely to damage them.
However Tesla needs a stupendous amount of battery to get its range. Hence we get large cars where they have chosen to isolate the entire pack at the floor of the car. They made the choice to dispense with a central tunnel which could hold more batteries, not use the front engine area or any trunk space.
There were to reasons for their design choice, first because of the sheer amount of batteries needed and second because of pack swapping. I don't think the trade off is worth it.
They can use some of the front "frunk" area to have a raised stack of batteries where the traditional firewall is and do similar in the rear. The could have a raised tunnel down the center line and reduce the need of having batteries so far forward. There were many choices available but they wanted swappable packs.
What statistic I want to see is, how many penetration events have their been versus how many fires. How many cells have to be penetrated to create a fire? Is there some flaw that another component is failing during these minor accidents that is causing the fires?
There are 6,000 collision related car fires in the US every year, out of 250M vehicles. There have been 3 collision related Tesla fires out of ten thousand vehicles in roughly six months of total drive time. That works out a 25x higher fire risk for the Teslas.
It's nice to see actual numbers and risk rates. I had dismissed the Tesla fires as insignificant. Your post has made me think I should take a second look.
77 comments
[ 0.19 ms ] story [ 133 ms ] threadThis kind of reminds me how the media reports every single airplane crash. How about report every car crash with fatalities. According to Wikipedia over 2,500 per month.
Gasoline gars don't catch on fire in non-fatal accident with any frequency
Now let's talk numbers.
Ford Fusion fleet, in the US: 1,520,000 vehicles. Tesla Model S fleet, world: 18,300. If 300 Ford Fusions caught on fire in a 6 week span I guarantee NHTSA would be investigating.
>Gasoline gars don't catch on fire in non-fatal accident to begin with.
That can't be true. I live in a relatively small town and I've seen 2 car fires(or the aftermath) in the past year, as a result of an accident.
There were 187,500 car fires in 2011 (http://www.nfpa.org/research/fire-statistics/the-us-fire-pro...), that's just over 3600 a week, which is about 0.07% of the total cars (~254M, http://en.wikipedia.org/wiki/Passenger_vehicles_in_the_Unite...).
3 of 18,300 Model S vehicles have caught fire. That's 0.02%.
As for running over debris, again it is not important because who cares if gasoline cars have different causes for their fires vs the causes of the model S. What matters is the percentage that actually do catch on fire with a large enough population.
After that, the next question that should be asked is, what is the difference between the danger of these fires. Are gasoline fires more or less dangerous.
https://www.google.com/search?q=car+fire&tbm=isch
http://www.ebaumsworld.com/video/watch/83524899/
There are pressure relief valves on the tank that prevent that from happening in the vast majority of cases.
They burn, sure. They almost never explode.
If your tank was punctured while driving it's quite likely the first you'd know was when you ran out.
However, last year the 2013 Ford Fusion was recalled (over 89k sold) after 13 fires caused by engine overheating: http://usnews.rankingsandreviews.com/cars-trucks/best-cars-b...
It's not clear whether the Tesla catches fire more often than other cars from road debris, 2 instances could be coincidence (I've never run over something that caused actual damage to my car after years of driving, crosses-finger).
However, that is the government's job, to examine the fire and determine whether they should have been prevented.
Straight from the article.
The volt had one, rather famously when it was tested by the NHTSA. In that test (which was pretty ludicrous if you dig into the details) as in the recent Tesla crashes, I don't think they're really reflective of the engineering quality of the car.
> The comparisons to the Leaf and Volt are meaningless if none of those vehicles have been in similar accidents. You can't just point to numbers and imply that a lack of reported fires means they're safer.
So...the question is...well, why should we assume that the problem is underreporting...that it's just Tesla's that are getting all the attention? That's not even true:
http://www.nytimes.com/2011/12/22/business/tsunami-reveals-d...
> DETROIT — Nissan inadvertently gained some valuable insight into the durability of its electric car, the Leaf, when about two dozen of them were destroyed in the tsunami that ravaged Japan in March. > > None of the cars caught fire, and their batteries remained fully intact, shielded by an airtight steel exoskeleton and two other layers of protection that surround the 660-pound packs.
“Considering how they were tossed around and crushed, we think that is a very good indication of the safety performance of that vehicle,” said Bob Yakushi, the director of product safety for Nissan North America.
Nissan’s decision to encase the Leaf’s battery in steel may help explain why federal safety regulators investigating postcrash fire risks in the Chevrolet Volt do not have the same concerns about the Leaf. General Motors packages the Volt’s battery cells on a T-shaped steel tray with a plastic cover.
We should expect more fires given enough time. I wonder for how many they will continue to report on every single one?
Nissan Leaf has sold 3x more vehicles than Tesla.
Are you saying that objects don't hit the underside of Leafs? (Perhaps they are repelled magically?) Or that objects are somehow more attracted to the underside of Teslas cars? Or the Leaf accidents don't count because the Leaf owners didn't file a report that nothing happened?
I'm not saying that the Tesla is necessarily better performing, just that two different types of people are the likely owners.
NB: Just pondering.
Since these are high-energy impacts, you'd have to be traveling at highway speed to have a damaging impact.
It might be more interesting to compare how a similar car style would behave (think high performance sedans).
I was surprised to see them because I don't consider where I live to be a big Tesla type area.
This isn't disastrous. Problems like this routinely emerge with new designs. And at least the flammable materials stay in one place (unlike a gasoline engine). But still, a fire's a fire, and I'm sure Tesla's engineers are working overtime trying to figure out a solution.
What could the solution possibly be? Puncture the battery and it catches fire, and the battery covers almost the entire bottom of the car. Practically speaking there's only so much steel you can put on the bottom of the car.
It seems that Model S spreading the battery across the bottom of the car may simply be a defective design.
The conditions that cause the fires seem to be narrow and well-defined enough that a clever workaround is possible. Maybe add an extra layer of sacrificial armour? Or move some cells away from the vulnerable region.
Where is this claim coming from? Wired? I can't find any source. To me it seems like some writer for Wired figured they could get some clicks by speculating.
The National Highway Traffic Safety Administration “absolutely has to investigate” the Nov. 6 Tennessee incident, Clarence Ditlow, executive director of the Center for Auto Safety, based in Washington, said in a phone interview.
I'm not a car expert but I imagine it's hard to protect all the pipes that are carrying the liquid. And that may be why these fires are starting... some puncture causes the liquid to release. Just a theory.
There's only so much you can do with a punctured lithium ion.
The presence or absence of that liquid won't affect the outcome of what's happening in these accidents, which involves driving steel into the battery packs, compromising the batteries and causing a short circuit.
150,000 car fires per year / 52 weeks is 2880 fires per week. There are 135900000[1] vehicles in the US, which gives 21.2 fires per million-vehicles week. With 19000 Teslas on the road and three fires in six weeks we have 0.5 (fires / week) / 0.019 (million vehicles) = 26.3 fires per million-vehicles week. So even if this is typical for the Model S (and it doesn't seem to be), the cars aren't much more likely to catch fire than ordinary cars. Also, I haven't accounted for the fact that the Model S's probably get driven more than your average passenger car, both due to low per-mile costs and the fact that they're more likely to be new and thus valued by their owners. Taking this into account would probably bring the Model S's fire-rate below gas cars again.
[1]: http://www.wolframalpha.com/input/?i=number+of+cars+in+us
http://www.autonews.com/article/20130806/RETAIL/130809922#
For example, I wonder what happens when you group the Tesla S in with cars that can do the quarter mile in under 13 seconds... http://www.wreckedexotics.com/
But they still are too expensive for me.
The second and really much larger reason this article is spin is that they are painting this as a major issue without giving enough information to quantify it. I try to quantify it in this comment and the results are the opposite of how Wired is spinning this issue. Here's what they said:
While there are an average of 150,000 car fires in the U.S. each year, with only 19,000 Model S sedans on the road, having three vehicles catch fire in six weeks has garnered plenty of attention.
So this is comparing two data sets:
A: % of Tesla Vehicles that catch fire. Dividing 3 by 19,000 leads to a fire incidence rate of 0.0158%.
B: % of all vehicles that catch fire. For this data set, Wired helpfully tells you the numerator of 150,000 vehicles catch fire in the U.S. but completely leaves out the denominator.
Without giving the complete information for data set B, it's difficult to determine if Tesla's fire incidence rate is out of the ordinary and warrants the speculation of, "Could Spark Federal Inquiry". So in an attempt to find the denominator, I discovered:
http://en.wikipedia.org/wiki/Passenger_vehicles_in_the_Unite...
Overall, there were an estimated 254.4 million registered passenger vehicles in the United States according to a 2007 DOT study
Seems reasonable to use this as the denominator. If anything, since not every registered car is regularly driven, it is likely a high estimate and therefore understates the percentage.
But calculating it, 150,000 divided by 254,400,000 equals a fire incidence rate of 0.059% for data set B. So the average passenger vehicle catches fire 3.73 times as often as a Tesla Model S. Now, caveats:
This is comparing an annual number (0.059%) for some other time period. Without thinking much about it, you might believe there is weight to Wired's assertion that this is "3 vehicle fires in 6 weeks". The Model S has been driving on the road for much longer than 6 weeks. In fact, retail deliveries started about a year and a half ago. While technically accurate, to frame the data as "3 fires in 6 weeks" is misleading due to cherry picking data.
http://www.teslamotors.com/blog/model-s-fire
In my experience (as a firefighter), most vehicle fires occur in older (>10 years) vehicles.
There's nice spin you could put on this, which is likely true: people who have Teslas drive them more, because they are fun, cool, excellent - and have 4 times the range and power. Secondly, with more power, you will drive faster, hence greater likelihood of accident, and of greater severity.
But Tesla's gotta fix this. For PR reasons, and to be making a great car. There's always going to be unforeseen problems with new tech, the only issue is how you handle it.
It's a bit of a shame though: low slung batteries give low center of gravity and great handling and stability. Thicker armour adds unnecessary weight. Perhaps armour that deflects rather than stops? Perhaps a different failure mode for the batteries; built-in extinguishers? (i.e. the issue is fire, not car damage. Ideally, solve both).
I hesitate to say 'it's not possible', but:
-Until a new battery formula becomes viable, we're pretty much stuck with batteries that light up when abused. Hopefully new formulas come online soon, but who knows wether or not they'll actually be safer.
-Chemical fires are pretty tough to extinguish via conventional means.
While gas powered cars must, out of necessity, carry a tank of flammable liquid around, which is fairly unsafe, the gas tank is not in the front of the car. No other electric car manufacturer puts their batteries so far forward under the hood, either. The Tesla also has very low road clearance.
No one would ever suggest lining the bottom of a car with a gasoline tank, fortunately we don't have too because the power density is so high. Hence they can put tanks where a collision or road debris damage are least likely to damage them.
However Tesla needs a stupendous amount of battery to get its range. Hence we get large cars where they have chosen to isolate the entire pack at the floor of the car. They made the choice to dispense with a central tunnel which could hold more batteries, not use the front engine area or any trunk space.
There were to reasons for their design choice, first because of the sheer amount of batteries needed and second because of pack swapping. I don't think the trade off is worth it.
They can use some of the front "frunk" area to have a raised stack of batteries where the traditional firewall is and do similar in the rear. The could have a raised tunnel down the center line and reduce the need of having batteries so far forward. There were many choices available but they wanted swappable packs.
What statistic I want to see is, how many penetration events have their been versus how many fires. How many cells have to be penetrated to create a fire? Is there some flaw that another component is failing during these minor accidents that is causing the fires?
- Firefighters will take longer to rescue injured from the passenger compartment
- The fire can possibly ignite spilled gas from other cars in a multi-car accident
- Insurance rates will be higher, as a collision is much more likely to total the car