An EV Battery Engineer From A Major Carmaker Explains Why Flooded EVs Catch On Fire

Floodedfireevexplainer Top
ADVERTISEMENT
Earlier this week I wrote about how Florida firefighters were having to deal with something new and troubling: electric vehicles, flooded with seawater from Hurricane Ian, have been bursting into flames, and the resulting fires have been extremely difficult to extinguish, leading to unusual methods like shoving cars into water-filled ditches, using about ten times the amount of water to put them out compared to a normal fire, and then parking the charred carcasses very far apart because the damn things will re-ignite, like demons. In that story, I speculated a bit why this is happening, but to really understand would take an actual engineer. Thankfully, one has reached out.
This engineer works at a major OEM, working with batteries and EVs. They requested anonymity, because OEMs can get weird about things their engineers reveal to the public, so we’re honoring that request. We think this knowledge is important enough for the general public that we want it out there, so here we go.
Take it away, anonymous engineer!
Rednsudivider

Why Salt Water Flooding Makes EV Batteries Burn

Part 1: Salt And Shorting

As the Fire Marshal and other commenters pointed out, the problem begins with salt bridges short-circuiting bus-bars and other high-current components in the pack – essentially exceeding the creepage and clearance distances because not only is the salt water itself more conductive than air or air-filled potting, but the salt residue left behind can be much more conductive, forming a path for current to flow.
This part can happen in anything with a battery or voltage potential – it’s why ICE cars that have been flooded and resold often have “quirks,” as some components in various ECUs have shorted and subsequently burned up or changed their behavior because the circuits around them have changed (via the residual salt shorting circuit traces). Its just that your typical lead-acid battery at 12V doesn’t have enough voltage to drive sufficient current through a salt bridge short to heat up enough to explode. Drop a wrench across the terminals and sure, it’ll blow up, but the salt isn’t conductive enough for a typical 12V battery to flow enough current to heat up enough/off-gas enough to explode.
The salt bridges are specific to the issue with salt-water flooding. Other sources of electrical shorts resulting in thermal runaway not specific to flooding could be manufacturing errors/damage in the cells (the primary cause of the Chevy Bolt fires, and probably the other LG cell fires as well), deposition dendrites that grow from the cathode or anode (depending on the cell temperature and voltage history) to pierce the separator, physical abrasion of separators or other insulating materials from the cells swelling and shrinking with each heat and charging cycle (secondary Bolt fire cause), or, of course, physical damage.

Part 2: Hard And Soft Shorts

BEV propulsion batteries on the other hand have ~300V or more (Hyundai’s newest EVs are at 800V), which is a lot more electrical “pressure” (to use a hydraulic analogy) to drive current through an electrical short, resulting in much more current (flow) and thus more heat generation. This heat will be generated in the cells themselves, and in all the electrical bussing that is part of the circuit, as well as in the salt bridge.
In the case of a “soft,” relatively high resistance short, this may take a while before the temperature gets critical, but since the battery has a lot of capacity and is pretty well sealed off with little heat transfer to the environment with the car off (as in no active cooling systems operating), the heat has nowhere to go, and things still heat up.
A “hard”, low resistance short makes this much worse since most Li+ chemistries can output much more current than is safe for the cell (and the restrictions to prevent this are in the battery control module, motor controllers, inverters etc, all of which are off), heating the cell up.
If the pack is filled with water, it is harder to build up enough heat since water is very good at absorbing heat, warming up and then turning to steam before venting from the pack, but eventually it boils off (possibly leaving behind more salt bridges). Either way, the cells get too hot through their own heating and being heated by the electrical bussing carrying the current until…

Part 3: Something Gives

Typically in Li+ cells, this something is that the graphite* in the anode (even in so-called “silicon anode” cells, graphite still makes up 80-90% of the volume) breaks down, then separator sheets inside the cell deform or melt, causing a small short between a single cathode and anode layer (monocell) within a cell. This short begins generating heat, melting more of the separator and breaking down the electrolyte, making the short worse and the cell internal heating worse.
*It’s not the graphite anode itself but the binders and the electrolyte interface over the graphite that begins decomposing first. The result is the same – you get a collapse of the anode, and this paves the way for a short when the separator begins melting, but technically the graphite decomposing isn’t really accurate. This process also off-gasses, adding to the flammable mixture waiting for a good O2 source to combust.

Part 4: Chemistry

Now we get into the cell chemistry effects a bit more. For example, LFP (Lithium Iron Phosphate, LiFePO4 etc) cells are considered safer because the chemistry contains less energy (thus less energy can go into either current to further heat the cell or be chemically converted into heat once the cell is on fire, heating other cells), and the LFP cathode chemistry seems to retain bonds to oxygen for longer at higher temperatures. Therefore, the LFP chemistry resists thermal runaway until higher temperatures, and has more trouble achieving those temperatures due to its lower energy density, so more mass to heat up with less energy than other chemistries.
In contrast, NMC (Nickel-Manganese-Cobalt oxide cathode) or NCA (Nickel-Cobalt-Aluminum oxide cathode) are more energy dense, which means more range for a given weight or volume, but more severe thermal runaway behavior. Variants of these two chemistries are also more commonly used at least in the Americas & Europe because they allow for more range, more power etc – China has a lot of LFP production and use in vehicles including their own LFP prismatic-can version of the Tesla Model 3. Anyway, any of these chemistries can break down at high enough temperatures when the anode, separator, cathode, and electrolyte all degrade, these things just happen at different relative times and temperatures with different chemistries.

Part 5: Why The Fires Can Re-Ignite

The electrolyte, separator, and cathode at least all typically contain some oxygen. This means that they can ignite, even in the absence of air (made worse by any Li metal that forms within the cell, which will react with almost anything). When burning, the lack of enough oxygen for a stoichiometric reaction means that some of the gasses released can burn more if they find a new source of O2.
This causes the hottest cell to rupture (if a pouch) or vent (if a cylindrical or prismatic can), venting hot combustion products or partial products, still-burning particulates, and bits of molten aluminum and copper from the cathode and anode current collector into any air gap within the pack (hopefully there is a manifold designed to carry this stuff away from other cells & electrical connections, but not always).
All this hot, flammable debris consumes all the O2 in the pack, but it’s still not enough for stoichiometry to be happy, thus you see fires outside of the pack fueled by H2, CO, and hydrocarbon chains generated by the thermal runaway, and constantly re-ignited by the molten bits of aluminum, copper, etc. that form the current collectors and other pack components.
Douse the whole pack in water and you can cool things off enough to stop the fire, but the internal short from the salt bridge may still exist (and has probably been joined by other shorts from the molten metal and other debris from the cell or cells that ran away already), which then starts the process over, reigniting the whole mess.
All of this is basically a long-winded way of saying that salt water in high voltage packs spells a bad time, but current EV design has a long way to go to prevent thermal runaways from propagating through the pack, and the high energy density of Li+ cells (even LFP) makes this considerably harder than with NiMH or lead-acid batteries.

About the Author

Mercedes Streeter

View All My Posts

52 thoughts on “An EV Battery Engineer From A Major Carmaker Explains Why Flooded EVs Catch On Fire”

  1. Wonder if we might eventually see a mandated requirement for EVs to be equipped with a standardised mechanism to discharge the battery pack(s) after a crash, fire or flood damage.
    A lot of industrial equipment with significant electrical battery storage (eg. Some Active harmonic filters, Variable speed drives, etc) will discharge to the earth connection when they lose power, to limit risk of electrocution or other exciting electrical quirks that come from dumping a bunch of electrical energy into a dead system with no control.
    Obviously don’t have a suitable permanent earth connection on a ground. But perhaps something emergency services could be equipped with as EV fleet grows – a load cell or portable electrical grounding system they could connect to my mythical and entirely unproven emergency discharge point to dump majority of the stored energy out, reducing risk of electricity and fire.

    1. Trouble is, 100kWh is a LOT of energy to dump into anything in a short time, a load cell that can comfortably cope with 100kW is a small towed trailer full of heating elements and a big cooling fan, and that is going to take an hour to dissipate that 100kWh.

  4. “BEV propulsion batteries on the other hand have ~300V or more (Hyundai’s newest EVs are at 800V)…”

    Imagine being the poor sap tasked with breaking these cars apart for recycling.

  5. Being an insurance adjuster, I have total lossed cars that were flooded, due to learning my lesson where I had one that I tried to fix (under about 65% of the value of the vehicle) and then going back over that car for months due to one electrical gremlin after the other. I ended approving for more than the car’s value to fix, since once you start fixing an insured’s car you can’t stop part of the way. Very expensive lesson learned.

    You also get the raw sewage in the car, which has a unique smell all of its own. Once you have smelled sewage you won’t forget it. It always smell the same, so I could walk into a house and know it had a sewage overflow the minute I walked into the basement.

  7. Has anyone ever tested applying the foam used for airplane crashes for EV crashes as it smothers flames? How about storage with a system similar to restaurants with grease fires? Or how about spray foam insulation which blocks out any air , isn’t flammable, and can be peeled off when needed?

    2. It is. Remember, I’m LV and HV qualified.

      Both LV and HV systems do use watertight connectors and sealing at potential entry points, or place intrusions within weathertight areas.
      As an example, the Chevy Bolt EV (both generations; both the T-pack and the floor pack) locate the HV connections outside of the weather-tight area at the very front of the vehicle and very low to the ground. Well below the underhood electronics in fact. But the connector is double-gasketed. There’s an outer housing for the pass-through which has a watertight gasket, and then the HV cable connections themselves have a gasket as part of the connector.

      But these assemblies are rated much like your ‘water resistant’ watch; they can only keep water out for X continuous hours at Y, Z, and A variables. You can soak the hell out of the underbody on a Chevy Bolt. You can drive it through several inches of water. You can drive it in a monsoon. Snow and ice from salted roads can get stuck all around it. And in those conditions, the seal will hold.
      In fact, nearly every sensitive or high current connection on your car which may be exposed to water regularly has similar waterproofing. Even those that don’t necessarily look like it. Those little ‘accordions’ inside your O2 connector? Those actually form a multi-layer water-tight seal. The electrical connectors on modern cars generally have excellent weather resistance or water-tightness.
      But when you start exposing it to high salinity water for hours or days – especially when that water is full of other nasty chemicals that break down those rubber seals – it’s going to fail. That’s well outside the design envelope.

      And as everyone will tell you: the safest thing to do if you think there will be flooding is to completely remove the car from the area. Cars are not designed to sit in flood waters, whether they’re ICE, PHEV, or BEV. That’s miles outside what reasonable for anything. If they were, flood totaled cars wouldn’t be a thing. Firefighters would take the big trucks on flood rescues, not low-draft boats.

    3. Water has a pesky habit of getting wherever you don’t want it, so trying to completely seal all electrical components (even just the high power circuits (since not actually HV)) in a vehicle that still needs to have certain flexible/moving points and actually be used and maintained would be quite an issue.

  9. So the engineer left out the very important fact that salt water is actually a very small part of the problem and also not fully accurate.

    It’s flood water. Which is inherently contaminated. It’s not pure salt water. It’s a disgusting, corrosive, toxic mix of freshwater, salt water, whatever’s on the ground, biological agents, chemicals and toxins galore, and nothing fun at all.
    This toxic mix from hell is why flood cars with interior flood water are instant total no matter how ‘light’ the damage. Insurers do not want to cover claims for flesh-eating bacteria.

    That toxic mélange destroys any sort of seal in short order. Take your pick as to mechanism – your guess is as good as mine. Nobody knows what exactly is in that ‘water.’ We don’t want to. All we need to know is that between pressure, corrosion, and contaminants, your seals are done. Seriously. Don’t go playing in flood water.
    We also know this water will corrode contacts and metals. We don’t know which, what extent, any of that. Again: it’s mystery liquid. There’s ammonia and ammoniums, propylene and ethylene glycols, new and used petroleum distillates, salts, sulfurs, bromides, and all manner of conduction-enhancing corrosive brews.

    Then there’s the matter of seal and gasket designs. Seals and gaskets are designed to keep a specific type of fluid inside a specific type of metal or plastic and shrug off a monsoon. They are NOT designed and they are NOT capable of handling long-term submersion. That would require a TOTALLY different design of everything; different metals, different plastics, different gasket materials, the works. If you soak a sealed rear differential in pure isopropyl alcohol for 24+ hours, then you check what’s inside, you’re going to find contamination from the isopropyl alcohol.
    Going through large puddles and even fording rivers? A-okay. Submerged uninterrupted in a liquid for days? Doesn’t matter what the liquid is, that seal is done.

    “Oh well just fix the seals! Easy!” We just covered that. It’s not easy. You’re now talking about a battery housing made from high nickel alloys. Stuff like UNS N06601 AKA Alloy 601 AKA Inconel. Easily over triple the price. Twice the thickness and switching to highly insensitive butyls which also cost 2-3x as much.

    And ultimately the fundamental problem with BEVs is inherent to the design. Everyone puts the batteries below the floor. This means they are the lowest point on the car and the most susceptible to submersion in even passable water. That’s right folks – even standing water which a Tesla Model 3 can (not entirely) safely drive through is more than enough to submerge the battery and seals.
    How much water exactly? Let me put it this way: it doesn’t even have to reach the bottom of the doors on a BEV. We’re quite literally talking an amount of water that wouldn’t bother a Jeep Wrangler, and even a Toyota Prius would likely survive.
    Why? Wrangler is obvious – the water didn’t reach anything but wheels and tires. But on the Prius? The secret is simple: the HV system is mounted far higher. Flood waters only touched the floor pan, oil pan, and maybe transmission and axles. The actual HV battery is mounted in the back seat area, above the top of the spare tire well, and inside the weathertight interior.

    But again: it’s an inherent design flaw that cannot be overcome by anything. Because it’s about weight. The battery is by far the heaviest component of a BEV, and can even weigh more than the unibody. If you want your 4 door sedan with flat floors, and you don’t want it to weigh 10,000lbs thanks to reinforcements, and you also don’t want it to handle like a plate balanced atop a twig?
    Structural battery floor. That’s the only way it works – regardless of ride height! Ride height is just setting the body higher on the axles with a higher COG. The Hummer EV still uses a structural battery floor which is the lowest part of the car. Get the water over the axle hubs, and that Ultium pack’s now an Ultimate Inferno pack too.

    1. Your info about flood water is spot-on. As a storm water professional, I’ve seen people in flood situations wading through it like it was a day at the beach. These people are crazy.
      The thing is, if the ground is flooded, that means the wastewater conveyance system (sanitary sewers) is flooded too. Whatever material people flush is now in direct contact with a person wading around in shorts and flip flops. Same goes for septic systems, industrial chemicals & wastes, you name it- anything that is on the ground or underground is now suspended in that flood water. That includes bird poop, which the surface of the earth is covered with.
      Even breathing it is bad, let alone wading in it.

      1. It always astonishes the hell out of me to see anybody just playing in that shit. I’ve seen people jet-skiing in it on TV and just… “WTF?”

        I’m not a storm water professional by any stretch. I’m IT and automotive. I live in an area that basically never floods either. But once, we had a storm sewer collapse during torrential rains. This led to localized flooding. And it’s a fairly modern suburban area.
        The area’s old enough, people absolutely dumped used motor oil, transmission fluid, ethylene glycol, pesticides, and paint right into the brickwork storm sewers. For years. (And well past the point where they knew better. I saw people doing it into the 90’s.)

        And the fucking stench. Oh. My. Gods. We’re talking maybe a whole four or five inches of water over a thousand square yards at most. And it smelled worse than the contaminated fluids container at the shop. The one we dumped literally anything that wasn’t clean used motor oil in.

  10. I wonder how Tesla’s new design would hold up with the 4680 batteries. They are basically encased in a very rigid foam, but I do not know how watertight it is but it covers the bus bars and all electrical connectors on it as well.

    1. It won’t. Full stop.

      One, you’d be a fool to trust Tesla used a non-flammable foam. They’ve outright lied about their battery cells and been caught lying about them, and carried on lying.
      Second, foam does not magically solve for the fact that you have an enclosed space filled with liquids and fatally corroded batteries. If the foam is absorbent, it will actually make things WORSE. Because now you have fatal ruptures AND new shorts AND more corrosion.

      1. and for anyone with a Dodge with a foam filled rear quarter panel you know the foam traps condensation so that body panel rusting from the inside out would just as easily be the anodes and cathodes on the batteries under the foam.

  11. The part of this that was most interesting to me was the bit at the end about why they reignite. It’s one of those things that seems obvious once it’s been pointed out: you can put the fire out, but the conditions that caused it are still present. You still have a big mass of fucked-up, shorted-out Lithium cells, and some of them still have enough energy in them to start a fire. So you can smother the fire and cool things down to the point where the burning stops, but once things dry out there’s nothing stopping another cell from kicking the whole process off all over again. Kinda neat, in a way.

  12. I’m still not sure that this is a real problem or just hyperbole from antidotal accounts. The statistics are pretty clear that the risk of a vehicle fire is much greater for hybrid and liquid fuel vehicles than for pure electric vehicles. https://www.autoinsuranceez.com/gas-vs-electric-car-fires/ Yes, EV fires are harder to put out but the number of fires per 100,000 cars is 60 times less than a gasoline vehicle. How many gasoline powered cars burned due to shorted batteries after the hurricane?

    2. I think that in this case, it’s less that EVs are more fire-prone overall and more that EVs present a new class of hazard in the aftermath of natural disasters. ICE cars can and do catch fire, but not typically right after they’ve been flooded and are just sitting there amongst the wreckage of a hurricane-ravaged neighborhood. Also, the fires they cause are really quite nasty—hard to extinguish, and prone to randomly reigniting. This doesn’t seem to be something that had been considered much before, but it’s going to be more and more of a problem as both EVs and flooding continue to become more common.

      1. Nope. They are far more prone to fires and fires are more destructive. Both are true.

        What you see in movies is bullshit. Getting a typical gasoline car to actually catch fire is VERY hard. Set aside flooding. To get a gasoline car to catch fire you need several things to happen. One, you need to have a significant fuel leak. Two, you need to have a sufficiently hot ignition source or significant amount of vapor in an enclosed space. Three, you need to combine these two things.
        This is actually a LOT harder than people think.
        Underhood fires occur because you have fuel, enclosed space where vapors collect, and extremely high temperatures. That’s the issue.
        If you take a car that just got it’s entire rear half sheared off in a collision and has poured the entire contents of it’s gas tank over the road? Fire’s responding “just in case” more than anything. Liquid gasoline doesn’t want to burn easily. It’s the vapors that are on a hair trigger.

        BEVs (note I’m talking BEVs alone, NOT PHEVs) require only one thing happen to become an impossible to extinguish fire: damage a battery cell.
        That’s it. If you damage a battery cell in the BEV’s motor power pack, you have fire risk. You don’t need any other thing to occur. You don’t need to puncture the steel housing, you don’t need to apply voltage, you don’t need to add water. All you need is one damaged cell out of hundreds or potentially thousands, and you have a runaway exothermal reaction. Otherwise known as: BIG GODDAMN FIRE.

        The Tesla cultists love to throw around the false claim that numbers prove ICEs catch on fire more frequently. They don’t. They’re lying with statistics. As a raw number, more ICEs catch on fire, because there are more ICEs on the road. Duh.
        It’s exactly as absurdist as claiming that because you have 50 hens and 5 roosters, all chickens lay eggs.
        The question they refuse to answer is: what is the number of fires per miles driven as a share of BEV and as a share of ICE. That is, for each of each type on the road, how many fires occur as a percentage of the population?

        I’ll give you a hint: EVs in 2018 (the data I gave above) had a total of 208,000 new registrations out of a total of 279,100,000 total registered cars in the US.

          1. Try a Plymouth Valiant where the owner repaired a leak in the fuel line with gasket sealant that was susceptible to degradation when exposed to gasoline. Has anyone ever seen an engine block glowing orange right after driving it? I have.

    3. EV fires aren’t news because they’re more common than other kinds of fires. They’re news because they’re different than other kinds of fires in ways that people don’t expect. We’ve got a pretty well ingrained understanding of how fire works. And when fire stops working like we expect, it’s _terrifying_!

      It will be interesting to see how much of the difference in the prevalence of fires in EVs vs gasoline vehicles changes over time, and whether the difference is largely down to the fact that the fleet of EVs is significantly newer on average.

      1. Yeah, this. “Don’t let your EV get flooded because it’ll catch on fire” is the kind of warning that most people are going to have a very puzzled reaction to.

        1. You’re right but EVs not caught in floods are also bursting into flames in huge numbers without being close to a hurricane. Or ramming other vehicles while operating on autopilot.

    4. Don’t forget to control your car fires for age of vehicle.

      EVs are still new and expensive enough to be far better maintained than the average ICE that catches fire.

    5. Did you read the article?
      Its just that your typical lead-acid battery at 12V doesn’t have enough voltage to drive sufficient current through a salt bridge short to heat up enough to explode. Drop a wrench across the terminals and sure, it’ll blow up, but the salt isn’t conductive enough for a typical 12V battery to flow enough current to heat up enough/off-gas enough to explode.

      1. I did read the article. It is HARDER (not impossible) to make lower voltage batteries catch on fire due to these types of shorts. Per the article’s last paragraph: “makes this considerably harder than with NiMH or lead-acid batteries”

      3. Yeah we seem to be getting alot of EV zealots from the old site since everyone else stopped reading their rants. I expect the political crap posts to start soon.

    6. That is terrible data because it doesn’t seem to account for the fact that there are hardly any EVs older than like 5 years. New cars do not catch on fire very often.

    7. Yeah not sure where you got your math skills from but they called and want your diploma back. You are confused on likelihood of an ICE fire vs probability of an ICE fire in comparison to an EV. Sure there are twice as many ICE fires but that is from the fact that there are 1900 times more ICE vehicles than EVs.

Leave a Reply