Comment by lolc | Hacker News Reader

Comment by lolc | Hacker News Reader

I can't relate to per-year risk percentage numbers. But the inverse is quite illustrative I find. Here is the expected time between events for each category according to the paper:

General house fire: 360 years

ICE fire: 1100 years

EV fire: 4200 years

Battery storage fire: 20'000 years

Tumble dryer fire: 27'000 years

PV system fire: 71'000 years

This way I can understand that while I might see my house burn, the battery storage is very unlikely to pop in my lifetime. Because these numbers are not very accurate anyway, I've rounded to two digits. The authors note that the number for battery storage in particular are very unreliable.

On the other hand: just looking at probabilities of events does not tell you how lethal or destructive the event will be. And lithium fires are very bad in both categories. Comparing to tumble dryers is not very useful when you factor in the properties of the fire. Quite possibly the danger to life is higher with an EV compared to an ICE, even if the ICE is assumed to catch fire at four times the rate of the EV. In addition, the wear-out failure rate of EV is to be discovered yet, because the fleet is so young. The EV rate may still climb. Or maybe battery manufacturers improve and the danger will be reduced? Some things do get better after all!

I find this data quite amusing. In another life I cleaned up after things like fire damages. The vast majority of them were kitchen fires and the only one I can specifically recall that was not cooking-related was some dude's hot rod that blew up and set the garage on fire.

> The vast majority of them were kitchen fires

Without any personal or professional reason, I've seen many of them as a casual onlooker. I remain terrified of open pot deep frying over a stove.

But the insurance assessors I've talked to recently do mention a new trend in battery fires - though usually caused by a combination of bad quality, excessive quantity stored and sloppy practices bordering on the abusive.

So in a housing development with 36 homes, we expect one to burn in 10 years? Am I reading that right?

If so, that doesn't sound right to me.

No, all 36 burn in year 360.

That sounds pretty close to right. House fires don't necessarily burn down the whole house with flames to the sky. My home had 2 pretty serious fires growing up (over 20 years): 1/ car fire from grease/oil/gas leaking from our diesel car engine after driving in very heavy rain, 2/ unattended hair dryer overheated (don't know if it was on or off but it was plugged in).

> So in a housing development with 36 homes, we expect one to burn in 10 years? Am I reading that right?

Yep. And - anecdata - it is about correct... if you count all the unintended fires that didn't escalate to a full blown dwelling fire. The classic example is oil overheating and flash-combusting in a pan or pot, and people making that exponentially worse by trying to extinguish it with water instead of just putting a damn lid on it.

Another infamous example used to be (old) people smoking in their wing chair while watching TV and falling asleep with the lit cigarette then setting the furniture or carpets alight. That one has been very effectively remediated by strict fire resistance requirements on furniture as well as requiring cigarette manufacturers switch to self-extinguishing paper.

Seeing a fire truck responding to my direct neighbourhood every 10 years sounds about right.

The PV stats seem low to me since Amazon for example started removing solar panels from their warehouse roofs due tu multiple fires in multiple different places.

"The rate of dangerous incidents is unacceptable, and above industry averages"

Amazon took their rooftop solar offline temporarily to inspect and safety work after a string of fires in in 2020. They replaced some of the more shoddy installations, but Amazon warehouse rooftop solar is back online for years since.

Good to know, thanks!

I wonder what the root causes of these bad installations were.

There was a lawsuit a few years ago by Walmart about panels installed by SolarCity in the early 10s. During this time the company was facing financial issues (their model was leasing panels), so did a lot of cost cutting.

The main causes of PV fires are arcing due to poor installation, e.g. loose connections or damaged components. PV is not like household mains power where you have high current for a few minutes (e.g. turning on a kettle), it's like that for hours at a time. You need to be much more careful.

On a residential PV install, each 'string' from the panels to inverter could easily be 10 or more amps at peak output.

The years perspective is useful. But I'm not sure it's useful to compare the danger from different sources of house fires. The bad ones happen at night, and then have more than enough time to cause catastrophic damage whether it starts from a dryer or a candle or whatever. (There are some instructive videos on YouTube of how quickly fires spread in rooms)

After a few years of arguing with people about nuclear safety I'm not sure that this sort of perspective is useful. People make a gut-feel check on whether they feel safe and then just rationalise it. The actual level of risk doesn't seem to factor in to the conversation.

If we're going by the numbers this should probably disqualify the use of home battery storage on safety grounds. We're talking something like 1 home/year catching fire in a large town, those are some crazy levels of damage. If it was 1 home/year being hit by no-harm-detected-but-higher-than-expected radiation people would be seizing the numbers as evidence that a reactor needs to be shut down.

The thing with nuclear reactors is that they don't go off that easily or that often, but when they go off, the impact is both very large (Chernobyl spread to Bavaria!) and very long-lasting in its impact (Chernobyl is still a no-go zone, and to this day you have to inspect and throw away game and shrooms collected in Bavaria if it is impacted by radiation).

Coal power plants have a much larger impact, new ones have been cleaned up, but in the 1960s they were putting lots of radioactive waste into the air which rained down around the world and is still there.

Nothing comes for free. If you have 100 bad things each with an odds of happening once every 100 years, its not going to be a pretty century.

Okay, but if your house is on fire the chance of your battery storage catching on fire goes way up.

Sure when the fire expands into other rooms the battery storage may go off too. By that time I hope I'm upwind of the whole debacle. Many "general" fires do not go beyond one room though.

Reply on news.ycombinator.com