Hacker News Reader

Comment by A_Person | Hacker News Reader

It's always surprised me how many people do not know that water is incompressible! For example, when someone is filling a scuba tank, it's common to put the tank in a large, often metal-sided container of water. When asked why, the shop staff member will typically say: "In the unlikely event of the tank blowing up, the water will absorb the force of the explosion." But in fact, water being incompressible, it will absorb approximately 0% of that force - which is then transmitted undiminished to the metal walls - which then explode like a giant hand grenade, promptly maiming or killing everyone standing nearby! The purpose of the water is instead, to absorb the heat from adiabatic compression, and keep the tank cool.

View on HN

That's just wrong. The the blast from the tank would have to displace water before of the energy of the blast can exit the side or top of the container. That spreads out the energy from the blast and the container would redirect much of that energy upwards both of which reduces the effective radius of the blast.

The shopkeeper is still somewhat wrong though: it's redirecting the energy, not absorbing it.

Regardless, Mark Rober and The Backyard Scientist have a pretty great video demonstrating exactly what is going on here: https://youtu.be/W4DnuQOtA8E

I'm no scientist and you may well be right. But I'm skeptical of your statement that "much" of the blast would be directed upwards. If you look at the pictures of tank explosions, they often show the entire room flattened right to floor level - not a big hole in the roof and minor damage everywhere else.

> It's always surprised me how many people do not know that water is incompressible!

Interestingly enough, they're right (though they may not know it) since the article is about its ability to compress under the right circumstances!

But only by 3% - so I still won't set up my chillout lounge right next to the tank fill station! :-)

It's definitely both the cooling (otherwise the cylinder would get hot, just like a can of duster gets cold), and the safety thing. The walls of the container would generally be more than enough to withstand the acoustic shock and send that energy upward as a burst of water, but your real enemy is shrapnel from the cylinder, which the water will promptly slow to non-lethal speeds. I'll agree that a guy doing it in a plastic tank or a smallish metal one would be asking for trouble.

On the flipside, water is used for the periodic pressure testing of scuba tanks, precisely because it is essentially incompressible. So if a tank does blow up during testing, you have a very small expansion and thus very small energy release.

If you draw a curve of gauge pressure as a function of volume, the energy release is the area under the curve, and as the curve steepness goes to infinity (i.e. compressibility goes to zero) the energy released starting at a given pressure goes to zero.

I did know that, but I'm not sure how that applies. In your scenario, the tank ruptures, there's virtually no water expansion, so nothing happens. In my scenario, the tank ruptures, the internal gas content instantly expands by (say) 250 times, creating a massive force which I assume is transmitted virtually undiminished, through the water, to the sides of the container, which promptly explode outwards. I do accept that some of the force will go upwards, but I believe the container will still explode. Not really trying to disagree, just trying to get my head around it :-)

I'd have to do the math (and know the material properties and thickness of the tank walls) before saying one way or the other, whether that scheme is effective.

What it could also be doing, is keep the tank from going off like a missile through the building (added inertia of the water plus it takes much longer for the tank to tip sideways).

> "...the breakthrough came when we realized that we should not assume water is incompressible..."

The shop staff was right, water is not incompressible (under important conditions that don't apply to their scenario!).

Reply on news.ycombinator.com