[Beowulf] Swimming in oil..

Lux, Jim (337K) James.P.Lux at jpl.nasa.gov
Mon Feb 11 10:16:51 PST 2019

Stu reports swimming, but perhaps he was really more wading.

A significant problem with large vats of liquid, whether used for cooling electronic equipment, or just storage, is that if the density is significantly less than that of water, you don’t float.  Humans are just slightly positively buoyant in water (with full lungs).  Change that to oil or corn syrup or scotch whisky with a density of 0.9, and it’s like having 5-10 kg of weight on you, and that takes a lot of work to stay on the surface.
This is a well known hazard in the petroleum processing industry (aside from the fact that the air above the tank’s liquid surface is probably full of all manner of unhealthy things and not oxygen) – you fall in the big tank, you die.

Diala AX (a HV insulating oil I’ve used) has a specific gravity of 0.885, and is somewhat more viscous than water (not a lot) – if you fell into it, and couldn’t support yourself by standing on the bottom or equipment within the tank, you’d need to be rescued pretty quickly.  The increased viscosity would also mean that it’s more work to keep “treading oil” to stay above the surface.

USP white mineral oil is about 0.85 g/cc.  We had a thousand gallon tank of this where I used to work, and there was a whole discussion about safety – it was a wide flat tank, so in theory, if you fell in, you could stand up (except that the tank was polyethylene and it *is* oil.. there were questions about whether you could stand up on the slippery surface)

Of course, you can get oil in all densities – the stuff they use for road surfacing is quite dense.

Fluorinert FC-40 (which I’ve also used) is quite dense – 1.85 g/cc – you’d float well above the surface like a cork.  A quick glance at the line card for Novec® shows they’re all pretty dense - 1.4g/cc is the least dense.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.beowulf.org/pipermail/beowulf/attachments/20190211/e8118346/attachment.html>

More information about the Beowulf mailing list