Cryogenic Vessel Design Tips 1

[elbrownos]

Hi All
I'm designing an LNG vessel as an onboard fuel supply for natural gas powered trucks.
First prototype is performing ok but there is plenty of room for improvement. Mainly excessive heat inleak throught the feedthroughs and inner vessel supports.
Incidentally it's insulated with 3M glass microspheres + 10 micron vacuum. This seems to be performing well but MLI seems to be the way to go.
Any tips or pointers toward books or papers on the subject would be appreciated.

Cheers
elbrownos

 

[iainuts]

Hi elbrownos. Most of what you are asking is typically proprietary so you won't find a lot of detail on how to minimize heat leak in books. That said, Randal Barron's book, "Cryogenic Systems" is something of an industry bible.

http://www.amazon.com/Cryogenic-Systems-Monographs-Cryogenics-Randall/dp/0195035674

Bottom line, the design of inner vessel supports can be done using basic principals, such as by calculating the thermal conductivity through supports assuming the inner surface is at low temperature, the outer surface is at ambient temperature, and there is a thermal gradient between with no other heat transfer (such as convective heat transfer) off the sides of the support. Note that the thermal conductivity of most materials varies considerably depending on temperature, so I would suggest getting real properties data for thermal conductivity for the supports you use from NIST.

http://www.cryogenics.nist.gov/MPropsMAY/material properties.htm

More later if you're interested......

 

[elbrownos]

Thanks iainuts
That's the problem I'm finding - nobody wants to give away their secrets.
The book looks good, I've just ordered it.
The tank is around 27" OD so there isn't enough room in the vacuum space to do anything complicated.
Mind if I ask what your experience in this is?

 

[iainuts]

My experience is limited to what I've seen our suppliers provide. As such, I won't provide specifics of what they do, just guidance around basic principals to apply.

 

[Compositepro]

Glass microspheres in vacuum insulation seems a very odd choice. The spheres are filled with air which will slowly leak into your vacuum. Fumed silica would be better choice, I think.

 

[iainuts]

Nice catch Compositepro, I didn't know what those things were. (star for you) So microspheres have AIR in them??? Not only will air slowly leak out, but with any kind of thermal contraction/expansion of the inner, the balls are going to be ruptured. Very bad choice, for sure.

If MLI isn't used, perlite is the next best material. The only issue with Perlite is that as the vessel shrinks and expands (which it generally does to some degree, not to mention some motion caused by this being on a truck) the perlite will crush and end up as a powder at the bottom of the annular space leaving a hole at the top. This also effectively becomes 'concrete' which will damage piping in the annular space.

For a tank like this, MLI is the only reasonable material to use.

 

[elbrownos]

Microspheres are a ligitimate vacuum insulation with many advantages over MLI and only slighty worse performance.
My problem with MLI is that all information about its application is proprietary.

http://www.techapps.com/microsphere-insulation-programs.html

 

[iainuts]

Thanks elbrownos. In regards to MLI, Barron's handbook should help you out there. But I still wouldn't use microspheres for the reasons I mentioned before, they'll get crushed, drop to the bottom as a powder just like perlite, and leave you with holes at the top and a relatively incompressible powder at the bottom that will cause undue stress on piping.

 

[elbrownos]

They have an isostatic crush pressure of around 300 psi. I've got a paper on tests with microspheres in vacuum insulated flexible hoses, at extreme bend radii there was some loss of vacuum due to crushing but for vessels they are more than strong enough.
For me they are great because I can concentrate on the vessel design without worrying about how to wrap MLI. Down the road I will squeeze that extra bit of performance by exploring MLI.
I have a vacuum gauge on the prototype vessel so if there was a degradation of vacuum I would know about it.

 

[Compositepro]

That is an interesting application for microspheres. The literature does not answer obvious question of whether the spheres are very porous so the insides can be evacuated quickly, or if they are so impervious that they do not leak out slowly. I'm curious.

 

[elbrownos]

The spheres are impervious, they contain low pressure O2 & SO2

 

[iainuts]

I had a look at the web site you point to. I see they claim to have addressed the compaction issue of perlite.

Currently, cryogenic tanks using perlite can suffer from compaction caused by thermal cycling and/or vibrations, resulting in costly thermal inefficiencies. Microspheres have a fluid-like behavior that fills voids and does not compact while in use, thus reducing boil off and eliminating the need to re-insulate.

They also reference some testing done and a patent. I'd be interested in seeing what testing they've done. I don't see anything on their web site.

After giving it some thought, I could see how spheres might provide a bit more 'flow' inside the annular space due to their spherical nature when compared to perlite which is relatively jagged and tends to create 'shear stresses' between particles when they need to flow, or move about, in the annular space. Most of your large, cryogenic tanks today use perlite. Those that don't use MLI, mostly on hydrogen and always on helium tanks. I've not heard of microspheres being used in general industry. It's obviously very new but probably also limited to aerospace at the moment.

One other thing you might consider to reduce heat leak if you really need to do so would be to add a shield which can reduce heat leak by a factor of 10.

 

[iainuts]

Couple other thoughts... Have you verified the vacuum after the vessel has cooled down and been in operation over a period of time? Do you see the vacuum rising and does it ever get as high as 10 micron? It should drop considerably after the vessel is filled and cold and shouldn't rise much over time. If you start out at 10 microns, I'd think it should drop to less than 1 micron after the fill.

What pump out port are you using? Have you verified it can handle significant amounts of vibration? We've found many off the shelf units leak under vibration and have designed our own for use in portable equipment. I've also seen pump out ports with pipe thread on the vacuum side for the thermocouple gage tube. Those pipe threads are also prone to leakage.

 

[elbrownos]

Just checked the vacuum and it's less than 1 micron (gauge is reading 0.000 Torr).
I couldn't get it below 80 microns warm so clearly there was water vapour which is now frozen. I thought this might be an issue with the microspheres as they were sitting around in our store for a few months unsealed.
I'm using a Cryocomp V1000 port as found on Chart LNG tanks, with a 1/8" NPT side port for the vacuum gauge.
The microspheres do flow quite easily, especially when excited by a pneumatic vibrator. They also turn to smoke so you need a sealed system for handling them.
When you say shield, do you mean a vapour cooled shield?
It looks very effective but perhaps not feasible for a small tank.
Barron's arrived and looks very useful.
There's a list of reports linked at the bottom of that webpage.
Have you got any experience with getters?

 

[Compositepro]

I do have experience with microspheres. They do flow almost like water. It seems that inner wall of your insulation space is acting quite well as a getter.

 

[iainuts]

The Cryocomp V1000 should be fine but I wouldn't use the 1/8" NPT port for a production unit. They are notorious for leaking. Just put it on your test rig since that's where you need it anyway.

Regarding shield, yes a shield that uses the LNG taken out of the tank, be it gasseous or liquid. It would be expensive to integrate into the design I know, so that's the trade off.

Regarding getters, I'm not that familiar with the various types. As Compositpro mentions, the inner vessel will do a great job pulling any water out.

One other tip is to heat the inner and outer vessels when pulling a vacuum, preferably around 150 to 200 F to help get the water molecules out.

 

[HDS]

"The spheres are impervious, they contain low pressure O2 & SO2 "

Are there problems with corrosion when the gases leak out?

Aerogel beads might work.

 

[elbrownos]

It's great you're all so interested in microspheres but what I actually need help with is the design of inner vessel supports and piping feedthroughs for minimal heat transfer.
Due to the small size of the vessel (approx 1" vacuum space) there is no room to do anything intricate.

 

[elbrownos]

Aerogel looks like a good option, similar performance to microsopheres but perhaps easier to handle.
The boiloff is around 5% per day, corresponding to around 20 watts of heat inleak.
What's a good target, 1 or 2% per day?
Level measurement is another challenge. Capacitance gauges seem to be popular but thermally I can't think of a good way to install one.

 

[iainuts]

Compare your results to a typical liquid nitrogen cylinder, 200 Liters at 2% per day NER which is around 6.6 watts. You should be able to get it down to 5 watts without a shield, but you may not have room with only 1" of annular space. What percentage of your 20 watts is due to thermal conductivity? And have you assumed a constant thermal conductivity for your analysis or are you using real data?

You might want to see what liquid cylinder manufacturers use for a level sensor. Basically, it's a spring loaded float. Here's a web site for example:

http://www.chartindustries.com/Content/distribution/products/packaged-gas/dura-cyl.aspx