High strength metals for external pressure vessel

[TeejT]

I have to design a corrosion resistant cylindrical pressure vessel that withstands high external pressures. I am limited to a certain wall thickness which, according to testing already performed, is insufficiently strong if the tube is made out of stainless steel. There is some machining required in reaming the ID of the tube to tolerance to accept an O-ring plug. What material options do I have to improve the pressure rating? Inconel, titanium? Any other suggestions? Again, the ends of the tube need to be machined to an ID tolerance spec and roughness spec to accept the O-ring plugs, so the material should be relatively easily machinable, or at least not impossibly difficult to machine. Thanks!

 

[rb1957]

guys, there are Way better (stronger) steels out there ... fcy = 27ksi for steel is Really low, almost annealled.

look up (google) MIL-HDBK-5, you can easily get 160-180 ksi (ftu, that is,) steels; fcy >100ksi (i think). High Strength Steel should be cheaper than Titanium, and similar (or stronger even; 300M, ftu = 280ksi) than Ti (6Al4V for example, ftu = 160ksi).

i believe pr/t works for external pressures too, only the resulting stress is compression.

 

[TeejT]

rb1967,
I need a metal with good corrosion resistance. Stainless steels are much weaker as far as I know. Correct me if I'm wrong please.

 

[TGS4]

rb1957 - the primary failure mode for external pressure is buckling. In elastic buckling, as pointed out by MikeHalloran above, is governed by the elastic modulus and the Poisson ratio.

It is only when you get to plastic buckling where the yield and ultimate strengths start to play a role. In most applications, it is elastic buckling that will dominate, so high strength or low strength will make no difference...

 

[mechengdude]

How bout Nitronic 40 or 50

 

[rb1957]

stainless steels go up to 180ksi (301 full hard, for example).

i think this is a very long cyclinder ... (L/r)^2 >5r/t
L = 24", r = 1", t =0.1" ... 576>50

then fcr = nu*0.25*E/(1-u^2)*(t/r)^2

nu is a plasticity correction, call it 1
u is poission's ratio, call it 0.3
fcr = 0.25*30E6/(0.91)*(0.1/1)^2 = 82ksi
pr/t = 10000/0.1 = 100ksi
so you'd need more than 0.1" thick of a stainless steel with fcy > 100ksi (@ 200C, something like 80% of RT)

t = 0.12, fcr = 111ksi, pr/t = 83ksi