`From experiance,
For Vacuum conditions in standard c stl/s stl pipework, when does the size and wall thickness become a problem?

TJN,

The obvious answer, of course, is that vacuum becomes a problem at high D/t ratios. So the simplest answer is, "The problem occurs when the pipe diameter is large and/or the wall thickness is thin." Another factor is the length of the piping or the distance between stiffeners expressed as the L/D (Length/Diameter) ratio. However, if you're looking for a specific D/t or L/D ratio at which the problem occurs, I don’t think you’ll find one.

As you know, B31.1 and B31.3 both refer you to the ASME Boiler Pressure Vessel Code (BPV) to calculate vacuum conditions (or more specifically, straight pipe under external pressure). (I don't know how non-US codes deal with the issue.) The BPV Code is primarily concerned with vessels of known length. That presents a problem when attempting to calculate pipe. For instance, if you were writing a piping spec, you’d want to be certain that the piping would be safe at any length the designer chose. Fortunately, as the piping system gets longer the effect of external pressure diminishes: While the effect on a system with an L/D ratio of 2.0 is much greater than that of an L/D ratio of 4.0, the difference between L/D ratios of 25.0 and 50.0 is negligible.

Let me give you some specific examples. I recently calculated the required minimum wall thickness for 40”, 32” and 30” SS 316L piping. Corrosion allowance was 0.03” and mill tolerance was 12.5%. I assumed the maximum length to be 40 feet. The D/t ratios I give below take the corrosion allowance and wall thickness into account. The results under full vacuum were as follows:

40” pipe, D/t = 212:
0.250” wall thk: FAIL
0.375” wall thk: FAIL
0.500” wall thk: PASS

32” pipe, D/t = 169.5:
0.250” wall thk: FAIL
0.375” wall thk: PASS

30” pipe, D/t = 158.9:
0.250” wall thk: FAIL
0.375” wall thk: PASS

As you can see, the critical D/t ratio for the given systems was somewhere around 200, but to be certain you have to slog through the somewhat arduous steps in the Code.

Hope that helps.

If you think the method is laborious consider this.... its based upon research done with manufactured metal cylinders circa ~1935 after all the test data was assembled voila what you see in the code which was a great simplification compared to the original research!