Stress analysis of Jacketed Piping

Consider employing a decent stress engineer and save yourself some cost in the long run.

One important consideration is that the tensile restraint of the outer, usually cooler pipe is not so great as to buckle the expansion of the internal pipe. What you're really making is a kind of an axial bimetalic thermometer, without the "bi". Since it doesn't take much temperature difference to get a large stress and a big axial compressive force in the interior pipe, which is easily restrained by axial tension of the outer pipe, buckling can easily occur and the inside pipe will contact the outside pipe at the buckle point. Loads must be keep smaller than the Euler critical buckling load for the length between restraints or internal spacer distance.

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus

a234f56,

With permission, post this in the COADE Inc.: CAESAR II forum.

The other thing that I typically do is double the corrosion allowance for the inner pipe, since it is exposed to (admittedly different) process fluids on both sides of the wall. This exacerbates the concern raised by BigInch. However, I usually see the hot fluid on the jacket side (e.g., sulphur) and the warm-up procedures typically require the jacket fluid to come up to temperature first, so for me, usually the core pipe is in tension.

Regards,

SNORGY.

Hot on the outside could be worse since the column lateral buckling would not be restricted by the outer diameter and therefore not be a self-limiting phenomenon with secondary moments increasing any eccentricity of column axial loads.

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus

There are many particular problems that are not immediately evident in design of jacketed pipe, whci can lead to rapid failure. Such as ... design of closure plates or end flanges, proper modelling of pipe in stress analysis software. Pressure limitation in jacket (testing) to avoid collapse of the inner.

The guy asked about stress analysis of ... I still say, put it in the hands of an expert.

Ya. He's trying to do that now.

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus

REFERNCE C2 Applications Guide 6-4, 6-5

Jacketed Pipe
Jacketed piping systems are input by running the jacket elements directly on top of the core elements where the two are concentric.

A very simple way to generate a jacketed pipe model is to run through the entire core and then duplicate the core piping using a proper node increment (such as 1000). This will produce a second run of pipe, which will be modified to build the jacket model. For the jacket, change the pipe size, temperature; bend radii, etc., to finish the model. Then attach the jacket and core by changing the node numbers and adding restraints.

Typically, the end caps connecting the core to the jacket pipe are much stiffer than either the core or the jacket. For this reason node pairs like (10 and 1010), (25 and 1025), (35 and 1035), and (40 and 1040) are often joined by using the same node for each, i.e. the displacements and rotations at the end of the core pipe are assumed to be the same as the displacements and rotations at the end of the jacket pipe.

Internal spiders offer negligible resistance to bending and axial relative deformation. Node 15 might be connected to node 1015 via a restraint with connecting node. For an X run of pipe, rigid restraints would exist between the two nodes for the Y and Z degrees of freedom.

The +Y support acting on the jacket at node 1020 does not cause any stiffnesses to be inserted between 20 and 1020. Node 20 is included in the model so that outside diameter interference can be checked at the 20-1020 cross sections. Should there be any concern about interference, or interference-related stresses at the 20-1020 nodes, then restraints with connecting nodes and gaps can be used to approximate the pipe-inside-a-pipe with clearance geometry.

Since CAESAR II constructs the jacketed piping model by associating nodal degrees of freedom, the program really does not know one pipe is inside of another. Therefore the following items should be considered.

If both the jacket and the core are fluid-filled, the fluid density of the jacket must be reduced, to avoid excess (incorrect) weight.

If wind or wave loads are specified, the wind or wave loading must be deactivated for the core, or else the core will pick up wind load.

The core pipe should probably have its insulation thickness set to zero.

Somebody ask about Caesar?

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus