Bellows Squirm and Externally Pressurised Joints: Reversing the Pressure

When a long run has to absorb a large expansion through a single joint, the first idea is always to make the bellows longer: more convolutions, more stroke. Past a certain point that stops working. The longer the bellows, the less able it is to stay on its axis under pressure — and it buckles sideways.

How squirm happens

In an internally pressurised bellows the fluid presses on the inner face of the convolutions. Each convolution tries to open under that pressure, and together they produce a tendency to extend along the length of the bellows. If the bellows is short and stiff, its own rigidity balances that tendency. If it is long and slender, the situation resembles a slender rod loaded in compression: above a certain pressure the smallest geometric asymmetry starts to grow, and the bellows leaves its axis in an S shape. This is squirm.

When it happens, the consequences are:

• One side of the convolutions is over-strained while the opposite side is compressed; stress is no longer distributed evenly around the circumference.
• Axial capacity is lost: instead of using its stroke, the element deforms by bending.
• Stress concentrates at the end welds, and cracks start there.

The hard part about squirm is that it is not gradual. It is not ageing but an instability: it begins once pressure passes a critical value. So “it has been fine so far” is no assurance — the same bellows can leave its axis abruptly when pressure rises a little.

Reversing the pressure

In the externally pressurised design the fluid does not pass through the bellows. The bellows is enclosed in an outer shell, the fluid flows through an inner pipe, and pressure acts on the outside of the bellows. Mechanically the difference is fundamental: pressure no longer tries to open the convolutions but squeezes them from outside and keeps them in line. The mechanism that would push the thin wall off axis is gone. Think of a wire in tension rather than a rod in compression — a wire in tension does not buckle.

There is a second benefit: the bellows sits outside the flow path. Turbulence and entrained solids do not settle into the convolutions, and the flow passes cleanly through the inner pipe.

What it buys in practice

With the squirm limit removed, the bellows can be made longer. The result is one element absorbing a far longer axial stroke — and on long runs that translates directly into the number of supports.

Consider the alternative. When a single standard bellows does not have the stroke, the run must be divided: a separate joint for each section, and for each joint a separate main anchor to direct movement into it. Every anchor is a force transferred into the structure and a piece of civil work that needs room — in a basement, a service tunnel or on a pipe rack. One externally pressurised element removes a part of that chain.

Hence the typical applications: long, high-pressure runs. District heating distribution, steam mains, power plant feed lines and tunnel crossings where expansion approaches the metre range.

Guides are still essential here

The externally pressurised design protects the bellows against squirm, but it does not hold the pipe on its axis. For any axial joint to work as intended, the line has to be held straight by guides; otherwise the expanding pipe wanders sideways and imposes angular rather than axial movement on the joint.

This is a common assumption error with externally pressurised products: “squirm resistant” describes the stability of the bellows itself, not the stability of the line. A long-stroke joint on an unguided line cannot use its own advantage.

Where the decision starts

If the calculated movement exceeds the stroke of a single standard bellows, there are two routes: divide the run, or use one externally pressurised element. The decision starts geometrically rather than economically — is there anywhere on this run to build an anchor? If not, the question has to be solved on the product side.