USING FEA TO MODEL GASKET WEAR
In the sterilization system and the processing equipment and systems being sterilized, thermal cycling presents a further
problem. When the test manifold was disassembled after thermal cycle testing, a certain amount of wear on the face of the
gasket seal was observed. The surface was roughened and some of the gasket material was gone. It was concluded that the wear
and fretting of the gasket material was probably caused by the radial expansion and contraction of the gasket during heat-up
When the gasket expands radially into the bore of the tubing, it is no longer constrained between the faces of the two ferrules.
The unconstrained portion of the gasket is free to expand back to its original thickness. The gasket material was characterized
and the shape of the unconstrained portion of the gasket was modeled using finite element analysis (FEA) techniques.
The shape is bulbous. During expansion, the gasket extrudes radially, and the unconstrained portion expands axially. During
contraction, as the gasket moves back into the constrained space between the ferrules, the surface of the unconstrained portion
of the gasket is dragged over the relatively sharp metal corner formed by the bore of the ferrules and their flat faces. This
explains the wear that was observed. The gasket material that is worn or scraped off can end up inside the system, eventually
in the fluid stream.
When gaskets are subjected to high temperature and high compressive loads, some gasket materials can take a compression set
and become loose. Compression set is the tendency of an elastomer to lose its memory under stress and not return to its original
shape when the stress is removed. A common industry practice to address this challenge is to retighten all of the fittings
after thermal cycling or sterilization. During thermal testing, regular leak tests indicated that the ISO 2852 fittings needed
to be retightened after every fifth thermal cycle through the first 15 cycles, because the gasket had taken a set.
These results can explain why containment is lost immediately following a sterilization cycle. In such situations, a usual
maintenance procedure would be to retighten the clamp even tighter, which would cause further extrusion and a larger dam,
magnifying the other kinds of problems discussed.
A FITTING SOLUTION
Another fitting design that provides solutions to the issues discussed also was tested. In this fitting (Figure 1B), the configuration
and cross-section of the gasket and the face of the ferrule are different from the ISO 2852 fitting.
The gasket consists of two parts—the rib and the crown—each having a specific function. The rib portion is a rectangular shape
with flat faces. When clamped between two ferrules, the seal is made at the rib. The function of the large mass of material
in the crown of the gasket is to control the amount of gasket extrusion toward the bore of the fitting. The faces of the ferrules
are machined to accept the crown of the gasket and align the two ferrules for assembly of the connection. A metal-to-metal
stop is provided at the maximum outside diameter of the ferrules to limit the amount of load that can be applied to the gasket
during initial make-up and prevent overtightening.
The gasket was configured to maintain proper "squeeze" over its complete cross section. At initial make-up of the connection,
compressive force is applied to the gasket with the same type of clamps used with an ISO 2852 fitting. Controlled extrusion
permits a small amount of extrusion into the bore of the fitting, creating a stable bore-line seal, and avoids undesirable
concavity at the seal point. The majority of the extrusion is taken up in the crown contained in the chamber formed between
the faces of two ferrules. The chamber is not completely filled, to accommodate expansion of the gasket material during thermal
These fittings were assembled into a manifold identical to the ISO 2852 manifold discussed earlier and tested in exactly the
same way. The results were quite different and are shown in Figure 2. None of these fittings required retightening during
the thermal testing. Flow in this fitting also was modeled with the CFD technique, based on a velocity of 5.5 ft/sec.
The small amount of controlled intrusion at the gasket seal looked no different than the inside of a full penetration buttweld.
There was no constriction of the flow through the connection and no increase of velocity through the connection (Figure 5).
The CFD model showed no entrapment zone downstream of the gasket seal.