On the design of a novel single piston effect seal for cryogenic valves

Factors impacting seal performance

• Seal design and materials
• Groove design and finish
• Inter action valve –seal
• Valve mechanics
• Pressure
• Medium

What makes it than so difficult to perform well at cryogenic temperatures? 

While cooling down, the polymers, used as jacket material, would shrink by a factor 10 times higher than the surrounding groove material and in addition the seal material becomes increasingly harder. These two factors contribute in not a small amount to the difficulties seen at cryogenic sealing.

At liquid nitrogen (LN2) temperature the yield strength and the E-modulus of typical PTFE or PTFE compounds increases 10-fold, the hardness as a consequence, increase also by a factor 10. In addition, if the spring would not stop it, the outer seal lip would shrink approximately 2%.

These radical changes work negative on the seal performance. 

The shrinkage force of the outer lip works opposite to the original spring force, reducing the seating stress and sealing capability of the outer lip. Contrary the inner lip seating stress increases because of the shrinkage force.

The shrinkage because of cryogenic temperatures creates a second issue. Since it is impossible to manufacture a helical spring with a “0” gap at the outer diameter, the shrinking polymer will shrink into these gaps and reduce over a short distance the seating stress. One can try to reduce this gap as far as possible by manufacturing a “0” gap coil. However once assembled in the seal cavity a small or bigger gap will be formed. An attempt to remedy this problem by using an overlapped spring may well do more harm than solve the problem. At the overlap, a small triangle shaped void is created. This allows the polymer to shrink minimally more at this location, so that a kind of wave form load distribution is formed. 

Where at room temperature and higher temperatures a relatively low spring load (1 to 5 N/mm circumference suffice, this not the case anymore at lower and cryogenic temperatures.

The increased hardness of the polymer jacket requires a far higher spring load compared to higher temperatures. At LN2 temperature the hardness of commonly used polymers comes close to the hardness of copper at room temperature. With this kind of hardness the spring will be far too weak to plastically deform the Polymer jacket. And there is not much to do about it. A strong(er) spring will eventually counteract the shrinkage of the outer seal lip in a better way than a weak spring but it will fail to come anywhere close to the yield strength of the polymer jacket at LN2 or LNG temperature. It needs to be said that a stronger spring also yields a lower elasticity. Also the available groove size, limits the increase in spring load and flexibility. In addition a stronger spring also hinders the movements of the dynamic seat ring. Secondly a high spring load would embed partially in the seal lip, thereby reducing much needed elasticity and also its actual load.

There is a way to go around the difficulties encountered in the classical radial seal design. The major problem in the existing radial seal is the outer lip. Differential shrinkage, hardness increase and therefor lack of sufficient and uniform seating load create tightness problems at the outer lip.

When the valve reaches an equal temperature the seating load at the OD lip will not significantly increase. 

With this shortcomings a highly polished inner and outer seal groove finish is a primary requirement. The spring should have a high enough load to counteract the shrinkage force of the seal and seal lip. But the spring load is also limited to a certain level to avoid too much cold flow at higher temperatures and to avoid a too high friction versus the dynamic seat.

By splitting up the sealing task of the seal into a radial part, sealing the dynamic seat, and an integrated face seal, sealing the body, many, if not all of the bespoke problems with the classical radial seal solution can be resolved. 

 

In the below sketch a possible example of such an integrated radial-axial seal is given.  The seal jacket may be machined as one integrated seal or an interlocking radial and face seal. The cover holding the seal, locked in place, in the valve body needs to be sufficient rigid to counteract the system pressure.

The test fixture as shown above, is set up in such a way that the standard seal and the special patent pending radial-axial design is tested under exactly the same conditions.

The pressure is applied to both seals via port P1. Port 3 allows the leakage measurement over the outer diameter of the seal, whereas port 4 allows the leakage measurement over the seals inner diameter. Via port 3 the leakage over the standard radial seal is measured.

A temperature sensor (not shown in the above sketch) is located 1 mm away from the radial outer diameter of the novel seal design.

To cool down, the test fixture it was placed in cool box. The temperature of -100°C and -170°C was created solely by evaporation of LN2. In order to reach -196°C the test fixture was submerged in LN2. Enough time was let to assure an equal temperature over the complete test fixture.

Seal ID 228 mm, Cross section 3/16”

Test medium in all cases: GHe

Prior to cryogenic testing a Helium sniffing test was carried out.

This table reflects the results with a standard shaped novel SPE seal and slightly improved version of a standard SPE seal.

In total 3 different sets of seals were tested. Without an exception the novel SPE design showed the lowest leak rates.
 

One of the standard seals was a 3/16 inch helicoil spring energised seal, the other a V spring energised seal.

These two standard seals at -196° and 100 bar showed a leakage exceeding 1000 ml/min. (max capacity of flowmeter).

1) The novel SPE seal design offers potential to zero leakage at normal LNG operation conditions.

2) The design requires an additional part for the valve.

3) The extra cost related to this design can be offset by;