Route to resistance - hybrid construction
Hybrid construction could prevent rapid gas decompression in elastomer seals. Peter Warren, Andrew Douglas and Stephen Winterbottom of James Walker Technology Centre, Cockermouth, UK, outline the patent-pending concept.
Special elastomers are designed to resist this phenomenon, and there are standards against which they are tested. The most commonly used is Annex B of Norsok M-710 Rev 2, Qualification of non-metallic sealing materials and manufacturers, issued by the Norwegian Technology Centre. The test regime soaks sample ‘O’ rings, typically in a CH4/CO2 gas mixture, for 72 hours at a minimum pressure of 15MPa, then subjects the samples to 10 rapid gas decompression (RGD) cycles over 10 days. The level of damage is assessed against a clearly defined set of criteria.
With oil and gas extraction moving into increasingly hostile environments, the need for low temperature sealing is becoming more important. Elastomers that have low glass transition temperatures (Tg) are based on modified polymers with reduced mechanical properties, and do not exhibit good RGD resistance. Moreover, RGD resistance at specific test temperatures becomes more difficult to maintain as the cross-section of the seal increases. Slowing down the rate of decompression, particularly over the last 6.9MPa, can reduce the level of damage, but is unrealistic in many real-life situations.
This presents two problems to consider – producing elastomer seals that will work at low temperatures, yet meet the requirements of Norsok M-710 Annex B, and developing a technology that limits RGD damage in larger cross-sections.
Extensive research into e l a s t o m e r compounding technology has shown that it is unlikely that stepchange improvements will be achieved with conventional compounding systems, which is why alternative approaches are being explored. Although the work on hydrogenated acrylonitrile-butadiene (HNBR) is the main focus, other polymers have also been studied.
An HNBR for use at low temperatures will have a low acrylonitrile (ACN) content of about 17%, plus an additional monomer to reduce the crystallinity. When compounded to give the best possible properties, it has a nominal hardness of 85 International Rubber Hardness Degrees (IRHD). Compared to a medium ACN grade of about 36% without the extra monomer, the difference is considerable, particularly at the minimum Norsok test temperature of 100°C:
To illustrate this point, a 6.99mm cross-section ‘O’ ring moulded in low ACN material can split right through when tested at 15MPa in a 90/10 gas mixture of CH4/CO2 at 100°C with a 72 hour soak, followed by ten decompressions at 3.75MPa/minute at 23 hour intervals:
When tested under the same conditions, ‘O’ rings moulded from the medium ACN elastomer would be undamaged. The low temperature sealing performance of this, however, is quite different:
Elastomer seals are normally fitted at temperatures above their Tg, and therefore readily deform to fit the housing and form a seal. If the surface layer has a lower Tg than the core, in theory, the components should continue to seal down to the temperature capability of the outer layer.
This reasoning has led to the concept of the hybrid seal, that exploits the excellent RGD resistance of medium ACN in the core, and the low temperature capability of low ACN at the surface:
When tested for static sealing properties, the seal with a low ACN outer layer, at nine per cent of the radial thickness, sealed efficiently down to -36°C, whereas one with an outer layer at 17% of the radial thickness sealed down to -46°C. The RGD performance of the hybrid with the thicker outer layer – on ‘O’ rings of 6.99mm cross section – was significantly improved, with only minor splits observed at the interface in the areas of maximum stress that could be eliminated with improved bonding:
The low-temperature sealing performance of the hybrid is related to the thickness of the outer layer. Increasing this will get closer to that of seals moulded solely from the low ACN elastomer, while maintaining improved RGD performance. The principle of dual Tg elastomers can be applied to products other than ‘O’ rings.
Further improvements to RGD performance can be made by introducing energy-absorbing domains through the seal cross-section. Where there are inadequate mechanical properties to resist RGD, failure is mostly in the form of severe cracks or fractures. The level of retained properties at the application temperature required for RGD resistance is the subject of speculation. However, the ability to restrict crack growth will limit the level of damage should the limitations of the elastomer be exceeded.
Elastomers with high tear-resistance tend to have reduced resistance to compression set. The challenge is to design a product that incorporates the advantages of ‘dead’ energy-absorbing high hysteresis elastomer, whilst retaining acceptable sealing stress over time at the application temperature.
Using the polymer for the low ACN compound, a formulation was developed to maximise hysteresis and subsequent energy dissipation. To obtain a balance between energy absorption and stress relaxation, the energy absorbing (EA) material was employed in thin layers within a host elastomer – in this case, the standard low ACN compound.
To study the effect on overall performance, a 0.5mm layer of EA material was moulded into standard angled crescent tear samples where both components had a nominal hardness of 85IRHD. A high level of compression set was designed into the compound to enable the effects of poor stress retention to be studied.
The curve above shows the tear test results. The first peak is the failure of the ‘low’ ACN compound, which drops to the force required to initiate tearing in the EA component. Calculating the area under the combined curves gives a 100% increase over the test without an EA layer. The crosshead travel to reach failure also doubled with the inclusion of the 0.5mm layer.
Further work showed that limiting of EA layer thickness helps to retain the sealing force over time. Also, careful placement of the layers in key areas of the cross-section maximises their effectiveness, with points of location differing with various product configurations.
The image, right, shows an early practical example of an EA layer in an ‘O’ ring after it was subjected to extreme RGD conditions. The cracks in the outer layer were stopped by the EA layer. Locating the layer more towards the OD of the seal, or introducing a second layer, would limit the damage still further.
Finite element analysis to identify areas of maximum stress will assist in designing seal constructions. Where improved sealing performance at low temperatures is required for seals of larger sections, a low Tg outer layer can be used.
Peter Warren, MIMMM, Head of Materials Engineering, James Walker Co Ltd, Cockermouth, Cumbria, CA13 0NH, UK. Tel: +44 (0)1900 898277. Email: email@example.com