INTRODUCTION

Despite their relatively long history, projects still experience problems with gaskets. This presentation took a step back to the basics, covering the technical background behind the materials and the main design issues relating to gasket compression loads and the tunnel watertightness. Gasket requirements change with each project, and just copying what happened on the last project does not guarantee success.

There are two key aspects of a gasket’s performance requiring different considerations:

  • Gasket compression and its impacts – tolerances that make gasket larger, the groove smaller or cause the gasket to be forced into more intimate contact.
  • Watertightness – tolerances that make the gasket smaller, the groove bigger or cause a widder gap between segments and therefore reduces the load across the gasket.

The term ‘gap’ does not necessarily equate directly to a space between concrete contact faces but more of how the compression of the gasket changes.

‘WATER IS THE ENEMY’

The expression ‘water is the enemy’ is true for many reasons:

  • Ground can enter along with the water, which could be catastrophic as the lining relies upon the external ground support to maintain its stability. A gasket can be argued to have a function related to maintaining the stability of the lining.
  • Tunnel operations and internal equipment can be affected, including the need to provide and maintain drainage and pumping systems. Water flows bringing in fines can cause silting or sintering within the drainage system.
  • Ice might be a problem in some locations, or water on the surface can change breaking characteristics or traction.
  • Durability. Concrete with flowing water on the surface or with ‘wet & dry’ cycles is less durable than dry concrete. Flowing water can bring in and replenish damaging salts that affect the concrete and the protection it provides to any reinforcement Sustainability and carbon zero targets are accelerating the use of low carbon concretes and these binders are generally more susceptible to carbonation as they are lacking in Portlandite, which acts as a pH buffering agent in the highly alkaline pH range. As cements and binders change, we will need to re-think the durability aspects and how water and wet/dry cycles impact upon the long term lining durability.

Controlling or stopping water from entering into the tunnel is therefore about more than comfort and convenience, and the humble gasket offers a relatively easy and reliable method to ensure safety and structural stability.

TYPES OF GASKETS

There are a number of gasket types and numerous sizes and combinations available. The basic style is a section of ‘solid’ rubber with voids that allow for compression (see figure 1). Different design ‘types’ include:

  • Glued in gasket – the profile is glued into a groove formed during segment manufacture
  • Cast-in gasket – developed mainly to reduce the likelihood of water flows between the gasket base and the concrete, this type is placed in the mould and cast directly into the segment and held in place with anchors. Anchorage systems are continuously developing including a fibre facing at the base to improve fixity, which will have less impact on reinforcement cover and potential edge spalling, with better concrete compaction around fibres than with the anchor legs
  • Hydrophilic – these are generally thinner ‘solid’ pieces of rubber with the ability to slowly increase in volume when exposed to water. Exposure conditions (the water chemistry) dictate the amount of expansion that will take place. Their degree of expansion decreases over successive wet and dry cycles so they should not be relied upon long term where conditions change. Also, there is now some concern about the potential for microbial attack on this material, and that would need to be explored in detail for any specific project use
  • Composite gaskets – these are also relatively common, either with a separate gasket and hydrophilic element, or with the hydrophilic material bonded onto or inserted into the solid rubber portion
  • Double gasket systems – some projects are using these systems

Gaskets – considerations: material

While there are specific differences for the pure hydrophilic gasket, the talk considered the design and testing requirements for all types of gaskets.

The majority of gaskets are made up of a ‘solid rubber’ element with voids, produced through a continuous extrusion process with the material then passing through a series of microwave and curing chambers to ‘fix’ the material. With rubber, the curing is also induced by the addition of a ‘crosslinker’ in the formulation, so curing is not entirely a mechanical process and relies upon the material chemistry and the mix design. The degree of crosslinking determines the rigidity and durability, as well as other properties of the material.

The material is made up of a number of ingredients within the main rubber compound – of which there are several types, examples being Ethylene Propylene Diene Monomer (EPDM) and Nitrile rubber (NBR).

The different ingredients make up a usable and durable product and include the main compound, fillers, oils, crosslinkers and other polymers. Each final mix will exhibit different behaviours and chemical resistance.

EPDM for example is a stable, general purpose material that can be used in the majority of tunnel gasket cases. NBR has good mechanical properties and better resistance to mineral oil-based lubricants and greases than EPDM. This doesn’t mean EPDM cannot be used in the presence of hydrocarbons – the design life, concentrations in the surrounding environment, and level of exposure all go into assessing the best value material.

The materials are classified based upon their basic chemistry to the classification in ASTM standard D-1418, and can have varying characteristics. For example, EPDM rubber can contain between 45% and 85% of propylene and be cured in different ways, providing it with a range of properties, and still be technically classified as EPDM to the standard.

Not all gasket rubbers are the same, and even referring to a generic type, such as EPDM, does not fix the characteristics of the gasket material.

The base gasket material (not the final gasket profile) exhibits the following behaviours:

  • Compression Set (see figure 2) – the ability of the material to return to the original shape following a period under load. Upon unloading, there would be an immediate partial elastic response followed by a further, slower recovery, related to time as well as temperature of the material. Poor compression set equates to limited recovery, for example plasticine. However, for tunnel gaskets significant unloading is not expected and stress relaxation is probably the more important relationship.
  • Stress relaxation (see figure 3) – relaxation of the material when under load, or the gradual loss of force required to maintain the same deformation.

To a non-polymer material specialist, it can be difficult to differentiate between Compression Set and Stress Relaxation. It is, therefore, probably easiest (although technical incorrect) to think about them together for a gasket situation. For a given profile, together they determine the energy that is stored within the compressed gasket at any given time, and it is the energy stored in the system that provides the water sealing capability.

Gaskets – considerations: profiles and tolerances

Both material formulation and the void arrangement and size (i.e., the design of the gasket profile) can affect the compression characteristics and hence the compression loads and the watertightness. It is likely that the highest compression loads will occur when the two gaskets are in line (zero lipping) – but this is not guaranteed because it is the internal structure of the struts and voids that partly dictate behaviour.

Making the gasket into a ‘window frame’ shape to fit onto a segment requires the cutting of the gasket into lengths and a joining process at corners, achieved by vulcanizing the rubber in a mould. The result of this is a solid piece of rubber rather than voided as in a typical profile, which will have different behaviour characteristics. It is a potential hard spot, although manufacturers are developing corner profiles that are relatively ‘soft’ compared to historical profiles. It is, however, something that requires consideration when designing the gasket and for the testing regime.

The rubber itself, for our considerations, can be considered to be almost incompressible, although stress relaxation is still important. The behaviour of the gasket in place is determined by a number of factors, including the material, profile design (the size, void volume and arrangement), and also by all tolerances related to gasket manufacture, segment manufacture, ring build and ring deformation under load.

For gaskets, it is not uncommon for specifications to omit the tolerance class for the manufactured product. However, the width of the gasket is also important as are the void sizes within the profile. Fortunately, most manufacturers meet the requirements of BS ISO 3302 (see table) and will supply a product to class E2 unless specified otherwise.

Note that the tolerances are dependent upon individual dimensions, so just specifying a height tolerance does not define all other tolerances, and for cast-in gaskets it is the protrusion height that is important, i.e., the relative level of the nib controlling its position in the mould.

When dealing with any tolerance it is important to take specific care with reference surfaces. For example, dimensioning the groove depth from the recess concrete rather than the contact surface of the joint faces should be avoided or there will be cumulative tolerances. It is important to consider which dimension is important to the design and the gasket characteristics, and then to specify that dimension.

For the segment tolerances, the ring length, the deviation of the joint planes, and the position of the groove can impact upon gasket performance.

Joint plane deviation is an interesting aspect. Ideally, the groove depth would be referenced from the concrete contact faces, but this is not always the case. There is also no easy practical way to make an appropriate measurement to every point on the surface. Any checks are, therefore, commonly made at the location of the concrete immediately adjacent to the groove and the deviation of the joint plane away from this point has the potential to impact upon the overall geometric arrangement and relative positioning.

Therefore, although there is an argument that this is not a tolerance that affects the gasket compression in particular, it is normal to take it into consideration. Finally, for ease of discussion concerning ring build/ deformation, consider long-term deformation of the lining within the same tolerance as the ring build, but be aware they are not the same and they are possibly ‘owned’ by different parties – for example the construction ring build tolerance is owned by the contractor, the deformation of the tunnel under load is probably owned by the designer or possibly the client. ‘Time’ might also be relevant – the construction build tolerance will occur instantaneously, while the deformation under load might take months or even years and its inclusion in any test may require additional consideration.

For a deformed and loaded ring there are several consequences for the longitudinal joints. For rigid segments there will be some rotation at the joints where the gaskets are located – this can be in either direction depending on where the joint is in the ring and the impact will therefore be different when considering gasket compression and watertightness (see figure 4).

There is also an additional movement or compression of the gasket that occurs as the joint is loaded and the concrete strains locally. This result develops from the segment lining analysis, but it is relatively small and as long as the assessment has been generous in all other ‘worst case’ considerations it can possibly be ignored in the compression test. This strain usually happens over a long period, but the short term stress relaxation of the gasket is quite high and usually makes this additional compression affect insignificant.

There are also construction tolerances affecting the circumferential plane, which will impact upon the watertightness design where it results in ‘gaps’ in the circumferential joint. However, as these impacts reduce the compression loads of the gasket they can be ignored in gasket compression deliberations.

Gaskets – considerations: summing the assessments

The different tolerances need to be ‘summed’, and the assessment is different for the compression load assessment and the watertightness assessment. It is also important to check both joints.

Compression load assessment – everything that increases compression in the gasket or makes the gasket bigger and the groove smaller is of interest. Some tolerances double up as they apply to both segments.

The main tolerances are:

  • Gasket manufacturing tolerances
  • Gasket groove tolerances (glued-in gasket), or for a cast-in gasket the protrusion above the joint plane (here the tolerance of the machined groove in the mould as well as insertion workmanship might be important)
  • Concrete joint deviation, where this can increase compression.
  • Joint rotation, where this increases gasket compression.
  • Watertightness assessment – tolerances that increase the gap or have the potential to reduce the compression load between the gasket faces are important. As above, some tolerances double up. The main tolerances are:
  • Same tolerance considerations as for the compression assessment, but in the opposite direction
  • Construction tolerances affecting the circumferential joint, for example ring plane, ring length
  • Loading interaction between the bolts/dowels and the gasket on the circumferential joint, resulting from the strain of the connector under load. Simple interaction diagrams between gasket load and connector deformation will help with establishing this effect. However, depending on the method of ring build, the lining in its final grouted and constructed condition may not be loading the connectors at all, and any gap is simply a result of physical tolerances and construction tolerances. In some cases this interaction can be ignored.

Typically, the watertightness test gap results in answers between 5mm and 7mm, and if the result is significantly less than this or significantly more, reconsideration of how all of the tolerances have been applied in combination and at what time in the life of the ring might be required.

Consideration of different ring build scenarios can get very complex, and it is also important to consider which tolerances might be applicable at the worst location within the tunnel, particularly for the watertightness assessment.

For example, the likelihood of all of the worst gasket, segment manufacture and ring build tolerances, along with the worst load deformation occurring at the same time and at the point of maximum water pressure might be very small, and a more statistical consideration might be appropriate (for example using something similar to the Root Sum Squared method).

Consideration of tolerances is not simple, and some practical judgement is required.

Gaskets – considerations: compression load

For the gasket compression, the load deflection test will normally be undertaken in steel rigs within a compression machine in a laboratory. Dimensions of the gasket sample used in the test need to be checked to make sure there is no double counting on some tolerances, or even ignoring some. For example, if the test sample is at the extreme maximum of the profile tolerances this could be taken into account when establishing the maximum load required to compress the gasket.

What is not covered in any test standard is compensating for the loss of air from the holes and voids in the gasket. The test rig should have closed steel ends to prevent longitudinal extension, but these are not air-tight. The difference between air escaping in the test, and what can happen with a sealed gasket on a ring, can underestimate the compression load and it is recommended that the gasket ends should be sealed with a thin layer (smear) of rubber.

The precise compression load is, however, probably less important than the demonstrable impact upon the concrete itself, where the consequences of edge spalling behind the gasket can be severe. Can the concrete be designed for forces that are not measured in a test where vertical load but not the horizontal force is measured? Although it is common to assume that the horizontal pressure is equal to the vertical pressure, this is not 100% reliable because gasket shape is also important – a thinner but wider gasket will exert less force laterally than a deeper narrower gasket, even if the vertical load is similar. Therefore, the concrete design and edge distance should be based upon testing in samples using similar concrete to that proposed in the works.

Historical demonstration from a relevant project that used similar concrete and tolerances, and exactly the same gasket, might also be acceptable.

In the very short-term the force required to compress the gasket reduces rapidly from a peak (see figure 5a). The maximum force is very likely to apply to the TBM erector and to the concrete itself, but as the connectors will not be installed, or at least loaded, for several minutes after the initial compression, lower than maximum loads might be appropriate for connector design.

A second graph compares the load with the amount of deflection (see figure 5b), with results being very sensitive to small changes near full compression. Note also that a typical load deflection graph will be provided with tolerances in terms of its reliability and accuracy, which also need to be considered in any design.

Gaskets – considerations: watertightness

For the watertightness testing, the relationship can be simply expressed as “the higher the compression load, the better the waterproofing”. As well as taking tolerances and the potential joint gaps into account (by which we really mean the reduction in compression force across the gasket), typical tests obviously need to consider the water pressure and the longer term behaviour of the gasket, i.e., its reduction in load due to stress relaxation, and uncertainties.

Some standards, including the BTS Specification, require the watertightness test to be undertaken with an applied factor of safety. It is, however, important to understand the background to any requirement.

The test pressure should be considered in terms of the known relaxation and other factors of safety relevant to the water pressure, for example will it fluctuate uncertainly with climate change.

Typically, testing adopts a factor of safety of 2.0 on the actual water pressure. This probably resulted from an assessment based on a typical relaxation of the gasket to 65% after 120 years of the initial compression load, and a general factor of safety of 1.3 to cater for testing inaccuracies and knowledge of the water pressure. These can obviously be varied depending on the gasket material and its actual stress relaxation, the project design life and any other uncertainties, or certainties, around the water pressure. The test pressure factor of 2.0 is a good guideline but might not be appropriate in all circumstances.

A typical watertightness test graph (see figure 6) will establish a relationship between the ‘gap’, the water pressure, and the offset. These graphs can obviously QUESTIONS AND ANSWERS be used in reverse – for example, a larger gasket than theoretically required might be more tolerant of larger offsets or gaps and hence tolerances might be relaxed. Also, during construction, should the segment/ring joint gap or lipping exceed specification, the graphs can be used to demonstrate acceptance if the water pressure at that location is low enough.

Gaskets – considerations: double and cross gaskets

Double gasket systems are becoming more common under the belief that they improve watertightness – but this is true only under certain circumstances.

For example, it is possible that somewhere along the tunnel the outer gasket will leak. Water then leaks into the small spaces in the joints between gaskets, and, in time, pressurises the full length of the tunnel. Somewhere else in the tunnel there will be another weak spot in the internal gasket and this will leak. The problem is that there is no relationship between where the initial leak occurs and where it is seen in the tunnel and any repair, therefore, can only address the inner gasket. This tunnel effectively has paid for 2 gaskets but is probably only working with one.

To resolve this, isolate areas of the tunnel with a cross gasket between the inner and outer gasket. Now when the outer gasket leaks it is constrained to a small area and if the inner gasket also leaks there is more certainty where the outer gasket has failed. It may not be possible to repair the outer gasket, but at least the problem has been significantly isolated. At present, fully vulcanised double gaskets, including at joints with the cross gaskets, are extremely difficult to manufacture, and the joints of the cross gasket might be glued, or a hydrophilic material used instead. This may not be ideal, but it’s better than no isolation at all.

Gaskets – considerations: fire

A final aspect that needs to be considered with double gaskets is the impact of a fire, if relevant, on the inner gasket. Tunnel linings are often asked to be designed for extreme conditions with temperatures up to and exceeding 1200°C. This results in high temperatures extending into the concrete where, once spalling of the concrete has been allowed for, the gasket may see temperatures in excess of 600°C-700°C.

Gasket material has a service temperature of about 150°C, and at 200°C the material would have melted. Loss of the gasket in a fire may not be a problem and could be an acceptable risk to the project.

However, burning and melting rubbers give off smoke, may contribute to the generation of heat and propagation of the fire, and could also give off toxic fumes. Projects with internal gaskets and fire load requirements need to take into account the material contribution to these aspects, along with all other internal materials when assessing the safety of the system and the ventilation requirements.