The problems encountered by lifeboat designers seeking extra speed have been mentioned before in these pages, as boat design is never as straightforward as it may seem. Increased speed is not just a question of bigger engines or even just of greater strength or a different hull shape, all are woven into a complex dilemma for the designer. But there is one factor which returns to haunt the naval architect- weight. The key to the RNLI's latest fast lifeboats is their Fibre Reinforced Composite construction, producing boats which are light enough to attain the required speeds and strong enough to withstand the stresses they impose.

These materials may appear expensive, but as part of the continuing research which the RNLI is conducting into the longterm life of the materials one independent report has come up with an interesting conclusion - their cost is barely more than that of conventional glassfibre from the viewpoint of residual strength after a period in service.

The RNLI is carrying out research in conjunction with the University of Southampton, the DTI and Lloyds, and also at Bournemouth University where Peter Pangbourne, a qualified Quality Control Engineer and an experienced yachtsman, tested the materials while researching the use of acoustic emissions - effectively 'listening' to the noise made by materials as they flex -as a form of non-destructive testing.

Testing timesi: o I or fast craft higher speed can be obtained either by reducing weight or by greater engine power - but additional power increases weight in the form of heavier engines and a heavier fuel load for a given range.

High speed means higher hydrostatic and dynamic pressures on the hulls, which in turn calls for stronger, stiffer and therefore heavier glass reinforced plastic (GRP) hull panels and structures - which is obviously incompatible with the higher speed requirement.

Weight saving and stronger hulls can be obtained by using lighter and/or stronger reinforcements based on or incorporating para-aramid fibres (such as Kevlar and similar materials with different trade names).

Conventional single skin GRP panel construction uses a resin-rich chopped strand mat (CSM) surface, backed-up by layers of glass woven roving (WR) and chopped strand mat using a marine grade polyester laminating resin.

The resin-rich CSM outer surface layers prevent the panel utilising the maximum strength of the underlying woven layers because the chopped strand mat is relatively weak, and a crack in the outer layer is taken as the failure criterion for marine laminates in service. Consequently, everything is done to minimise deflection, and so stiff, heavy laminates have tended to prevail - despite their poor utilisation of the ultimate strength potential of the structural woven rovings within the laminate.

Traditional panel design is governed by simple bending theory, and consequently thickness plays a major part in limiting deflection, which can lead to heavy craft.

Advanced panel design is governed by strength considerations, making use of large deflection theory, and thus better use of the load carrying capability of the laminate. This concept is similar to that used for designing aluminium hull panels which are allowed to deflect up the point of irreversible deformation. In theory composite hulls' panels can also perform at high levels of deflection but without permanent deformation. It is however necessary to limit deflection and hence surface strain to avoid surface cracking under fatigue conditions.

If the installed power in a particular craft remains constant the increase in performance associated with reductions in structural weight may result in an increase in the dynamic pressure to which the vessel is subjected and so it is vital that the fatigue properties of the hull laminates are understood.

To investigate this fatigue three samples were tested - '1010' a composite used in the RNLI's Severn class and costing £221.70 per square metre, and for comparison '1020' a similar material but using a different epoxy resin and costing £199.53 per sq m.

'2030' was also used for comparison, being a more conventional glassfibre laminate using polyester resin and costing £32.58 per sq m. A test rig was used to apply an increasing load, and when the panel failed the load, deflection and type of failure were recorded.COMPARISON OF STRESS STRAIN The test results shown in the graph - with the flexural stress calculated at the surface, assuming that the neutral axis is in the middle of the thickness - indicate that 1020 and 1010 are considerably stronger than 2030 (by 98.03% and 88.05% respectively) which would be expected.

The mean difference of the slope of the graph between 1010 and 1020 is 6% and the ultimate stress of 1010 is 4% greater than 1020.

COMPARISON OF FLEXURAL FATIGUE To test flexural fatigue the test pieces were deflected and released at a maximum rate of six cycles per second, at which no rise in temperature could be detected. This was repeated until the sample failed or one million cycles had been applied.

The resulting data showed 1010 and 1020 to be 56% stronger than 2030 - with a mean difference of 6% and a difference of 19% at one million cycles.

FLEXURAL FATIGUE LIMIT The Flexural Fatigue Limits follow the established pattern with 2030 at approximately half the strength of 1010 and 1020, although the difference between 1020 and 1010 is larger than the Stress Strain results. It would appear that 1010 is a better material where Flexural Fatigue strength is a requirement.

FLEXURAL STRENGTH AFTER AGEING Test pieces of each type of material were then 'aged' so that the strength of the materials could be compared after 'life in service' To do this two samples of each were deflected to the fatigue limit, the first over 1,000 cycles, and the second over 10,000. Samples that had not failed in previous tests and which had reached 1,000,000 cycles were also tested.

Testing these 'aged' samples showed the reduction in maximum stress to be 10% for 1010,16% for 1020 and 56% for 2030. Given the very small reduction for 1010 and 1020 it can be argued that they do not have an appreciable drop in strength due to cyclic 'ageing', while 2030, the conventional GRP, appears to age much more.

COST AND STRENGTH AFTER AGEING The costings for the three materials to give an equal strength after 1,000,000 cycles at their fatigue limit shows that although 2030 is still the cheapest the difference between 2030 and 1010 is only 14.5%. But, while 2030 would be the least expensive material the hull would be 30% heavier.

The increased strength of 1010 and 1020 over 2030 have several advantages that offset their additional cost - for vessels where weight is of primary importance and for vessels where strength is the primary consideration, such as lifeboats.

Although, without detailed analysis of total design requirements, it would appear that 2030 is the best material to use thereduction in weight and increase in internal volume gained by using either 1010 or 1020 may offset their additional costs by the reduction in size and cost of power plant. The increased range that this allows the vessel for a given strength and speed makes the vessel more viable and the lifetime costs will also offset the additional initial cost.

Overall it appears that 1010, the material used for the Severn class hulls, is very suitable for its purpose.At the limits...

Any material has a point at which it will eventually fail and, although the point at which 1010 and 1020 fail is far beyond the loads imposed on even lifeboat hulls, examination of the way in which failures occur can point the way to even better materials.lt should be remembered that that the failures described hereoccurred on a test rig applying unnatural loads to a small strip of material! During a flexural test three different types of failure can occur: on the surface under tension; on the surface under compression; or internal shear .Internal shear is the main cause of failure in 1010 and 1020.. As two different resins were tested with one fibre lay-up, the problem would seem to lie in the fibres. As 97% of this failure type occurred in the same position in the lay-up the reason could be the fibre orientation or the angular fibre change between layers.

The fibre matrix sheared along its crystal boundary, leaving the fibre to take the strain and virtually no matrix adhered to the aramid (Kevlar-type) fibre. The glassfibre in the same failure area showed considerably greater fibre matrix adhesion. In one sample over 70% of the the aramid fibres had separated from the matrix, whereas the glass fibres had not.The surface of the aramid fibre is considerably smoother than the glass fibre and it seems that the fibres are mainly held by mechanical means and not by micro bonding.

The glass fibres are half the diameter of the aramid fibres which gives the glass a 50% greater surface area for the same volume. In the failure area the fibres run at 90° to the stress plane and it seems possible that the matrix fibre interface is sheared along its crystal boundary, releasing aramid fibres and allowing them to move and bunch into a bundle to form a crude wedge, which is then driven between the still-firm fibres increasing the number of loose fibres and the size of the wedge until delamination occurs.

94% of 1010 samples failed by delamination, and a rather lower percentage with 1020 but they have near identical flexural fatigue curves, although 1020 is 6% lower. As the only difference is the resin it can be concluded that 1010's resin is more elastic in fatigue than the one in 1020, which would make it a more suitable matrix resin for a hull subjected to fatigue in its working life.

It is interesting that despite its already considerable strength 1010 could be made even stronger. At present it is not reaching its full potential strength and if by changing the orientation of fibres and/or the use of a different coupling agent the delamination could be stopped - perhaps at the expense of its ultimate theoretical strength - the strength of the material could be further enhanced.

The movement of unidirectional fibres might be reduced by producing a weave with 80/20 instead of the 100/0. Although this would reduce the strength in one direction, when the fatigue is 90° to it the 80 fibres would be held by the 20 coming under tension, stopping them from moving and hence stop the delamination..