WO2002022440A1

WO2002022440A1 - Composite joints

Info

Publication number: WO2002022440A1
Application number: PCT/GB2001/003912
Authority: WO
Inventors: William Brooks
Original assignee: Bae Systems Plc
Priority date: 2000-09-14
Filing date: 2001-08-31
Publication date: 2002-03-21
Also published as: AU2001284224A1; GB0022589D0

Abstract

A joint suitable for use with composite materials, comprising two adherends (20,40) joined by a chemical bonding agent wherein the joint further comprises a relatively low elastic modulus material (30) bonded between the two adherends sufficient to provide a reduction in stress concentration across the joint between the adherends. The invention also includes a method for making such a joint.

Description

COMPOSITE JOINTS

The present invention relates to chemically bonded joints and a method of providing such joints. In particular, the invention relates to joints where the main loads generally carried by the two component adherends are born in non- parallel directions.

Composite structures often require chemical bonded joints in their construction. A common problem arising in the bonding of such composite joints is the occurrence of stress concentrations at joint boundaries when a load is applied to the bonded structure. In particular where one or other of the components comprises a composite material, such materials are often prone to crack propagation through at least one dimension and are thus particularly vulnerable to localised through-thickness or shear stresses which may initiate rapid crack propagation causing premature failure of the joint. Laminated composites in particular suffer from this weakness, such stress concentrations are known to give rise to localised shear and peeling stresses through the thickness of the laminated parts to be joined. One specific example where composite bonding gives particular problems is in the chemical bonding of rib to skin joints common in the manufacture of aircraft and seafaring vehicles.

Conventionally, where a rib is joined to a skin, loads are applied to the rib tending to pull the rib away from the skin causing consequential through- thickness stresses in the skin. These stresses concentrate about the rib to skin joint and once they reach a critical level cause fatal through-thickness crack propagation and premature failure of the rib-skin composite structure. It is understood that these problems are a consequence of the relatively high stiffness of the epoxy adhesives conventionally used in the joints.

Conventionally these problems have been addressed by attention to the stiffness and shape of the composite components to be chemically bonded, these characteristics being optimised between the components to minimise resulting stress concentrations at the join. Another approach in the case of laminate structures has been to employ through-thickness reinforcement by means of bolts, rivets, staples, pins or stitching. Such techniques add complexity and cost to the manufacturing process and disrupt, undesirably, the in-plane properties of the skin. In addition, these techniques are of limited application in thick laminated structures (of greater than about 4mm) where, for a given surface area and through-thickness strength, the loads required to be transmitted are considerably increased.

As well as the through-thickness problems of the skin, there can also be problems with high in-plane skin tension strains causing through-thickness failures in the rib, even without a rib pull-off load being applied to the rib. This problem is commonly encountered in relation to chord-wise ribs in a transport aircraft wingbox. Once such failure has occurred within the joint, the joint itself becomes prone to failure under relatively low out-of-plane loads.

Chemical bonded joints are, nonetheless, considered preferable to other forms of joint manufacture, especially in aircraft and ship building applications, due to their relative ease of manufacture and low cost. In aircraft and ship building applications, chemical bonded joints are particularly preferred because of their ability to act in a secondary function as a seal against fuel leakage or condensation and the removal of arcing problems which are commonly caused where mechanical fastening devices protrude into fuel carrying vesicles.

It is therefore desirable to provide a means for improving the performance of chemical bonded joints which enables more even load transmission through the joint.

In a first aspect the present invention provides a joint comprising two adherends joined by a chemical bonding agent characterised in that the joint further comprises a relatively low elastic modulus material such as rubber or an elastomer, bonded between the two adherends so as to provide a reduction in stress concentrations across the joint between the adherends.

In a joint according to the present invention, when a load is applied to the bonded composite, the layer of low elastic modulus material compensates for localised concentrations of stress by increased deformation in these localised areas, so redistributing the load within the joint more evenly. The occurrence of dangerously high stress concentrations in the joint is thereby minimised and the overall load carrying capability of the joint is increased. A further advantage of the arrangement occurs during impact. The joint of the present invention is able to deflect further on impact and requires higher energy of impact to destroy the joint, consequently minimising the risk of failure in these conditions.

Further advantages are associated with the insertion of a low modulus elastic layer in accordance with the present invention. The low elastic modulus layer is able to help absorb tolerances in the thickness and profiles of the joint adherends during assembly providing a more uniform distribution of stress across the interface and consequently a stronger bond. Where joints need to be disassembled for repair or inspection, the low modulus material can be cut more easily than a chemical bonding agent, thus enabling simpler and cheaper disassembly. Hysteresis within the low modulus layer may provide improved attenuation of sound and vibration through the structure enabling noise and vibration reduction, a particular advantage in aircraft and boat applications. Sensors such as fibre optics or strain gauges may be embedded in the elastomer layer to enable monitoring of the loads in the structure.

Improved performance can be achieved by tailoring the strength and stiffness of the low elastic modulus material to react to specific conditions of loading expected during normal use of the composite structure.

Optionally, the layer of low elastic modulus material may comprise an interface sheet to be inserted between the adherends. Alternatively, the layer may be integrated into the joint tooling. It is preferred that the layer of low elastic modulus material is relatively thick compared to the normal thickness of the layer of bonding agent necessary to join the same two adherends by conventional means.

Conveniently, the low modulus elastic layer may comprise a chemical bonding agent removing the need for an additional step of applying a separate adhesive. One particular example of such a material is the fluoro-elastomer Viton®. The joint of the present invention is particularly suited to designs where the main loads generally carried by the two component adherends are born in non-parallel directions.

Optionally at least one of the adherends may be a composite laminate. One of the adherends may be a structure to be stiffened and the other said adherend may be a stiffener for joining to the structure. To allow the stiffener to be joined to the structure to be stiffened it is preferred that the stiffener has a relatively large interface surface area to assist in achieving a satisfactory bond with the structure to be stiffened. This area is known as the footprint area. To form the footprint area the stiffener may split into two split portions and curve progressively outwardly until each split portion extends almost parallel with a surface of the structure to be stiffened. The split portions continue to extend outward as relatively flat profile sections to ends of the split portions. Preferably the layer of low elastic modulus material is between 0.1 and 2 times the average thickness of the relatively flat profile sections of the split portion of the stiffener. The low elastic modulus material needs to be of a sufficient thickness to spread the load from the rib evenly through the whole surface area of the interface with the skin of the component to be joined. The other consideration is weight, especially in aircraft applications and an optimum for the thickness for the low elastic modulus layer has to be found taking both factors into consideration. It has been found in practice that layers with a thickness of between 0.2 and 1 times the average thickness of the relatively flat profile sections of the rib split portions produce good results. It is believed that for many applications the optimum thickness is between 0.4 and 0.6 times the average thickness of the relatively flat profile sections. However, it must be remembered that the thickness of the low elastic modulus layer must be arrived at on a case by case basis.

The joint of the invention is particularly suited to skin to stiffener applications. It is especially suitable for use in aircraft and boats, especially aircraft wing skin to stiffener joints. The stiffener may be an "--"-section or a "T"- section. The joint is particularly suitable for applications where the thickness of the skin is greater than 3mm. This enables larger vehicles to be constructed with better joint stability and strength, overcoming limitations in conventional vehicle design.

A particularly preferred application of the present invention is in skin to stiffener joints in aircraft wings where the skin has a thickness greater than 3mm. In this application, the thickness of the low elastic modulus material is preferably greater than 0.5mm to provide the desired level of load redistribution.

Rubbers tend to go into a glassy state at very low temperatures. If the joint is to be subject to very low temperatures an elastomer formulated with a low temperature plasticiser may be used as the low elastic modulus layer. Where the composite structure forms part of a vehicle, preferred materials for the low modulus layer include nitrile rubbers and fluoro-elastomers such as Viton® which are resistant to fuels and corrosives which may be present in the working environment of the composite structure.

In a second aspect the present invention provides a method for bonding a joint using a chemical bonding agent including the steps of:

- aligning adherend surfaces of components to be joined in a desired position

- interposing a quantity of low elastic modulus material between the adherend surfaces, and - curing the bonding agent such that the low elastic modulus material is bonded between the adherend surfaces.

The low elastic modulus material may contain the chemical bonding agent.

The invention will now be further described by way of example only and with reference to the accompanying drawings of which:-

Figure 1 is an end view of a test piece used to verify the concept of the invention,

Figure 2 is a schematic view of a rib to skin joint with an interface layer between the rib flange and skin used as a two dimensional finite element model, Figure 3 is the finite element model of Figure 2 modified to show a part rubber and part stiff noodle,

Figure 4 illustrates the results from the two dimensional finite element model in Figure 2 run with quasi-isotropic carbon material properties simulating a joint according to the prior art,

Figure 4a is a key for Figure 4 defining the stress contours,

Figure 5 illustrates the results from the two dimensional finite element model in Figure 2 run with a joint according to the prior art,

Figure 5a is a key for Figure 5 defining the stress contours, Figure 6 illustrates the results from the two dimensional finite element model in Figure 2 run with an all rubber noodle and interface,

Figure 6a is a key for Figure 6 defining the stress contours,

Figure 7 illustrates the results from the two dimensional finite element model in Figure 3 run with a part rubber and part stiff noodle, Figure 7a is a key for Figure 7 defining the stress contours, and

Figure 8 shows the results of testing an elastomer interface rib-skin joint bonded together with various interfaces.

The skilled reader will understand that the invention is not limited to the embodiments described and many others will become apparent on reading the following examples.

To verify the concept of the invention a test piece was made as illustrated in Figure 1.

A pair of extruded aluminium alloy flanges 80, 81 were bonded back to back in the form of a "T" section 10. This "T" section was in turn bonded to a Carbon Fibre Reinforced Plastic Non Crimp Fabric (CFRP NCF) laminate 20 of 12m thickness, using Redux 420™ adhesive. A low elastic modulus layer 30 of Viton® rubber (70 Shore hardness), 2mm thick, was interposed at an interface between the flanges 80, 81 and the laminate 20. The aluminium surface 11 was then anodised. All surfaces were degreased before bonding together using light pressure with G clamps.

The joint specimen was mounted in a test rig (not shown) such that the load was reacted over a span of 120mm. The specimen, 25mm wide, was loaded in an Instron™ testing machine until failure occurred. Load vs. Time at a steady displacement was measured and recorded.

At a running load of approximately 150N/mm, the specimen failed as a result of the Viton® rubber 30 disbonding from the aluminium "T" section 10. Neither the rubber nor a surface 21 of the laminate 20 showed any signs of failure or crack propagation.

Testing of the specimen showed that 150N/mm can be transmitted into the 12mm thick CFRP NCF laminate skin surface 21 without through-thickness failure of the laminate 20. Flanges made of aluminium alloy were used to isolate the region of interest, that is the composite skin/rubber interface. Given the superior bonding capability associated with a composite such as the CFRP NCF laminate over that of aluminium, the potential maximum load endurable by a composite to composite bond could be further increased.

The test specimen was 25mm wide with flanges of 55mm span, giving an area of 1375 mm². The maximum stress that could be transmitted through the rubber was 20 N/mm². If this load could be uniformly distributed, the load transferred would be 20 x 1375 = 27500 N or 27500/25 = 1100N/mm run. The FE model, based on a running load of 200N/mm, showed a peak through- thickness stress of 3 N/mm² which is easily within the capability of the rubber, adhesive and laminate inter-laminar strength. A 2D finite element model has been run of a typical 12mm thick CFRP

NCF laminate skin rib to skin joint, loaded by a 200 N/mm run out of plane of the rib combined with bending of the skin panel. This simulates a typical emergency landing fuel pressure load case in a transport aircraft wing box.

The model illustrated in Figure 2 comprises a rib 40 to skin 20 joint with an interface layer 30 of 2mm thickness between the rib flange 10 and the skin

20. To allow the rib 40 to be joined to another component a relatively large footprint area 41 is required to provide a suitable interface surface to achieve a satisfactory bond with the skin 20. To form the footprint area 41 , the rib splits at point 42 into two split portions 43, 44 and curves progressively outwardly until each split portion 43, 44 extends almost parallel with the footprint area 41. Curved transition sections 45, 46 extend from the split point 42 to points 53, 54. From here the split portions 43, 44 continue to extend outwards with a relatively flat profile in sections 47, 48 to ends 51, 52 of the split portions 43, 44 to cover the footprint area 41. As the rib splits into portions 43, 44 a void is created therebetween. The volume of the void formed between the rib split portions 43, 44 and the footprint area 41 on the surface 21 is known as the noodle zone and is infilled, conventionally with carbon. The infill is known as the noodle 50. The noodle is stitched or stapled 55 to the rib to prevent through-thickness splits.

The model was run with: a) Quasi-isotropic carbon material properties simulating an all rigid noodle 50 and an interface layer of adhesive 30; b) Rigid noodle 50 with rubber interface layer 30; c) All rubber noodle 50 and rubber interface layer 30; Finally, the model was run as illustrated in Figure 3 with: d) Part rubber 60 and part stiff noodle 70. The resultant stress concentrations are illustrated in Figures 4 to 7 in which;

Figure 4, illustrates the stress distribution foreseen for case a)

Figure 5, illustrates the stress distribution foreseen for case b)

Figure 6, illustrates the stress distribution foreseen for case c) Figure 7, illustrates the stress distribution foreseen for case d)

In Figure 4 (case a)), the peak in stress of 30 N/mm² occurs at the edges of the rib flanges 10 which, in use, could potentially lead to the start of through- thickness cracks which would cause immediate joint failure. Figure 5 (case b)), shows how effectively the rubber interface layer 30 accommodates the differential strains in the rib and skin areas, producing a nearly uniform stress distribution of approximately 3 N/mm².

In Figure 6 (case c)), the all-flexible noodle and interface shows the same uniform through-skin-thickness stress distribution, with bending of the rib flanges.

In Figure 7 (case d)), the part flexible and part stiff noodle interface shows the same uniform through-skin-thickness stress distribution, with bending of the rib flanges. Further testing was carried out with elastomer interface rib/skin joints (not shown).

A joint sample was designed and made using a pre-infused Non Crimp Fabric Resin Film Infusion (NCF RFI) carbon fibre rib. The joint incorporated the following features: i) A large rib foot flange radius with a noodle infill of similar stiffness to that of the rest of the rib, so that the joint pull-off load is taken by the flange areas in bending and distributed over a wide area. ii) Through-thickness reinforcement (e.g. using staples) to prevent interlaminar splitting of the rib flange. iii) An elastomeric interface layer between the rib flange and the skin to distribute the load from the rib into the skin as evenly as possible.

This is currently thought to be design best practice for rib/skin joints.

The carbon fibre rib was laid up using triaxial Non Crimp Fabric (NCF) with the unidirectional fibres going upwards (i.e. aligned with the pull-off load). Alternate layers of left and right handed fabric were laid up to give a balanced lay-up.

To spread the rib pull-off load over a large area of composite, a large root radius of 20mm was used. A stiffness-matched noodle of carbon fibre NCF was laid up and inserted in the void of the rib and stapled to the rib to prevent through-thickness splitting of the rib.

25mm wide 10mm thick strips of NCF RFI pre-cured skin and pre-cured rib flange were successfully bonded together using Redux 420 with the following interfaces:

1) Plain Redux™ bonded under G clamp pressure,

2) 2mm thick Viton™ fluorocarbon elastomer,

3) 1.5mm thick Nitrile rubber (cured),

4) 6mm thick Nitrile rubber (cured). Figure 8 shows the results of these tests. The basic joint design of the invention, with a Redux 420 bond, already produces a good joint strength (270 N/mm width) compared with the prior art, (e.g. see "The design of the spar- wingskin joint" Ralph D. Copes, R. Byron Pipes, Centre for Composite Materials, University of Delaware, 1978), which gives a typical load of around 100 N/mm width. The joint failed 1 ply down below the bondline, in the composite skin.

However, the inclusion of elastomer interface between the rib flange and skin has been shown to boost the strength considerably by reducing the stress concentrations in the joint interface between the rib flange and the skin. The best sample, using a 1.5mm thick nitrile rubber layer, failed at 384 N/mm of width, two plies down into the composite skin. There was no failure of the rubber, and the skin failed before the joint. The stiffness of the joint was not very much reduced compared to the straight Redux bonded version. The strength was 40% higher and the energy absorbed in failing the joint was approximately 3 times higher.

Bonding occurs between the low elastic modulus layer and the first ply of the adherend in contact therewith. It is this first ply thickness that is important.

The tests were carried out with a composite having a ply thickness of approximately 0.8mm. All three low modulus layers tested increased the strength of the joint considerably. This shows that low modulus layers between 1.8 times and 8 times the ply thickness can considerably increase the joint strength. For optimum joint design, joint weight must also be considered and there is a trade off between low elastic modulus layer thickness and joint weight.

Claims

1. A joint comprising two adherends joined by a chemical bonding agent characterised in that the joint further comprises a relatively low elastic modulus material bonded between the two adherends sufficient to provide a reduction in stress concentrations across the joint between the adherends.

2. A joint as claimed in claim 1 characterised in that the layer of low elastic modulus material is thicker than the layer of bonding agent necessary to join the two adherends by conventional means.

3. A joint as claimed in claim 1 or claim 2 characterised in that the chemical bonding agent is provided by the low elastic modulus material.

4. A joint as claimed in any preceding claim characterised in that the main loads generally carried by the two component adherends in use are born in non parallel directions.

5. A joint as claimed in any preceding claim characterised in that at least one of the adherends is a composite laminate.

6. A joint as claimed in any preceding claim in which one said adherend comprises a structure to be stiffened and the other said adherend comprises a stiffener for joining to the structure.

7. A joint as claimed in claim 6 wherein a portion of the stiffener upstanding in use from the structure to be stiffened, is split into two portions to co-extend with a surface of the structure to which the stiffener is attached, each of the two split portions having a relatively flat profile section and the thickness of the layer of low elastic modulus material being between 0.1 and 2 times the average thickness of the relatively flat profile sections.

8. A joint as claimed in claim 7 wherein a portion of the stiffener upstanding in use from the structure to be stiffened, is split into two portions, to co-extend with a surface of the structure to which the stiffener is attached, each of the two split portions having a relatively flat profile section and the thickness of the layer of low modulus material being between 0.2 and 1 times the average thickness of the relatively flat profile sections.

9. A joint as claimed in claim 7 wherein a portion of the stiffener upstanding in use from the structure to be stiffened, is split into two portions, to co-extend with a surface of the structure to which the stiffener is attached, each of the two split portions having a relatively flat profile section and the thickness of the layer of low modulus material being between 0.25 and 0.75 times the average thickness of the relatively flat profile sections.

10. A joint as claimed in claim 7 wherein a portion of the stiffener upstanding in use from the structure to be stiffened, is split into two portions, to co-extend with a surface of the structure to which the stiffener is attached, each of the two split portions having a relatively flat profile section and the thickness of the layer of low modulus material being between 0.4 and 0.6 times the average thickness of the relatively flat profile sections.

11. A joint as claimed in any of claims 6 - 10 in which said structure to be stiffened comprises a skin.

12. A joint as claimed in claim 11 characterised in that the skin is greater than 3mm thick.

13. A joint as claimed in claim 12 characterised in that the layer of low elastic modulus material is greater than 0.5 mm thick.

14. A joint as claimed in any of claims 11 - 13 characterised in that the skin comprises an aircraft wing skin.

15. A joint as claimed in any of claims 6 - 14 in which the stiffener is an "L"- section.

16. A joint as claimed in any of claims 6 - 14 in which the stiffener is a "T"- section.

17. A joint as claimed in any preceding claim characterised in that the low elastic modulus material is a nitrile rubber.

18. A joint as claimed in any of claims 1 to 16 characterised in that the low elastic modulus material is an elastomer.

19. A joint as claimed in claim 18 characterised in that the elastomer is a fluoro-elastomer.

20. A joint according to claim 18 or 19 wherein the elastomer is formulated with a low temperature plasticiser.

21. A joint as claimed in any preceding claim wherein the low elastic modulus material is resistant to corrosion in its surrounding environment.

22. A method for bonding a joint using a chemical bonding agent including the steps of: aligning adherend surfaces of components to be joined in a desired position interposing a quantity of low elastic modulus material between the adherend surfaces, and curing the bonding agent such that the low elastic modulus material is bonded between the adherend surfaces.

23. A method as claimed in claim 21, characterised by combining the chemical bonding agent with the low elastic modulus material.