LIGHTWEIGHT CONSTRUCTION FOR ADVANCED ... - CiteSeerX

60 downloads 268939 Views 2MB Size Report
construction have been the focus of further research and development. ... The paper first introduces conventional and more advanced types of ...... integrated into the hoistable deck itself or be located in the ship side between the web frames ...
LIGHTWEIGHT CONSTRUCTION FOR ADVANCED SHIPBUILDING - RECENT DEVELOPMENT P. Noury*, B. Hayman, D. McGeorge and J. Weitzenböck In this paper some recent advances in lightweight construction for modern shipbuilding are presented. Following a description of lightweight structures and their current use in shipbuilding, emphasis is put on naval applications of composites, and in particular on aspects related to cost benefit assessment and survivability. The following part reports on the use of adhesive bonding for superstructures of high-speed craft and passenger ships. It focuses on material selection, design and analysis, manufacture and application. Finally a recent application of laser-welded steel sandwich panels in hoistable car decks is reviewed. The content of the paper is based on the results of three ongoing research projects aimed at developing lightweight solutions for the shipbuilding industry the BONDSHIP and the SANDWICH European co-operative research projects, part of the 5th Framework, GROWTH, and the EUCLID RTP3.21 project, a European military research project. INTRODUCTION During the past 20-30 years there have been a great many developments in lightweight construction. During this period the use of aluminium alloys and fibre reinforced polymer composites in the ship building industry has steadily increased. Civilian applications include not only high speed vessels but also significant parts of the superstructures of large passenger ships. Military applications include mine hunters, fast patrol craft and superstructures of larger naval ships. Recently, pioneering lightweight solutions involving such features as adhesive bonding and novel types of sandwich construction have been the focus of further research and development. Some of these developments are based on innovative use of steel as well as the more usual lightweight materials. The areas of application have been extended to include more components of conventional ships, such as moveable car decks and ramps. Generally the aims are to improve safety and reliability in addition to saving weight and increasing efficiency of fabrication and maintenance. The paper first introduces conventional and more advanced types of lightweight structures used in the ship building industry. Recent developments in some ongoing European research projects are then presented and discussed. The topics addressed are naval applications of composites (with aspects related to cost benefit assessment and survivability of naval structures), the use of adhesive bonding in superstructures of high-speed craft and passenger ships (in particular material selection, design and analysis, and application cases), and finally the application of laser-welded steel sandwich structures for hoistable car decks. LIGHTWEIGHT STRUCTURES AND APPLICATIONS Why lightweight structures? Reasons for using lightweight materials and structural arrangements in ships (as in many other types of transportation vehicles) include the following: •

They allow a greater payload for a given size or weight of vessel.



They allow higher speeds to be achieved.



They reduce fuel consumption and environmental emissions for a given payload and distance travelled. 1

DET NORSKE VERITAS. * contact author’s address: UCTNO910 – Transport Systems Program, DET NORSKE VERITAS, Veritasveien 1, N-1322 Høvik, Norway

For ships with many decks (such as cruise vessels) the use of lightweight solutions in the upper decks helps to lower the centre of gravity, thus improving stability and permitting larger height/breadth ratios. In addition, some lightweight solutions (e.g. closed aluminium extrusions and sandwich configurations) are also compact, giving reduced space requirements and leading to smaller overall vertical distances between decks. The main lightweight materials The main lightweight materials used in ships are fibre reinforced plastic (FRP) composites and aluminium alloys. FRP is used in both single-skin and sandwich configurations; in single-skin applications there is usually a system of stiffeners as illustrated in Fig. 1, but unstiffened monocoque solutions are also to be found. (Here we use the term “FRP” to include the special case of glassreinforced plastics – GRP.) Aluminium alloys are commonly found in welded, stiffened plate configurations and in the form of extruded sections (both open and closed), but sandwich arrangements are also possible. Although not normally considered to be lightweight materials, high strength steels may also be used to reduce weight; these are to be found in stiffened plate and, recently, some sandwich configurations. There is increasing use of mixed solutions in which various materials are combined in one ship or superstructure. Current and potential applications of lightweight materials in ships are mainly related to high speed passenger and car ferries, patrol and rescue craft, smaller naval ships (e.g. mine countermeasure vessels), pleasure craft and sailing yachts. However, they are also used in superstructures of cruise ships and of larger naval ships (e.g. frigates). Furthermore they are used extensively in secondary structures and components for all types of ships, from masts and casings to moveable vehicle ramps and decks. As a rough guide, in the main hull structure, FRP is used for craft with length up to about 50 m, and aluminium for vessels up to about 120 m, while high-strength steel is mainly used for larger vessels. In recent years composites have been rarely used in hulls of new ferries; their use has been mainly confined to patrol/rescue craft, pleasure craft, yachts and naval vessels. The main reason for this is the severe restriction on the use of combustible materials that was introduced in the IMO Code of Safety for High Speed Craft in the 1990s. With new, approved fire protection systems now available, FRP has once more become a viable and safe alternative to aluminium for ferry applications. References 1 and 2 provide a good introduction to marine FRP composites and their applications. Much useful information about marine applications of aluminium alloys is provided in refs. 3 and 4. Materials and manufacturing methods: FRP composites FRP composites for marine applications are generally laminated composites. These consist of several layers of reinforcement fabric in a polymer resin matrix. In the case of sandwich construction there are two skin laminates with a core between that keeps the laminates in place and provides a shear connection between them. In reinforcement fabrics the main fibre materials are glass (E-glass, R-glass, S-glass), aramids (Kevlar, Twaron), carbon, polyester, high performance polyethylene (HPPE), and various hybrids/combinations. Various fabric formats are used, including chopped strand mat, continuous strand mat (randomly placed fibres), woven roving (plain weave, satin, twill), uniaxial, and “multiaxial” configurations (layers of uniaxials stitched together, also known as non-crimp fabrics). Matrix materials (resins) include polyesters, vinylesters, epoxies, various modifications of these resins (e.g. rubber-modified vinylester), and phenolics. In a pre-preg, layers of fabric are pre-impregnated with partially cured resin that will completely cure on being heated. Sandwich core materials (Fig. 2) include polymer foams, end-grain balsa wood, honeycombs and corrugated FRP cores. The main foams in use are PVC (various types), polymethacrylimide (PMI), 2

polyetherimide (PEI) and phenolic. Honeycombs are normally metal (mainly aluminium) or aramid paper (“Nomex”). Regarding manufacturing methods, the most common contact moulding method is hand lay-up, but automated lay-up systems are used for larger production, and spray lay-up in some cases when short fibres are being used. Greater compaction can be achieved by applying pressure, e.g. by vacuum bagging, or (for relatively small items) an autoclave. High fibre content can also be achieved by resin transfer moulding (RTM) or by vacuum-assisted resin infusion (e.g. the SCRIMP process). These compression and injection/infusion methods have also the advantage of being closed processes, so that emission of solvent vapours is much reduced. For cylindrical and similar shapes, filament winding is a convenient and economical production method. To manufacture sandwich structures it is possible to produce the skin laminates by one of the above processes and then bond these to the core. However, a common practice for hulls is to build up the core on a simple wooden framework and laminate directly onto the exposed side of the core, then remove the framework and laminate on the other side. If resin infusion techniques are used, another option is to infuse sandwich panels with face laminates on both sides in one step, and then assemble the panels to build a larger structure. In some cases it is even possible to infuse an entire sandwich structure or component in one step. Advantages and disadvantages of FRP An important advantage of FRP materials is that their properties can be tailored to meet the requirements of the structural application. If properly designed, FRP structures provide good strength for low weight, especially when optimised uniaxial or multiaxial fabrics are used. Also, FRP materials are readily formed into complex shapes, though it may be difficult to control the fibre directions in some cases. FRP suffers little or no corrosion if used properly. Such materials are virtually maintenance-free, giving low running costs. Stress concentrations are less critical than with metals, provided continuous fibre reinforcements are used. Hence fatigue cracking is less of a problem. Most FRP materials also have a low conductivity so that effects of fire can be more easily contained than with metal structures. Advantages for military applications include non-magnetic properties (needed where mines are a hazard) and transparency to electromagnetic waves (except with carbon reinforcements). Carbon reinforced plastics can have good absorption properties with regard to electromagnetic waves, giving good stealth properties. Advanced non-structural features such as sensors can also be readily built into FRP composites. It is also possible to provide better shock resistance with FRP than with timber, which was previously used for mine counter-measure vessels. Additional advantages of FRP sandwich include very good flexural stiffness and strength for low weight, a high margin against catastrophic failure or penetration because of the two skins, high buoyancy, good built-in thermal insulation, and the ability to build both large and small structures without costly moulds. Sandwich structures generally allow the lowest level of stiffeners to be dispensed with, giving smooth surfaces and a compact structure. Furthermore, the skin laminates are used optimally so that relatively expensive skin materials can be used without undue cost penalty. Disadvantages of FRP include high initial cost (except for smaller applications and items produced in large series), and in many cases a need for adequate fire protection, low elastic modulus, and low through-the-thickness strength (which can also make design of connections a complex issue). Aluminium alloys Aluminium alloys for use in marine applications are normally of the 5xxx series (with magnesium as alloying element) or, for locations such as decks that are not in direct, continuous contact with sea water, the 6xxx series (with magnesium and silicon). 3

Plates are normally strain hardened (cold worked), giving an “H” temper designation. Stiffeners and deck planks are generally extruded. Wrought aluminium alloys have a high strength/weight ratio compared to steel. Aluminium alloy structures generally have higher stiffness than corresponding FRP structures. Many grades are weldable, and many are also extrudable. Aluminium alloys, if chosen correctly, have a good corrosion resistance. The main disadvantage of aluminium is the severe softening of the heat affected zone (HAZ) that occurs during welding. This reduces both the static strength and the fatigue life. Aluminium alloys are also relatively expensive to weld. The combination of a high required heat input and a high coefficient of expansion leads to large distortions and shrinkage during welding. If the wrong alloy is selected for a given purpose, corrosion resistance may be poor. Aluminium alloys have a fairly low melting point (around 650ºC) and they soften at temperatures considerably below this, so that aluminium structures generally require extensive fire insulation. The tearing resistance of aluminium is relatively poor; however, this disadvantage is not reflected in the required scantlings of ships because the prescribed damaged stability cases take no account of the difference in tearing resistance of the various materials. NAVAL APPLICATIONS OF COMPOSITES Steel is the traditional material for naval ships as it is for civilian ships. This is due to the low cost and high strength of steel and because steel construction is well known to navies and naval shipyards. However, sometimes composite materials are chosen because they can offer properties that are particularly attractive for a particular application. For example, composites are used in mine countermeasure vessels due to the nonmagnetic properties of the composite and the superstructures of larger naval ships and in fast patrol craft due to the low specific weight and the possibilities for reduced signatures (stealth) offered by composite constructions (Fig. 3). Recently, it has been decided to adopt composite superstructures in future US Navy destroyers and cruisers, i.e. DD21 and CG21 (ref. 5). Hence, the composites are used because they are different from steel and offer advantages when used wisely in specific cases. Wise use of composites in naval ships requires that all differences be adequately accounted for. This implies that the detailed technical requirements to the composite structure should be different from that of a corresponding steel structure. Otherwise, an unbalanced composite design will be obtained, and it might even be impossible to take advantage of the desirable properties of the composites as explained in the next subsection. Therefore, a cost benefit assessment method has been developed that allows deriving a balanced set of requirements for a composite structure. This method is outlined in the subsequent subsection (ref. 6). The potential for improved resistance of traditional composite constructions is explored thereafter. Some directions for future research are outlined later in the paper. Introduction Composite materials are quite different from traditional materials and some of these differences represent advantages whereas others represent challenges to be overcome. The main differences are summarised in the TABLE 1. Many regulations provide prescriptive requirements that reflect what can be achieved with a steel structure. These requirements have evolved over time such that a balanced steel design is obtained when complying with the requirements. An example from civilian ships is the requirement in the SOLAS convention (ref. 7) that structural materials in many areas shall be non-combustible. This is of course always complied with if steel is used. However, it effectively rules out the use of composites although one may develop composite designs that provide fire safety that is similar to or better than accepted steel designs. This has recently been rectified by new regulations allowing the use of 4

solutions that are demonstrated to be equivalent to the traditionally accepted ones by a risk assessment (ref. 8). TABLE 1 – Summary of differences between steel and composites construction Property

Steel construction

Composite construction

Weight

High

Allows significant reduction in structural weight

Corrosion

Rusts in marine environment resulting in high maintenance cost

Very durable in marine environment, little maintenance

Combustibility

Non-combustible, will not contribute to fire or generate toxic fumes

Combustible, surface must be protected in fire hazard areas

Thermal conductivity

High, must be insulated to prevent fire propagation and to control infrared signature

Low, inherent insulation more than sufficient

Electrical conductivity

High, inherently provides electromagnetic shielding

Low, must embed conductive layer if electromagnetic shielding is needed

In a similar fashion, it is sometimes argued that composite designs for naval ships, in order to be equivalent to a traditional steel design, should meet the same requirements as steel designs to survive weapon-induced loads (such as internal and external blast). The hope would be to meet all the requirements that can be met with a steel structure and at the same time obtain all the advantages of the composite design such as reduced weight, reduced signatures etc. However, that is unrealistic. Direct application of current steel structure requirements to a composite structure would lead to a design that is "better" than the steel design in the sense that it fulfils all the requirements to a steel structure, some of them with a great extra margin. However, to meet all these requirements, weight and costs may have had to be added to the extent that the composite structure is no longer competitive. Furthermore, the composite design will be an unbalanced one. The "steel" requirements have evolved from in-depth knowledge of the potential and limitations of steel construction. Hence, a steel design can meet all the requirements, none of them with an excessive margin, if practical design solutions based on good engineering practice are adopted. Due to the different behaviour of composites, some of the requirements can only be met with great effort whereas others are met with an excessive margin. Therefore one could more cost effectively obtain a good design if a more balanced set of requirements was used. These requirements would have to reflect what can reasonably be achieved with a composite design. In order to compare a balanced composite design to a balanced steel design, it would be necessary to trade off such aspects as build costs, weight, signatures, speed, manoeuvrability, crew needs maintenance and repair costs etc. What is best depends on the capabilities required for the particular ship. For small fast craft, an all composite structure may be best. For a larger ship an all steel structure may be best, but a steel hull with a composite superstructure may be better due to reduced signatures and the possibility to locate sensors at a higher position without compromising stability. The above discussion highlights the need for a rational approach to 1) establish a balanced composite design and 2) compare that design to a steel design in order to make an informed selection. This is addressed in the next section.

5

Cost benefit assessment of naval structures (ref. 6) The goal of cost benefit assessment is to provide a means of selecting the most cost effective design options for improving structural performance. The starting point is a base design that reflects current state of the art and complies with existing safety requirements. The assessment of design improvement options considers the two aspects of survivability and cost. A measure of the structural performance of each of the design options is used to rank them and select the one that performs best. Survivability assessment A probabilistic approach is recommended for the assessment of survivability of the structure considering relevant combat threat scenarios (ref. 6). These threat scenarios should reflect the range of threat types and threat levels considered of relevance for the particular ship. They should be specified by a team with appropriate experts. The measure Si of survivability with respect to a certain specified threat of type i is defined as the probability of surviving an attack by that threat type. The measure si of susceptibility with respect to a certain specified threat of type i is similarly defined as the probability of not being hit if attacked by that threat. Rather than speaking of the event that the ship survives an attack by a certain threat i, it is useful to speak of the event of structural failure given a hit by that threat. The vulnerability measure Vi is defined as the probability of this event. With these definitions, it is possible to derive the following relationship (1) between these measures (detailed derivation is given in ref. 6):

S i = 1 − Vi s i (1) If a number of threats are considered, each with a specified attack likelihood Ai, the overall survivability may be defined as: S=

n

∑AS i

i

(2)

i =1

Although the scope of this article is limited to the structure of the ship, the above considerations are quite general in nature and may have a broader application. In the UK, the Royal Navy uses a similar approach, but also includes recoverability (the possibility of reinstating the ship’s functions after damage by appropriate damage control and repair activities) in survivability assessments that cover the whole ship system ( refs. 9, 10 and 11). However, since focus in this article is placed on the survivability of the structure, the following issues are required to estimate the vulnerability and thus the survivability of the structure: •

Methods to establish the load effects resulting from combat actions such as internal and external blast and load effects from underwater explosions.



Methods to establish the resistance of structural joints to weapon-induced load effects.

Another issue of particular concern is that of fire safety and survivability. As this is outside the scope of this article, it suffices to mention that the different behaviour of steel and composites implies that fire fighting strategies established for steel structures must be slightly changed to be effective in a composite structure: in a composite structure there is no need for or use of boundary cooling (a key element in fire fighting in a steel ship) and forced access to fire areas is required (ref. 12). Potter (ref. 13) concluded that an adequately protected composite structure for the helicopter hangar of a US Navy destroyer performs adequately in fire. Cost assessment

The cost assessment should consider relevant cost elements for the base design and cost savings and penalties of adopting promising design improvement options. The basic cost elements are: •

Manufacturing costs including costs of material, tools and labour. 6



In-service costs of inspection, maintenance, repair, modification, replacement and crew.

Furthermore, other differences between design solutions may in specific cases be translated into cost savings or penalties, and hence, directly included in the analysis. This may for example include benefits of increased speed or improved manoeuvrability, benefit of weight reduction or benefits of incorporating certain multifunctional properties into the structure. Selection of design performance improvement options

The starting point is an estimate of the total cost C0 and the survivability S0 of a naval ship using the base design solution. The structural performance P0 of this base design can be computed as: P0 =

S0

C0

(3)

The cost saving or penalty of implementing a structural performance improvement option is denoted ∆C1. A negative number indicates a cost saving. The ∆C1 estimate includes the elements outlined in the previous section. The survivability S1 is then estimated in the same manner as that for the base design as outlined before. The measure of the structural performance P1 of the improved design can then be computed as: P1 =

S1 (4) C 0 + ∆C1

The same exercise can then be carried out for other promising structural performance improvement options. Each option for which Pi > P0 will increase the performance of the structural design compared to that of the base design. If all promising means of cost reduction and vulnerability reduction are explored, application of the cost benefit assessment will lead to a balanced composite design that reflects what can reasonably be achieved with composites. In principle, the base design may be a traditional steel design and option 1 a balanced composite design. Hence, the cost benefit assessment method may be used to rationally decide whether a steel design or a composite design is best for a particular purpose. Survivable composite structures To obtain a survivable composite structure, the following issues are of primary concern: •

Prediction of the load effects resulting from weapon actions.



Prediction of the resistance of the structural elements to these load effects.



Understanding the behaviour of the critical structural elements and identifying how their resistance to the weapon induced load effects can be increased.

Weapon induced loads and load effects

The most relevant weapon induced loads are internal and external blast and effects of underwater explosions. Simple methods are available for predicting the response of traditional steel structures to blast loading (ref. 14). However, no such methods applicable to composite sandwich structures are known to the authors. Therefore, more sophisticated finite element simulations or blast trials are needed to establish the load effects due to blast in a composite sandwich structure. The response to underwater shock, whether steel or composites are used, is complicated and detailed assessment requires dynamic finite element analyses of the full ship structure.

7

Traditional designs and possible improvements

Even without numerical results, it is possible to identify some key challenges for typical naval composite structures. In this section metal composite joints for joining a composite superstructure to a steel hull and composite sandwich T-joints will be discussed. A traditional T-joint is shown in Fig. 4. This type of joint has a record of successful service experience in civilian ships. It is strong in compression and longitudinal and transverse shear, and even in tension the strength may be sufficient if not very thick or strong cores are used in the base panel (ref. 15). Its main weakness lies in its behaviour in bending caused by transverse loading on the attached panel: an edge bending moment can only be resisted by compressive stresses in one of the sandwich face sheets and tensile stresses in the other. The latter result in peel stresses that will tear the overlaminate off the base panel thereby causing fracture of the joint at relatively modest transverse loading. This is not much of a problem in civilian applications where such bending is avoided. However, in a naval vessel, this is precisely the kind of load effect one should expect from an internal blast event. Hence, one should expect that a sandwich structure with such T-joints would be very vulnerable to internal blast loading. The strength of the joint is strongly dependent on the detailed design at the edge of the perpendicular panel. Such issues as rounding of the panel edge, radius of the fillet and lay-up of the overlaminate have significant influence on the joint strength. Improvements can be made along two paths: •

Make the connection to the base panel stronger. An example is the inclusion of triangular fillets (ref. 13).



Make the connection rotationally flexible such that little or no bending will result from the transverse blast pressure.

Traditional metal composite joints such as the one illustrated in Fig. 5 have been studied in the past (ref. 16). Many variants of this type of joint are possible (refs. 17 and 18). Blast loading will produce transverse bending and shear across the joint. An underwater explosion will cause an initial compressive shock loading on the joint followed by a phase with tensile loading. These axial loads will be associated with bending due to the asymmetry of the joint. The traditional joint (Fig. 4) possesses the following characteristics: •

All important load effect components (tension, compression, transverse shear and bending) cause high stress concentrations and peeling at the edge of the steel insert. These stresses are likely to be the cause of failure. Furthermore, the contribution to these stresses from all the modes of loading implies that joint failure is governed by the combination of all these load effects.



Upon fracture of the bonded joint, the integrity of the structure is lost.

This suggests that more survivable joints may be developed where less peel is generated at the critical spot and where some sort of mechanical interlocking would prevent the superstructure from detaching completely from the hull in the event that the bond fails. The previous section showed that some weapon-induced loads, internal blast in particular, challenge the well known weakness of composite construction: the out of plane properties. These properties are highly dependent on the manufacturing conditions. This implies that full control of the manufacturing is even more important for naval composite ship structures than for civilian ship structures. A few specific issues are worthy of mentioning: •

The surface preparation of a composite surface prior to joining determines the resistance to peel stresses occurring in the joint. It is recommended to sand the surfaces completely prior to joining. To allow sanding without damaging the load bearing reinforcement, it is recommended to have a chopped strand mat surface layer of at least 300 g/m2. Substituting sanding with the use of a peel ply is not recommended unless acceptable performance is demonstrated by testing of representative joints made in shipyard conditions. 8



The surface preparation of the steel surface prior to joining determines the resistance and durability of the metal composite joint. A key issue is the implementation of a procedure that can be followed and controlled in an industrial shipyard environment. It is recommended to grit-blast the steel surface immediately followed by priming with a suitable primer and then leave to cure. This primer should be sanded immediately prior to bonding of the joint. To provide for sanding, one may incorporate a glass fire chopped strand mat within the primer layer.

Failure to comply with this may not be a problem in a civilian application where severe weapon induced loads are not anticipated. However, in a naval ship, such failure will inevitably compromise the survivability of the structure. USE OF ADHESIVE BONDING ON SUPERSTRUCTURES OF HIGH-SPEED CRAFT AND PASSENGER SHIPS Background Adhesive bonding is receiving considerable attention from the shipbuilding industry as a new joining method for lightweight materials. Aluminium alloys and composites are commonly used in superstructures of passenger ships and high-speed craft; smaller vessels are built completely out of these materials. The aim is to save weight but also to make production more efficient. Current joining methods such as fusion welding of aluminium or riveting or bolting are quite expensive and do not always give satisfactory long-term performance. Some of the attractions of adhesive bonding are that it does not change the properties of the base material, it permits the joining of dissimilar materials, it does not require hotwork with all the associated problems of heat distortion and fire hazard, and it gives great flexibility in the building schedule as one can even bond next to heat sensitive equipment or cables. There are already some applications of adhesive bonding in shipbuilding. Large window surfaces are common design features of modern passenger ferries and cruise ships. These windows are usually bonded with elastic adhesives. This minimises stresses in the glass. Moreover, misalignments of the steel structure can be evened out by varying the adhesive layer thickness. Wacker (ref. 19) gives a good account of bonding of windows in ships. Another recent application of adhesives is the bonding of seat rails to internal decks in fast passenger ferries (ref. 20). The motivation here was to use thinner and lighter deck scantlings below the minimum thickness recommended for welding. Reavey (ref. 21) reports on the use of adhesive bonding in hovercraft. These craft were built using aircraft bonding technology and design practices. The author concludes that the durability of the bonded aluminium joints was excellent , but the joints were expensive. Material selection Selection of a suitable adhesive for the planned application requires careful consideration of the geometry, loading condition and manufacturing requirements the joint is expected to experience. The substrates or adherent and also in some cases the surface finishes (paints, primers, etc.) are usually predetermined by the application case. These materials and surfaces can be determined using a standard questionnaire. The results of this survey can be used to also compile a list of requirements for selecting the adhesives. Based on these requirements adhesive suppliers may be asked to recommend suitable adhesives (Standard lists are available to help define requirements for the joint, see for example Espie et al. (ref. 22)). The list of requirements should include: •

Materials / surfaces of joint



Environment in production and in service



Geometry and load of joint



Application of adhesive



Temperature in service



Curing conditions

9



Required lifetime of joint

It is quite likely that at this stage relatively little is known about the strength requirements of the joint as one has no or only a vague idea about the detailed joint geometry and loading. Hence the adhesive selection is mainly based on general factors such as the anticipated service environment and manufacturing requirements. Next screening tests are carried out to select materials for use in the joint design process. The aim of the screening test programme is to reduce the vast number of possible combinations of adhesives, primers, paints and other surface preparations available for different adherent materials such as steel, aluminium and composites. Tests are chosen to obtain relevant test data in a cost and time efficient manner. Common to all tests is that the specimens are simple and cheap to produce. Typical tests used here are the bead peel test (only for flexible adhesives), lap shear test and the Boeing wedge test. The ageing conditions for the tests are determined by the in-service environments identified in the list of requirements. Design and analysis of joints Successful design of adhesively bonded joints requires an understanding of the mechanics of the joints, knowledge of the geometry and materials of the parts to be joined, the service loads, the environmental conditions and the requirements that exist for inspection of the finished joint (ref. 23). With this information one can design the local geometry of the joint and decide how to assemble and (if necessary) to jig the joint. An outline of the material selection was given in the previous section. Information on how to design local geometry and do (simple) stress analyses can be found for example in refs. 19, 23, 24. Generally speaking there is a need for better documentation of the whole design process for shipbuilding applications. The theory has been developed in many cases (e.g. in the aircraft industry) but it is still very difficult for the end user to carry out joint design for ships. There are many closed form solutions available for the analysis of lap joints (e.g. refs. 24 and 25). These can be used to estimate the bond strength. However, these formulae are not suitable for complex joint geometries, the only option here is to carry out an FE analysis. FE analysis requires a great deal of experience where special attention has to be given to failure criteria (there are no general failure criteria), non-linear material modelling, the modelling assumptions and simplifications and the boundary conditions. In many cases it is not possible to predict the joint strength by numerical modelling without additional calibration experiments. A practical approach is therefore to make full size samples of the joint instead and test them. Some important aspects which should be considered in the joint design are •

Strains due to mechanical or thermal loading.



Creep: especially important for flexible adhesives which can only take very little static load.



Design for production: ensure easy application of surface preparation and adhesive and assembly of the joint (access!).



Corrosion: sufficient electrical insulation from the adhesive to avoid corrosion between dissimilar materials, e.g. steel-aluminium.



Inspectability of joint: allow access for ultrasonic probes, joint can be inspected visually.

The following example shows results from ongoing research (ref. 26) on joining steel sandwich panels by adhesive bonding. Fig. 6 depicts a bonded joint between two steel sandwich panels. The adhesive used is a 2 component room temperature curing epoxy adhesive. The strength of this joint was predicted using simple formulae and the FE method and was compared with experiments:

10



Nominal-shear-stress-approach: 38.4 MPa x 100mm x 32mm x 2 = 245 kN. 38.4 MPa is the failure load of lap-shear samples with the same substrate material, surface preparation and adhesive.



FE simulation: the adhesive reaches the critical stress at an applied load of 160 kN. The critical stress inside the joint was obtained from lap-shear tests. An FE analysis of the lap-shear joint was carried out at the failure load which in turn yielded the critical stress used in the sandwich joint (Fig. 7).



6 samples were tested in a tensile testing machine and the following average failure load was obtained: 149 +/- 25 kN (Fig. 8).

In this case FE simulation yields the most accurate predictions of the failure load. The analytical method (nominal shear stress approach) over-predicted the strength by a large margin. This can probably be improved by using a more sophisticated analytical method (e.g. ref. 25). However, the biggest uncertainty is not the stress distribution in the joint but determining the forces acting on the joint. Note on the use of NDE/NDI There are currently no non-destructive examination/ inspection methods (NDE/NDI) which can measure the bonding strength of a joint or detect kissing bonds (see for example ref. 27). Hence NDE/NDI is utilised as part of an overall quality strategy for the production process. Ultrasonics can be used e.g. to confirm that a bondline is correctly filled with adhesive. The bonding strength is ensured by the correct choice of substrate material, surface preparation, adhesive and process parameter. The performance of this material system is demonstrated by (accelerated) ageing tests. NDE/NDI is used to confirm that the joint was produced according to the process specification. LASER-WELDED STEEL SANDWICH PANELS IN HOISTABLE CAR DECKS Metal sandwich panels (Fig. 6-7) are today increasingly being used in ships, rail, road and building applications. The first marine transport applications of all-steel sandwich panels were developed by the US Navy in the late 1980’s (refs. 29, 30 and 31). These applications have included bulkheads and decks in accommodation areas, deckhouses and miscellaneous other structures. Similar types of structures made of stainless steel have been used in Japan for high speed craft. Weight savings up to 30-50 % have been reported when comparing the novel panels with conventional stiffened panels (ref. 32). In Europe research related to all-metal sandwich panels has been carried out in Britain, Germany and Finland. The German shipyard, Meyer Werft, has performed theoretical and experimental investigations on the behaviour of laser-welded metal sandwich panels and their manufacturability (refs. 33-35). Studies of the design, optimisation and manufacture of all-steel sandwich panels have also been carried out at the Ship Laboratory of Helsinki Technical University (refs. 36-39). Various manufacturing techniques, such as resistance and spot welding and adhesive bonding, have been used for the production of all-metallic sandwich panels. As an alternative Meyer Werft has been developing and optimising a production technique based on laser welding. This technique offers high productivity and low heat input, which makes possible to connect thin metal sheets to form a light, strong and durable structure with minimal distortions. Sandwich panels are now used in various applications (e.g. decks and bulkheads) on board cruise ships and ferries. The main advantages of the laser-welded sandwich panels compared to conventional stiffened plates are as follows: •



Weight reduction up to 50% 11

50% less space consumption



Improved noise and vibration damping



Easy and fast assembly



Enhanced fire safety and heat insulation





Significantly improved crash resistance

Flexible use through modular solutions for panels and joints



Reduction of manufacturing cost by accurate pre-manufacturing

Based on previous research results a large European research project named SANDWICH (ref. 40) was initiated in 2000. The project has been aiming to further improve the properties of laser-welded sandwich panels, and to develop novel steel-composite lightweight sandwich panels for primary loadcarrying structures in ships and land transport by combining a laser welded steel structure with lowdensity foam cores. The following sections review some aspects of a real design case study for a hoistable car deck fitted with laser-welded sandwich panels. Hoistable car deck One of the principal advantages of using hoistable car decks is to increase the flexibility of a ship’s carrying capacity. Such types of deck enable vessels to carry two or more levels of light cars when the deck is in operating position. Alternatively, when the car deck is stowed, high vehicles and trailers can be transported. These decks are typically hoisted by a system of wire ropes running over sheaves and actuated by hydraulic cylinders. The cylinders (also called jigger winches) can either be integrated into the hoistable deck itself or be located in the ship side between the web frames (Fig. 9). These hoistable car decks are about 12 m wide and 25 m long overall. The average weight is 30 t when made out of conventional stiffened plate structures. For this particular design the motivations for using laser-welded sandwich panels were weight saving, reduction in the number of parts and weld lengths, and achieving a flat deck without any fairing work. The technical problems associated with the novel panels were numerous: local buckling of the top plates due to excessive bending, the fitting of lashing pots for flush decks, the integration of sheaves and hydraulics, the panel distortion caused by welding of thin panels (25mm), and unavailability of asymmetric standard beams (with small top flange and large bottom flange). The overall height of the deck was limited to 350 mm and one of the challenges for this particular design was to accommodate the jigger winches within this space. Loads, arrangement and design criteria In comparison to fixed car decks carrying trucks and trailers, hoistable car decks are only designed to carry light cars. The loads exerted on moveable car decks vary depending on the type of vessel. For this particular vessel a distinction was made between car carriers conveying new cars, which contain no luggage and have empty fuel tanks, and ferries conveying cars, which are fully loaded. The vertical acceleration was also taken into account for the calculations. Its value varies according to the position of the deck, and depending on the classification society, its formulation also varies. For this design a load of 200 kg/m² and a vertical acceleration of 1.25g were assumed - corresponding to a distributed load of approximately 0.0025 N/mm². An important factor in the structural arrangement of the hoistable car deck was the web frame spacing. Generally a transverse girder and a support are located at each web frame position on the outer edge. On or near the centre line of the ship, similar supports and pillars would be used to support the structure. A typical spacing is approximately 2.5 to 3 m. The maximum allowable deflection of the panels was set to L/200, where L was the girder spacing. For the grillage structure supporting the panels the maximum allowable deflection was also set to L/200, where L was the deck span. The formulations of allowable stresses set by the classification societies are different. This difference can be ascribed to the dissimilarity in the load assumptions and the vertical accelerations. For this design DNV’s allowable stresses were chosen, see TABLE 2 (all stresses given in MPa). The local loads induced by the wheels are specified by the 12

owner and the induced stresses must be below required allowable stresses. The local analysis is not presented in this paper. TABLE 2 - Permissible Stresses

TABLE 3 - Material properties

LR

DNV

BV

E

210.000 MPa

ν

0.3

100

160 x f1

150

G

80.800 MPa

ρ

8 × 10-9 t/mm³

90/k

-

90 x f1

80

σV 180/k

-

√σ²B + 3τ²

180

GL σB 140/k τ

Basic calculations General

The basis for the initial design was a standard I-profile panel with 2.5 mm plating and 4 mm I-profile stiffeners, spaced by 120 mm and with a height of 40 mm. The top plating of the ramps was built from 3 mm studded plating to prevent surfaces from becoming slippery. For these ramps Z-profiled stiffeners were then used instead of conventional I-stiffeners because of the manufacturing restrictions imposed by the laser welding process. The design was then optimised but a few limitations applied. For example the stiffener height could not be optimised because of a 40 mm minimum value. The stiffener spacing was also fixed because of the risk of local buckling induced by the wheel loads. The cover plate thickness was fixed by the minimum thickness requirements (i.e. 5 mm) and risks of welding distortion. As a result the design variables were the grillage span, the load and the panel height. For this design the scope was to optimise the sandwich panel and not the grillage structure. The number of transverse girders was thus reduced to a minimum. The girder spacing was fixed to 6 m, which corresponded to about every second web frame. The optimised value for the stiffeners was found to be 25 mm. In order to accommodate the jigger winch and to reduce the overall height of the car deck, a rectangular cut-out was made in the deck. At this position a conventional stiffened plate was used to provide the necessary height for the driving mechanism. The drive unit did not interfere with the supporting structure and had its own housing enabling a standard module solution. Due to the large girder spacing, only one cut-out was required for the cylinder and even here it was only for the shaft. Finite element analysis (FEA)

The aim of the structural analysis is to demonstrate by direct analysis that the permissible stresses set by the classification society are not exceeded in the given load cases. As no direct calculation method was available for the structural design of this structure, finite element modelling was used. The FEA software used for the calculation was NISA V9.0. The type of analysis was static and linear. The element types were 4- or 8-noded shell elements (Fig. 10). One ramp and two platforms (Fig. 9) were modelled successively. Due to their quasi-symmetrical layout, only one half or one quarter of the structure was modelled. The steel material properties used were as given in TABLE 3. In accordance with the DNV Rules for Ships (ref. 41) the allowable stresses were calculated to be σn = 160 N/mm² for the direct stresses and τn= 90 N/mm² for the shear stresses. The Z-profiles were not modelled but were assumed to behave as I-profiles. The sandwich panels were modelled with a mesh size of 120×75 mm (length × width mm). The longitudinal girders used for supporting the panels were modelled. The top flange (40×6 mm) was omitted in the model. The transverse girders were modelled as positioned, i.e. exactly below the sandwich stiffeners. However, the square tube used for the connection was not modelled. The webs of the girders were in most cases 290×6 mm but with tapers in some cases. In the vicinity of cleats and cut-outs the web thickness was sometimes increased. The thickness of the bottom flange of the girders was 20 mm. Different widths were used for the bottom flange of the girders. 13

Ramp 1

The ramp is hinged and is actuated by four vertical hydraulic cylinders. For stowing purposes, the ramp has a horizontal cylinder, which tensions cables pulling the ramp up below the next deck. To accommodate the horizontal cylinder, a few transverse girders have cut outs. The large cut-outs were modelled and the smaller ones were omitted. There are vertical supports at the ramp’s hinge and at the hydraulic cylinder mounting points. The uniform pressure design load exerted upon the ramp by 62 t of car load (31 cars of 2 t each) according to DNV rules is given by (5). The deadweight of the ramp is taken into account by using a body force given in (6). In both (5) and (6) an acceleration factor of 1,235 specified by the DNV Rules is used. The weight of the flap at the end of the ramp is assumed to be as follows at each longitudinal girder, see (7). P=

F 62.000 × 9,81 × 1,235 = 0,002331 N / mm² (5) = 25.300 × 12.750 A av = 9.810 × 1,235 = 12.120 mm/s² (6) F flap =

1.400 × 9,81 ≈ 2.300 N (7) 6

Results from FE analysis included contour plots of vertical deflection and stresses (longitudinal and transverse normal stresses, shear and von Mises stresses) in the plating, the sandwich stiffeners and the girders. Two contour plots of stresses in the plating and in the girders are presented in Figs. 11 and 13. Results showed that the permissible stresses set in the DNV Rules are not exceeded for this load case, except locally in one transverse girder (Fig.13). Here the transverse tensile bending stress exceeds the permissible stress in the bottom flange at the intersection between the transverse girder and the innermost longitudinal girder. This was considered as a hot spot. TABLE 4 - Summary of load cases values Load Case

F(t)

A (m2)

P (N/mm2)

fa (-)

av (mm/s2)

Fflap (t)

Ramp 1

1

31 x 2

25.3 x 12.8

0.00233

1.235

12.120

1.4

Platform 2,3

1

30 x 2

25.3 x 12.8

0.00231

1.269

12.450

0

Platform 2,3

2

0

-

0

1.269

12.450

0

Platform 2 and 3

For Platforms 2 and 3 two load cases were considered. The first load case corresponded to the platform in operating position, referred as Load Case 1; and the second corresponded to the platform in stowing position, referred as Load Case 2. For Load Case 1 the loads considered were identical to the loading conditions used for Ramp 1; only the number of cars, the area and the acceleration factors differed. In Load Case 2 only the deadweight of the platform was considered. TABLE 4 summarises the values used for the load cases. The load cases for Platforms 2 and 3 were identical but their geometries differed significantly. Results from FE analysis also included contour plots of vertical deflection and stresses in the plating, the sandwich stiffeners and the girders. Two contour plots of stresses in the plating and in the girders are presented in Figs. 12 and 14. Results showed that the permissible stresses set in DNV Rules are not exceeded for Load Cases 1 and 2. A comparison between a conventional structure and the steel sandwich structure indicated that, for the final designs of Platforms 2 and 3, weight saving was about 10%. For Ramp 1 weight saving was only a few percent. 14

This application has demonstrated that laser-welded sandwich panels offer weight saving, improved flatness and accuracy, high stiffness and strength (both local and global), a reduced space requirement, the possibility to integrate (within the structure depth) lashers, sheaves and hydraulics. FUTURE RESEARCH CHALLENGES For FRP composites the main research and development challenges today concern material characterisation and the need to ensure that the properties assumed in design are compatible with those achieved by the real production process. The situation is made especially difficult by the constant introduction of new or modified component materials and new processes. There is particular uncertainty about the influence of production defects and damage, especially on the compression strength of laminates with carbon or other high-strength fibres. There is considerable scope for development of faster and more reliable processes for non-destructive inspection of composites, particularly sandwich structures, and for methods of ensuring greater consistency in production. Other topics requiring R&D concern improved fire performance, including lightweight fire protection systems, development of reliability-based design codes that take proper account of the multiplicity of failure mechanisms that are possible with composites, proper characterisation of non-linear structural behaviour including buckling, and re-cycling. For aluminium alloys, the main challenges concern joining methods that are less costly and detrimental to mechanical properties than conventional welding processes. Indeed, for all lightweight materials, joining of both similar and dissimilar materials by novel processes such as adhesive bonding provides a series of challenges that are well worth facing because the potential returns are very great indeed. Lightweight structures tend to be more flexible than conventional steel structures. This means that buckling, vibrations and hydro-elastic effects such as springing and whipping may have to be addressed to a greater extent than with conventional structures. Provision of adequate impact, collision and grounding resistance also provides some research challenges, particularly for aluminium hull structures. Many new challenges are presented by the recent move in international regulations towards less prescriptive requirements for ship design. The use of formal safety assessment and probabilistic design approaches requires extensive modelling of accident scenarios. With accurate models it will be possible to give credit for novel, advanced solutions and make these attractive. However, this requires a thorough understanding and characterisation of the materials used and of the behaviour of the novel structural arrangements containing them. CONCLUSIONS In this paper a review of some of the recent progress in lightweight construction for advanced shipbuilding has been presented. The motivations for using conventional and more advanced types of lightweight materials and structures have been reviewed; commonly lightweight structures are used to increase payload, to reach higher speeds and to lower fuel consumption. In naval applications composite materials have been chosen because of the added benefit offered by attractive properties for specific applications. However, the wise use of composites requires an understanding of the differences between steel and composites structures. The specific detailed design requirements of a composite structure differ from those of a corresponding steel structure. Based on this principle, a cost benefit assessment method has been developed to compare steel and composite designs in order to make a rational selection and establish a balanced composite design. In addition a review of survivability of naval composites structures together with focus on design improvements, with particular reference to novel joint designs, also been presented.

15

The use of adhesive bonding on superstructures of high-speed craft and passenger ships aims to save weight but also to make production more efficient. The selection of a suitable adhesive for the planned application requires careful consideration of the geometry, loading condition and manufacturing requirements the joint is expected to experience. For the design and analysis of joints, closed form solutions are only available for simple lap shear joints. For common joints FE analysis is the only tool available for structural analysis and requires a great deal of experience, especially with failure criteria. Laser-welding technology offers high productivity and low heat input allowing production of lightweight, robust and durable structures with minimal distortions. An application of laser-welded sandwich panels for hoistable car decks has been presented, with some emphasis on design aspects. These decks have recently been built and are in use today. This application has demonstrated that laser-welded sandwich panels offer some weight saving, improved flatness and accuracy, high stiffness and strength, and a reduced space requirement. Some of the main future research challenges in the field of lightweight construction will be concerned with material properties, robustness of production processes, effective non-destructive inspection, fire performance for FRP composites, and joining methods for aluminium alloys. The shift of international regulations from prescriptive requirements towards formal safety assessment and probabilistic design approaches will facilitate the use of novel, advanced solutions by giving them proper credit and making them attractive. REFERENCES [1] Smith, C.S.: Design of Marine Structures in Composite Materials, Elsevier Applied Science, 1990 [2] Shenoi, R.A. and Wellicome, J.F.: Composite Materials in Maritime Structures, Cambridge University Press,Vol. 1, Fundamental Aspects, Vol. 2 Practical Considerations, 1993 [3] European Aluminium Association: TALAT – Training in Aluminium Application Technologies (CD-ROM), 1999 [4] 3rd International Forum on Aluminium Ships - Haugesund, Norway, 1998. (also 2nd Forum, Australia, 1996 and 4th Forum, USA, 2000 [5] DD(X) homepage on http://dd21.crane.navy.mil/ updated 2002-09-27 [6] Hayman B., Echtermeyer A.T., McGeorge D.: Use of fibre composites in naval ships, in proceedings of the International Symposium, WARSHIP2001 Future Surface Warships, RINA, June 2001, London, UK. [7] International Convention for the Safety of Life at Sea (SOLAS), IMO, http://www.imo.org [8] SOLAS, Ch.II.2, Construction – Fire protection, fire detection, and fire extinction, July 2002 (in ref. 7) [9] Manley, D: Procuring for survivability, in proceedings of the International Symposium, WARSHIP2001 Future Surface Warships, RINA, June 2001, London, The UK. [10] Manley, D: The development of "Smart" requirements for ship survivability, NATO RTO Symposium on Combat Survivability of Air, Space, Sea and Land Vehicles, Aalborg, Denmark, September 2002 [11] Wright, D. J.: The integration of vulnerability targets into warship vulnerability, NATO RTO Symposium on Combat Survivability of Air, Space, Sea and Land Vehicles, Aalborg, Denmark, September 2002 [12] McGeorge, D., Høyning, B.: Fire safety of naval vessels made of composite materials: foire safety philosophies, ongoing research and state of the art passive fire protection, NATO RTO specialists' meeting on Fire Safety and Survivability, Aalborg, Denmark, September 2002 16

[13] Potter, P.C.: Survivability of composite structures for naval applications, NATO RTO Symposium on Combat Survivability of Air, Space, Sea and Land Vehicles, Aalborg, Denmark, September 2002 [14] Baker, W.E., Cox, P.A., Westine, P.S., Kulez, J.J. and Strehlow, R.A.: Explosion hazards and evaluation, Elsevier Science Publishers B.V., The Netherlands, 1983 [15] van Aanhold, J.E., Groves, A., Lystrup, A and McGeorge, D.: Dynamic and Static Performance of Composite T-joints, NATO RTO Symposium on Combat Survivability of Air, Space, Sea and Land Vehicles, Aalborg, Denmark, September 2002 [16] Le Lan, J.Y., Livory, P. and Parneix, P.: Steel/composite bonding principle used in the connection of composite superstructure to a metal hull, in proceedings of SANDWICH2, Gainesville, USA, March 1992 [17] Clifford, S.M., Manger, C.I.C. and Clyne, T.W.: Characterisation of a glass-fibre reinforced vinylester to steel joint for use between a naval GRP superstructure and a steel hull, Composite Structures 57, 59-66, 2002 [18] French patent 91.08.439 [19] Wacker G.: Kleben - Ein neues Fügeverfahren im Schiffbau. In Handbuch der Werften, pp. 1347. Schiffahrtsverlag Hansa, Hamburg, 2000 [20] Anon: Sonderfahrzeugbau: Kleben als Alternative zum Nieten oder Schweißen. kleben & dichten 42(7-8), 20-26. 1998 [21] Reavey, D.G.: Marine experience of structural adhesives in hovercraft. Journal(September/October), 18-21, 1981

SAMPE

[22] Espie A.W., Rogerson J.H. and Ebtehaj, K.: Quality Assurance in Adhesive Technology, Woodhead Publishing, 1998 [23] Adams R. D., Comyn J. and Wake W. C.: Structural Adhesive Joints in Engineering. Chapman & Hall. pp. 359, 1997 [24] Burchardt B., Diggelman K., Koch S. and Lanzendörfer B.: Elastic Bonding. Verlag Moderne Industrie. pp. 71, 1998 [25] Bigwood D.A. and Crocombe A.D.: Elastic analysis and engineering design formula for bonded joints”, Int. J. Adhesion and Adhesives Vol 9, Nb 4, 1989 [26] Brede, M.: Final report on analytical and FE modelling of joints and verification of easy-to-use design rules, Draft BONDSHIP project deliverable, to be published, 2002 [27] Weitzenböck J.R., Niese F. and Hübschen G.: Non-destructive evaluation and inspection of adhesively bonded aluminium joints. In Adhesion '99, pp. 93-98, Cambridge, UK, 1999 [28] BONDSHIP: Bonding of lightweight materials for cost effective production of high speed craft and passenger ships. http://research.dnv.com/bondship/, 2000 [29] Sikora, J.P., Dinsenacher, A.L.: SWATH structure: Navy research development applications”, Marine Technology, 27,4 1990, p211-220 [30] Wiernicki, C.J. Liem, F., Woods, G.D. and Furio, A.J.: Structural analysis methods for metallic corrugated steel core sandwich panels subjected to blast loads, Naval Eng. J., May, 1991 [31] Marisco, T.A., et al.: Laser welding of lightweight structural steel panels, Proceedings of the Laser Materials Processing Conf., ICALEO’93, Orlando [32] Kujala, P., Metsä, A. and Nallikari, M.: All steel sandwich panels for ship applications, Shipyard 2000: Spin-off Project, Helsinki University of Technology, Ship Laboratory, Otaniemi, M-196, 1995

17

[33] Roland, F. and Metschkow, B.: Laser sandwich panels for shipbuilding and structural steel engineering, Meyer Werft, Papenburg, 1997 [34] Roland, F. and Metschkow, B.: Laser welding in ship building – changes and obstacles, Meyer Werft, Papenburg, 1997 [35] Roland, F., Reinert, T. and Pethan, G.: Laser welding in shipbuilding – an overview of the activities at Meyer Werft, Proceedings IIW, Copenhagen, 2002 [36] Kujala, P., Metsä, A. and Nallikari, M.: All steel sandwich panels for ship applications, SHIPYARD 2000: Spin-off project, Helsinki University of Technology, Ship Laboratory, Otaniemi, M-196, 1995 [37] Kujala, P., Klanac, A.: Analytical and numerical analysis of all steel sandwich panels under uniform pressure load, DESIGN 2002, Vol.2, Dubrovnik, 2002 [38] Romanoff, J. and Kujala, P.: The optimum design for steel sandwich panels filled with polymeric foams”, Proceedings of FAST 2001, Southampton, 2001 [39] Romanoff, J. and Kujala, P.: Formulations for the strength analysis of all steel sandwich panels”, Helsinki University of Technology, Ship Laboratory, Otaniemi, M-266, 2002 [40] SANDWICH: Advanced composite sandwich steel structures, http://sandwich.balport.com/, 2000 [41] DNV Rules for Ships, Part 5, Chapter 2, Section 7B, 303 ACKNOWLEDGMENT The authors would like to thank the following project partners for allowing us to reproduce the results: Markus Brede of IFAM, Bremen and Gregor Berns of MACOR, Bremen. The results presented in this paper are mainly based on the work carried out in the EU funded GROWTH projects, BONDSHIP and SANDWICH, and in the WEAO funded EUCLID project, EUCLID RTP3.21. FIGURES

Fig. 1 - Stiffened single-skin GRP hull structure (from Ref. 1)

18

Fig. 2 - Sandwich core materials a) polymer foam core b) honeycomb core c) corrugated core

Fig. 3 - Radford (USA), Skjold (Norway), Visby (Sweden), La Fayette (France)

19

Fig. 4 - Traditional composite T-joint (Ref. 15)

Fig. 5 - Typical metal composite joint

Fig. 6 - Bonded joint between 2 laser-welded steel sandwich panels

Fig. 7 - FE modelling of bonded joint between 2 laser-welded sandwich panels

20

Fig. 8 - Experimental strength values - 149 +/- 25 kN

Fig. 9 - Arrangements - Ramp 1 and Platform 3

Fig. 10 - Arrangement of nodes and elements in a steel sandwich panel

21

Fig. 11 - von Mises stress in sandwich top plate – Ramp 1

Fig. 12 - von Mises stress in sandwich top plate – Platform 3

22

Fig. 13 - Transverse normal stress in girder structure – Ramp 1

Fig. 14 - Transverse normal stress in girder structure – Platform 3

23