Proceedings of the 15th International Symposium on Yacht Design and Yacht Construction

15th International Symposium on

"Yacht Design and Yacht Construction"

Amsterdam, 16 November 1998

PROCEEDINGS

Edited by P.W. de Heer

October 1998

Organized by HISWA - Nationaj Association of Watersport Industries in The Netherlands, the International Trade Show for Marine Equipment METS 98

Printed by:

SIECA REPRO P.O.Box 117 2600 AC Deift The Netherlands

C1P-DATA KONHKLIJKE B1BLIOTHEEK. DEN HAAG

15th International Symposium on 'Yacht Design and Yacht Construction':

proceedings of the 15th International Symposium on 'Yacht Design and Yacht Construction', Amstetdam 16 November 19981P.W. de Heer (editor)- Deift University of Technology, Ship Hydromechanics Laboratory, The Netherlands

ISBN 90 370 0171-8

Subject headings: Yacht design, Yact construction

telefoon: +31 15 2561919

Recent Developments in the Velocity Prediction based on the

Delft Systematic Yacht Hull Series

J.A. Keuning, DeW University of Technology,

Shiphydromecha-nics Laboratory, Mekelweg 2, 2628 CD DeW, The Netherlands ₉₉

Table of contents

Programme I

Introduction 111

The Practical Application of the Vacuum Injection Techni4ue

in Boat Building. The injection of the hull of the Contest 55.

Hoebergen, Centre of Lightweight Structures TUD-TNO,

P. O.Box 5058, 2600 GB Deift, The Netherlands

i

Model Tests to Assess the Manoeuvring of Planing Cràft

Deaki, Wolfson Unit M. T.I.A, University of Southampton,

Southampton, United Kingdom ₂₁

The Development of Scantling Requirements in Support of the

European Boat Directive

F. Hartz, convener ISO Working Group 18, Hamburg, Germany 33

The Construction of a 65meter Wooden Barque for the Jubilee

Sailing Trust Employiiig Cold-Moulding Lamination Techniques and a Team of Mixed Skills and Physical Abilities

H. MacKenzie Wilson, Jubilee Sáiling Trust, Woolston, UK 57

Application of Advanced Computational Flùid Dynamics in Yacht

Design

M.M.D. Levadou, H.J. Prins, H. C. Raven, Maritime Research

Institute (MARIN), P. O.Box 28, 6700 AA Wageningen,

PROGRAMME

15th

International HISWA Symposium on Yacht Design and Yacht Construction Monday, 16 November 1998

Jade Lounge

08:00 - 10:00 Registration and Information

Room RIS

10:00 - 10:15 Word of Welcome

Moderator: Kim Hollamby

10,: 15 - 11:00

A. Hoebergen, TNO nstitute for Applied Physics, Deift, The

Netherlands

The Practical Application of the Vacuum Injection Technique in Boat Building

1 1:00 - 11:30 Break

11:30 - 12:15

M.M.D. Levadou, H.J. Prins, H.C. Raven, Maritime Research

Institute Netherlands (MARII), Wageningen, The Netherlands Application of Advanced Computational Fluid Dynamics in Yacht

Design

12:15 - 13:00 H. MacKenzie Wilson, Jubilee Sailing Trust, Woolston, UK

The Construction of a 65-meter Wooden Barque for the Jubilee Sailing Trust

Room RIS

14:00 - 14:45 J.A. Kenning, Deift University of Technology, Shiphydromechanics

Laboratory, Deift, The Netherlands

Recent Developments in the Velocity Prediction based on the DeW

Systematic Yacht Hull Series

14:45 - 15:30 B. Deakin, MTIA., 'University of Southampton, Southampton, UK

Model Tests to assess the Manoeuvring of Planing Craft

15:30 - 16:00 Break

16:00 - 16:45 F. Hartz, Convener ISO Working Group 18, Hamburg, Germany

The Development of Scantling Requirements in Support of the European Boat Directive

16:45 - 17:00 Closure

Jade Lounge

INTRODUCTION

The 15 -th International Symposium on Yacht Design and Yacht Construction has been organised under the auspices of'the HIS WA National Association of Watersport Industries, the METS Marine Equipment Trade Show and the Delfi University of Technology,Department of

Shiphydromechanics

The present symposium is the latest in a long row of HIS WA Symposia organised now over more than 30 years. Once again I think the organising committee has found an interesting mix between scientific and more practical topics of interest to all those active in the field of yacht design and

yacht construction . But also for all those who are interested in what is going on in this industry

the symposium once again offers a nice opportunity to keep informed.

The Organising Committee is very grateflul to all the authors, who were prepared to share their expert knowledge with the others active in the yachting community and who spend so much of their valuable time and effort in the actual oral presentation and in the preparation of the papers. I think that a symposium like this HIS WA Symposium is a very important and meaningful meeting were a diversity of people, with different backgrounds and present occupations, but all interested in yachting and what is associated with it, are willingly exchanging (new) ideas and their

knowledge and therefore all the effort that is put into organising it is very much worth doing so. Dr.ir.J.A.Keuning

The practical application of the vacuum injection

tee u

o 'que in boat building.

The injection of the hull of the Contest 55.

Arlen Hoebergen

Centre of Lightweight Structures TUD-TNO

PO. Box 5058

2600 GB Deift,. The Netherlands

Abstract

Closed mould injection techniques. (Resin,Transfer 'Moulding, RTM) have been successfully appliedforsmall products in large series. Thi has proven the advantages of a closed mould injection technique. The application of a closed mould injection teáhilique for large products in small series (i.e. boats) is hardly practised, although the technology exits! The risk of failure is often considered too high.

An introductioninto. the resin injection' technology is given. It will be shown that a large product needs to be injected using vacuum (with a resulting. maximum pressure difference of i bar). The following main options to optimise the injection fill time are identified:

Optimisation.. of reinforcement permeability: Choice of injectiOn strategy.

The general procedure for the mJection of a large product is given, focusing on one of the most criticalaspects: proper positioning and,securing of the city reinforcement and cores Finally, the injection method for thehull of the COntest55 is.described in detail as well.as the main production steps.

Although the RTM process and the vacuum injection technique already exists for quite a while, transferof the technology is well possible but mOst of the times not simple and straightforward. However, this paper shows. that a structured approach will. lead to the successful vacuum injection of any prodùct. To aid. disseminatión of the technology a process control tool to ensure the successful injection. of a large product .needs to be established

i. Introduction

Closed mould injection techniques (Resin Transfer Moulding, RTM) have been successfully applied for small products in large series [1, 2, 31. This has proven the advantages of a closed mould injection technique (better working conditions, less styrene emission from polyester.or vinylesterresins, betterquality products and less labour costs !). The application of a closed mould injection technique for large

products in small series (i.e. boats) is hardly practised, although the technology exists! Boat builders and other composite. manufacturers often want to switch to a closed mould (i.e. vacuum injeçtion) technique However, the risk of failure is often

considered too high. This is mainly due to a lack of knowledge of the technology and a lack of process control tools. This paper will show that a structured approach will lead to the successful vacuum injection of any product. Before going into details about the practical application and the injection of a hull of the Contest55an introduction into the resin injection technológy is given. The main variations in the technology and the most important processing parameters are discussed. The general procedure for the injçction of a large product :5 given, focusing on one of the most critical aspects: proper positioning and securing of the dry reinforcement and cores. Finally, the

injection method for the hull of the Contest55is described in detail as well as the main production steps.

2. The resin injection technology

RTM, VARI, SCRIMP, Infusion, Vacuum bagging [4,5,6,7,8].There are many words and many abbreviatións which try to describe a process which is basically very simple:

Dry reinforcement is placed in a mould, the mould aclosed and resin flows into the mould and impregnates the reinforcement. The driving force for the flow of the resin is a pressure difference.

Why do people use all these names and abbreviations? Because people 'invent' variations on the basic principle and want to give it a new and unique name. To fully appreciate the resin injection technology one needs to understand the basic principle and the possible variations.

The filling time of the injectiôn of a rectangular strip can be calculated [9], using equation (1).

* *12

fill_tlnle=2K)

With _.p porosity of the reinforcement

K _{permeability of.the reinforcement}

Ti viscosity of the resin

1 flOw distance (length of the strip)

applied pressure difference (constant during the injection)

Equation (1) clearly shows how different parameters influence the fill time and hence the cycle time of a product. The parameters can be split in:

Material properties: resin viscosity

reinforcement porosity and permeability Prodüct properties:

size, volume and shape (influenciñg flow ditance) Process properties

Pressure difference

Injection strategy (influencing flow ditance)

2.1 Pressure and mould type

The pressure difference applied and the size of the_{productalso determine the type of} mould (besides other aspects like product series and surface quality required). This gives us a first classification of resiti injection techniques as shown in table 1. Resin transfermoulding and_{vacuum assisted resin injection require stiff moulds to} prevent deformation of the mould due to the injection pressure. Vacuum injection however, can be practised using_{one flexible r semi flexible mould halve (either} singleuse fòil, re-usable foil or (silicon) rubber pads_{or thin fibre reinforced plastic} shells). Large products can only be manufactured_{cost effectively using a vacuum} injection technique, since the use of_{pressure would require very stiff and very costly} moulds. This however, implies thata maximum_{pressure difference of approximately} i bar can be achieved.

Table 1. Classification of injection techniaues based Øfl: ann1ied

+ : Indicates a pressure higher than atmospheric pressure

-i : indicates atmospheric pressure

O _{: indicates apressure belowatmospheric pressure (vacuum)}

2.2 Material properties

The driving force in vacuum injection is limited_{to i bar. In order to achieve a mould} filling process (and cycle time) that is fastenough _{it is necessary to optimise some} material properties. These properties are the viscosity of the resin and the porosity and permeability of the reinforcement. However, viscosity of the resin is usually limited to

100 to 500mPàs. The filling timeis directly_{related to the viscosity (see equation 1).} Consequently there is not much to be gained in _{trying to optiniie resin viscosity. The} porosity of most reinforcements used is in between0.5 and 0.85. The permeability of reinforcements however, can differgreatly. In addition _{core materials such as Balsa}

Injection technique

i -

---rr-- r

Applied pressure difference

Inlet port Outlet port

Resin transfer moulding _{+ i}

Vacuum assisted resin injection _{+ O}

relative high permeability. Accurate measurement of permeability is difficult [iø].. Approximate values of porosity and permeability of materials are shown in table 2. Table 2. Permeabilitiy and porosity of (reinforcement) materials.

A product is usually not made up of only one type of reinforcement. Different

materials are combined to get the. required lay-up. The. permeability of such lay-up can be roughly calculated from the permeability of the constituent materials using a.simple rulé of mixtures. However, this rule of mixtures. is not very accurate, especially when the permeability of the constituent materials differ greatly. This is due to

3-dimensional flow that occurs, which results in impregnation of the poorly permeable layer through resin flow from a highly permeable layer.. Knowing this effect, we can use highly permeable layers as resin feeder material. This can be applied either within the laminate (using Balsa, Rovicore, Multiniat or sometimes even a continuous random glassmat) and creating a resin, rich .core, or outside the laminate (using any

feeder material)., The feeder materials can be separated from the. actual laminate using a peel p.y. In this way this.'feeder .layercan be removed from the laminate, after cure. Cores applied in the laminate can be very useful in creating permeable channels allowing a quick injection of the laminate.

2.4 Injection strategy

The pressure difference is limited to i bar and material properties are only to be varied within limits Therefore, we need to optimise the injection strategy in orderto achievean acceptable filling time for a large product There are only three basically different injection strategies (shown in figure

I). Ali other injection strategies are always a combination of. the following three:

Edge injection Point injection Peripheral injection

The filling time is greatly influenced by the injection strategy. This shown by 8 examples of possible injection strategies for a flatrectangular plate of i 2 metre and 10 mm thickness (see. table 3 and figure 2). The fastest injectiòn strategy is about 25 times faster' than the

Point injection

Peripheral Injection

Figure 1. The 3 basic

Material . - . - .. . porosity

(-j

permeability [e-10m2]

Saertex'935 . 0/90/rañdom lay-up 0.63 0.6

Uni.filo 81.6 450 random mat 0.77 5

Rovicore4'50/D3/450 random mat + 'core'

83,

₅₀

Balsa 1" core 0.1.1 33

Feeder material 0/90 grid

95

. 100

1. Edge Injection short side

3. Peripheral injection

5. Point injection (2 points)

2. Point injection (1 injection point)

6. Edge injection with additional channel

r"çç1t,"rt

. .. t

-7. Sequential ,edge injection (3 channels) _{8, Edge' injection with grid of channels}

From these simple, examples_{wecan generally conclude that injection time is governed}

by:

Injectión distance.

Ratio of length of the injèctiòn channel and length of the flowfront. Table 3. Influence of injection strategy on fill time.

2.3 conIio

When we want to inject a large product we need to use vacuum injection (with a maximum pressure difference of i bar). We can identify the following options_to optimise the injection fill time.

1. Optimisation of reinforcement permeability:

s Use permeable reinforcement (random mat type materials) Use permeable core materials (óhannels)

Usepermeable feeder materials Choice of injection strategy.

How to inject a làrge product?

For any product that we want make we can identify the following general procedure: Define product requirements

Define processing requirements, Identify processing method

4 Describe and elaborate production steps

In the next chapter these four items will be discussed based on the hull of the Contest 55. _{Especially the processing method and the injectión strategy will be discussed in} detail in chapter 4. Before going into detall about this injection the overall production steps will be indicated. One aspect, which is of crucial importance, will be treated thoroughly: the positioning and securing of'the dry reinforcement. Another aspect, the selection and formulation of the resin, is alsocrucial but will_{not be treated in detall} here. However, one should bear mmmd that resin properties like viscosity, _geltime and peakexotherm can be critical and should be' thoroughly discussed with the resin supplier.

Injection strategy - ' Fill time

i Edge injection short side ₄₀₀

2 Point injection (1 injection point) ₂₀₀

3 Peripheral injection ₂₀

4 Edge injection long side ₁₀₀

5 Point injection (2 injection points) ₁₀₀

6 Edge injection with additional injection channels ₂₀

7 Sequential edge injection with additional injection channels 1 30

3.1 Production steps

Based on product and process requirements and the chosen injection strategy the production steps can be elaborated and fully described. Generally these steps are:

1. Mould preparation 2, Apply gelcoat

Apply first laminate, Ehand lay-up or spray-up) Position and secure thy reinfòrcement and core

Position and' secure injection channels and outlet ports and channels Seal mould with foil

Test for leakage and final check Inject resin

Cure resin

1O.Demould product

Most of the above indicated steps are quite straightforward and' easy but 'require dedicated personnel and accuracy. One of the most critical steps is the positioning and 'securing of the dry reinforcement. Initially this seems rather surprising. However, poor

securing can cause the dry layers to move or even fall down. Poor positioning can lead to highly permeabló channels. Because of its iñiportance this aspect is treated in detail below.

3.2 PosItioning and securing of reinforcement andcores

In the traditional hand lay-up or spray-up method the sticky resin also ensures that the reinforcement remains 'in place. Dry reinforcement layers on vertical mould parts however, can move, slide and even fall down. In addition the dry reinforcement layers are bulky and resilient. On 'applying the vacuum,in the mould this can lead to folds and wrinkles, mainly bccurring around edges. Depending on the size and' the

compression these folds and wrinkles can lead to highly permeable channels causing resin racetracking and the occurrence of cfry spots. Also cores must be positioned and secured properly for the same reason (i.e occurrence of channels in between different sheets oflbàlsa core).

There are 3 different:ways of securing: Temporary mechanical fasteners Permanent mechanical fasteners Adhesives

Clearly it is a requirement that these methods do not influence the laminate properties. The options given above are listed iii order of priority. Temporary mechanical

fasteners (such as. clips and clamps) are always preferred when possible. Otherwise permanent mechanical fasteners can be used (such as staples) which can be very convenient for securing reinforcement layers on cores.

Adhesives are very convenient to use as well however, they need to be used with caution because only little' is known about the short or long term effect on laminate properties. Therefore we have investigated several' adhesiveson both ease of

3.2.1 Experimental set up

Altera first evaluation.on ease of application and bonding strength 4 adhesives were selected out of 12. These 4 systems are:

multi-purpose adhesive spray in an organic solvent (cyclohexane) polyethylene hot melt

polyurethane hot melt cyano-acrylate adhesive

Firstly, for each adhesive the required amount (g/m2).to achieve a proper bond strength was determined. Secondly, test laminates were manufactured using the respective determined adhesive amounts. The build up of the test laminates was chosen to resemble the build up to be used for the manufacture of the deck. This test laminate build up contains:

gelcoat

hand lay-up layer (225 g/m2) adhesive Saertex 935 adhesive Rovicore 4501D3/450 adhesive Rovicore 4501D31450 adhesive Saertex 935

A reference laminate (without 'adhesive) was manufactured to be able to compare the results. In addition laminates were aged using rather severe cyclic temperature changes (6 hrs, -20°C and 6 hrs 80°C).

Subsequently, the following. materi,l properties were determined: Inter laminar shear strength (ILSS)

3-point-bending strength 3-point-bending modulus.

These material properties were determined at 23°C and 50%RH for laminates aged 0, 50 and 100 cycles, and in addition at 80°C for the laminates which were not aged.

3.2.2 Results

The test results are shown in figures 3. to 5. They. show the influence of ageing (0, 50 and. 100 cycles) and temperature.(80°C) forthe laminates with different adhesives on respectively the ILSS,, bending strength and bending modulus. Adhesive C and D showed clear decreases in interlaminar shear strength and bending, strength, especially at higher temperatures and after ageing, compared to the reference 'laminate. Adhesive A also showed a decrease, but less severe than C and D. Adhesive B showed hardly any difference in material properties compared to the reference laminateo This

adhesive was therefore selected forfurther testing_{to iñvestigate the sensitivity for the} amount of 'adhesive used. The results are shown in figure 6. From these results it is clear that more adhesive result in lower mechanical properties, although_bending modulüs is hardly affected, the ILSS and bending strength drops to about half the

strength. Twice or even three times this amount appears to be acceptable without a too large influence on the material properties. This encourages the use of this adhesive since accidentally spraying too much will occur occasionally and this will not result in a dramatic decrease of properties.

B C D

LamIns

Pego 3. ifh of adli,cs and agong on bn !LSS

3.2.3 Conclusion

It is rather surprising that_{a polyethylene hot melt is hardly influencing the material} properties, compared to the other adhesives. However, this can be explained by taking into account that the hot melt_{was sprayed on the laminate in small droplets. This will} result also in small droplets of PE in the cured laminate. These small droplets can be compared with voids, since there will be no interaction between the PE droplets and the resin. Because the droplets are always sprayed very evenly on the reinforcement the mechanical properties_{are not influenced severely. The clear ádvantage of the hot} melt against an adhesive spray is the absence of_{organic solvents, l'bis not only is} much healthier for the employees, it is. also rather contradictory to use a closed mould injection technique preventing styrene emission and simultaneously using an adhesive spray consisting primarily of an organic solvent.

I

U I

ri

IUU!.,U.F.

I

i I

I ;. U

I'.

I.

1riii

it

I[.]

I

11111

'an

LOU

bA.

b.,.&

mIChUIII.II pmnity

Figiac 6. flni iiInic oradboub, nimnigon nnicmmnboi p.opennia

'e 'e 4

I

I,'o «

I

nancy,i. .tqd and.00 4 I

I

Ii

I:

U1

Ii

IIIII

I:IItII

.

UL

lI1I

o

I. Ii LII}ii

[11 U

tE

i

i i i ri ii

i

I i i;uiii

IL IL ILi I.IIII

y. 'ruir

Ian

Ir: .

IL'.

1

1H

I

I

U LI

U

.

I

-uln. C Lambiate

Ftaa 5. Inlkanian of adlanMe niai agilig on kiniiain boning imdubs

u'c. A o D

LamiMt.

Figane 4. lnthieano o(ndluce niai aìig on boinini begsuungtli

Ian

an

4. The injection of the hull of the Contest 55

4.1 Product requirements and description

The product is the hull of aContest.55. A new yacht of the Conyplex shipyard. Outer dimensions were fixed (see table 4). Conypléx decided to try to realisea relatively simple mould: a female mould made. from. laminated wood WIThOUT

the help of a male plug '(i.e. a.female plug was made directly, which would act as the female mould directly). Consequently the hull would need filling, sanding and polishing to achieve the required surface finish.

Designer Dick Zaal was responsible. for the laminate lay-up. The proposed hand lay-up / spray-up laminate was slightly adjusted to facilitate the injection. The main laminates build up of the 'hull (laminate codes A to E) are shown in table 5. A description of the different materials used can be found in table 6. The first hull would be made as a Balsa sandwich shell construction (without stiffeners, skeg or stringer integrated). Integration of stiffeners,. skeg and stringer were to be realised at a second.injection.

Table 5. Main laminates build u

material I Lam A I Lam B I Lam C I Lam D : Lam E

outside M450 450 450 450 .450 i 450 Saertex 935 935 935 935 . '935 935 Saertex 935 935 935 935 935 935 Saertex935 ₉₃₅ ,935 935 .Saertex . . . 'Saertex935 . . . Saertex' 935" ₉₃₅ Saertex 935 ,.. . 935 Saertex 1128 . ', 1128 Saertex 1128 . ., 1128. Saertex 1128 . . 1128 M450 450 , 450 450 450 M450 450 450 450 450 450

Balsa,24 min Balsa Balsa Balsa ' Balsa Balsa

M450 450 450 450 450 450 .M450 . 450 450: 450 450 Saertex 935 ' 935 935 935 935 935 Saertex 935... 935 935 935 935 Saertex 935 . 935 Sacriez 935 . 935 Saertex 1128 . . 1128 1128 Saeriex 1128 . 1128' 1128 Saertex 1128 .. 1128' 1128 'M450 450 450 450 inside Tòtal glass' (g) 5540 6440 7375 15919 12758

Table 4. Contest 55 hull data

Length. hull

_{16.40 m}

Width hull

_{4.50 m}

Height hull

_{2.50 m}

Table .6. Materials used in the hÙll.

.4.2 Process requirements

Since the Contest 55 hull would be the first to b made with_{vacuum injection, it is} clear that Cònyplex was unfamiliar with the vacuum injection technique. Therefore, one of the main process requirements was to keep the process as simple as possible and easy to control. Furthermore1 the process should also be feasible for all other Còntest yachts.

4.3 Iijection method

4.3.1 Overall Injection strategy

The Contest 55 hull is both a long (16.4 m), wide (4.5 m) and high (.5 m)_{product to} inject in one shot. The height.results a hydrostatic pressure difference which can results in a lower driving force forthe injectiôn. There _{are three basic variations possible on}

the edge injection strategy to distinguish: 1. Edge injectiOn downward

2.. Edge injectiOn upward 3. Edge injection sideways

The injection downward is not preferred for two reasons. Firstly, bubbles in the resin and laminate will be entrapped more easily and secondly, there is .hiher riskon the occurrence of dry spots due to racetracking of the resin through.liighly permeable runner channels.

The injection upward can only be carried out with sequential injection channels (otherwise injection time will be too long). This however, would result in about 10

sequential injection channels each of which need to be controlled and opened at the right time.

Based on the requirement to have a simple process there was a clear preference fòr an injection with a grid of injectión channels in which the majority of the channels would allow for a sideways injection Tius would allow for_{one single resin inlet port (and}

Material Type and build-up . Supplier

M450 _{Random glass mat 450 g/m2}

Unifilo U816

Vetrotex

International S.A. Saertex 935 Bi-directional' glass fabric

0°-335 g/m2 1200 tex 900 -288 g/m2 600 tex CSM- 300 g/m2 2400 tex

Saertex Wagener GmbH & Co KG

Saertex 1128 . Bi-directional glass fabric 0°-480 g/m2 1200 tex.

90°-336g/rn2600tex

CSM - 300 g/n'2 2400 tex

Saertex Wagener. GmbH &. Co KG

Balsa 24 mm Balsa contourkore

thickness 24 mm

Baltek Corporation .

injection channel running fromthe_{sternto The bow (via the keel) and branches of} channels from the main channel in the keel to the flange (with the deck) The overall injection channel pattern is shown in figure 7

Figure 7. The injection strategy and position-of the injection channels.

43.2 A local injection problem

Based on the overall chosen injection strategy we must verify that there are no lOcal details which can cause the formation of dry_{spots or else disturb the injection The} Balsa core was sùfficieñtly :perflìeable. Therefore there_{wasnò additional feeder} material required on all laminates containing Balsa However,_{the laminate used in the} keel would have_{a low overall permeability. This.could cause diy spots on the keel} which were clearly shown using flowfront simulations Figures 8 and 9 show the

simulations done without (figure 8) and with (figure 9) additional feeder material on top of the reinforcement A small scale_{experiment was carried out to validate this}

simulation result A small part of the hull laminate was injected on a flat plate This part consisted of the keel laminate (partly with and partly without additional feeder material) and the balsa sandwich laminate shown_{in figure 10 It was injected using a main}

injection channel at the edge of the keel laminate and an injection channel branch in the middle of the laminate. The carefully registered flowfront.pròved thà occurrence of a thy spot on the part without the feeder material and the absence of a thy spot on the part with the additional feeder .materjàL

branch channel

Tjthouteeder

Figure 10. Test injection set up and resulting flow front.

Figure 8. Part of the hull with dry spot Figure 9. Part of the hull without dry formation at the location of the keel. Spot formation.

side view

flow front

Main injection channel

4.3.3 Final injection simulation

Finally the injection of the complete hull was simulated using the simulation program Pi7 (developed by TNO; future simulations are carried, out using RTMworx of

Polyworx, the successor of Pi7).. The hull and all laminate properties were modelled. Simulations were carried out to ptimise injection channel position and length. The final injection simulation is shown in figure 11.

FIUJNG TIME 10 M 1B$..3 34 12 10 28 26 2.4 22 lu IO 18 14 1.3 IO as 0G 04 02 Mfl O

Figure 11. Flow front sìmulation of the injection of the hull 'ofithe Contest 55.

4.4 ProdUction

For the production of the. hull ofthe Contest 55 the following steps were undertaken.

Mould preparation -

'.

:' . ' . '

The female mould, was checked on leakage and a release 'agént was applied.

pp1y gelcoat .. . . .

A transparent ISO-NPG gelc.oat was applied " . Apply first laminate '

A first laminate (450 g/m2 random mat) was applied with hand lay-up. This layer is applied for two reasons Firstly it protects the gelcoat when the dr' reinforcement is -positioned, secondly it provides a high quality layer ôf iso-phtalic polyester resin, while

the injection resin is an otho-phtalic polyester;

Position and securediy reinfôrcement and core

All layers of dry reinforcement and the balsa core are positioned and secured. Main methods of securing are clamping on the flange (see picture 12) and stapling. Only when these methods were not sufficient the PE hot-melt was used for securing the reinforcement locally.

Position and secure injection channels and outlet ports and channels,

The injection channels were positioned and secured with the PE hot-melt. The resin inlet tube was fixed to the main injection channel at the stern. Outlet ports (permeable strips) were positioned at the flange, in between injection channel branches. These outlet ports were 'connected' to a vacuum channel which was located, on both' flange sides from stern to bow. This vacuum channel was connected to the vacuum pump (with an overflow vessel').

Seal mould with foil

A double line of tacky tape was put on the mould flanges. The nylon foil was positioned in the mould' and sealed on the double line of tacky tape (see picture 13).

Test for leakage and final check

First vacuum is applied between the double line of tacky tape. This allowsa proper sealing of the foil on the mould. Secondly, a vacuum is applied in the mould. This is done very gently and slowly to allow the reinforcement to be compressed properly and the foil to be reposItioned where necessary. The scalings, all connections (outlet and inlet) and the foil are checked for leakage very thoroughly using an ultrasonic leak detector A leakage is sealed using tacky tape. Finally a final check is carriedout to ensure the correct position of injection channels, etceteras.

Inject resin

'The required amount of Reichhold ortho-phtalic polyesterresin (1200' kg) is mixed and the injection valve is opened. Resin 'flows into the injection channels and starts to impregnate the reinforcement (see picture 14 and 15). During the injection the product is carefully watched to' detect a possible leakage. If a 'leakage occurs in the foil this' can be sealed with tacky tape. The flowfront is watched as well and compared with the predicted i owfront. After 2½ hours the hull is fully' injected '(see picture 16) The resin inlet tube. is closed.

Cure resin

The resin 'gelation started after 4 to 5 hours. The peak exotherm was measured at different spots in the hull and limited to approximately 80°C.

Demould product

After the 'resin was cured the vacuum pumps were switched off, the foil' and injection channels were removed., After installation of several bulkheads the hull was demoulded (See picture 17).

5. Conclusions

In this paper the stepwise approach to realise the vacuum injection of a large product (the Contest 55 hull) is presented. Not all aspects have been covered, but only some major ones have been described. Although the RTM process and the vacuum injection technique already exists for quite a while and transfer of the technology is well possible, most of the times this is not simple and straightforward. The main reasons are:

There is (still) a lot of labour and craftsmanship involved. Consequently process control is difficult to establish.

Therefore, in order to be able to apply the vacuum injection technology it is essential that the manufacturer bears in mind the following:

Understand the basic principles of the technology.

Know the properties andrequirements of the materials youare going to use. Never JUST do something: KNOW why you are doing it and WHAT the consequences are onthe process and the final product properties.

Critical aspects to determine before starting are: Which materials to use?

How to inject the resin?

How to secure the reinforcement?

Whatcan go wrong during the injection? (AND: can you solve it?)

This paper has shown that a structured approach will lead to the successful vacuum injection of any product. To aid dissemination of the technology the main research area which must get attention is the realisation of a proper process control tooL This tool must be able to ensure the successful injection of a large product.

Acknowledgements

The work presented was funded through Senter (SM095103) of the Dutch Ministry of Economical Affairs. Much of the practical work has been carried out by dedicated employees of Conyplex under the supervision ofHan van Aalst and Corvan Zanwilet. This work was not possible without the close co-operation with the Dutch shipyards Friesland Polyester and Standfast Yachts, resiú supplier Reichhold and CVK. References

i Anonymous, "Resin Transfer Moulding: the designers choicà", Reinf. Plast., January 1994, 30-38.

2 _{Rogerson, T. "Is RTM right for you", Reinf. Plast. June 1991, 24-26}

3 _{Williams, D. "Resin injection yesterday's problem- today's solution"1 Reinf.Plast., June 1991, 46.48.} 4 Bunney, A. "Composite materials in maritime structures: production of ships with single skin

structures", Vol.2, Issue 6,91-116 (March 1993)

5 _{Lazarus, P. "Vacuum-bagging on the production line", Prof. Boatbuilder, Aug. 1994 18-24.}

6 Lazarus, P. "Infusion", Prof. Boatbuilder,Oct. 1994.

7 Seemann, W.H. "Plastic transfer moulding techniques for the production of fibre reinforced plastic structures", US Patent 4,902,215, 1990.

8 _{Seemann, W.H. "Plastic transfer moulding apparatus forthe production of libre reinforced plastic} structures", US Patent 5,052,906, 1991.

9 _{Gebart, B.R. e.a., "An Evaluation of Alternative Injection Strategies inRTM1!,47th Ann. Conf., Comp.}

Insu, Soc. Pl. md., February 3-6 1992, Session 16-D I h8.

10 Verheus, A.S. and Peeters, J.A. "The role of reinforcement permability in resin transfermoulding",

Picture 12. Securing the reinforcement and Balsa in the bow by means of clamping and stapling.

Picture 14. The resin starts flowing through the injection channels and impregnates the reinforcement.

I'

Picture 16. The hull has been fully inected.

Picture 17. Demoulding of the hull.

I

f

Model Tests to Assess the Manoeuvring of Planing

Craft

Barry Deakin.

Wolfson Unit M.T.I.A., University of Southampton

Abstract

This paper outlines the qualities which are required for good rnanoeuvring and control of high speed craft, and some of the design parameters which influence them. It describes a

system of simple tests which can be conducted on radio controlléd models, for design

projects with a modest budget, to highlight pOtential control problems or to:defme, boundaries of specific design parameters.

The papér is illustrated with data from recent commercial testing contracts.

Background

Planing craft designs have evolved and developed and, in general, good behaviour is assured with new designs deviating little from previous successful boats. Despite this, most people

in the industry will bave a tale to tell of a craft with an undesirable, or even dangerous,

handling characteristic.

While more craft are being operated at high speeds, and innovative designs are being

produced for specialised operational requirements, the incidence Of craft with manoeuvring problems is not diminishing.

The designer of a planing craft will influence the directional stability and control by his

choice of hull form, appendage design, and transverse stability. 'Unfortunately there are few reliable formulae or published data on which he can base ä prediction of these qualities..

Model testing offers a means of assessing a proposed design prior to production, but

conventional ship model testing techniques require detailed and accurate model hulls and appendages, and precise test execution. They bave developed to provide data for assessing and comparing standard manoeuvring characteristics in some detail. Justifiably these are seen as prohibitively expensive for the typical planing craft budget and, bemg developed for large slow ships, may not highlight the problem.

The Wolfson Unit. specialises in testing small models of fast craft, and 'has many years of experience in the investigation of their manoeuvnng through towing tank and simple free running model tests. Recent projects have included studies of the control of planing craft at an early stage of their design, when the test results might have a strong inflüence on some parameters.

One of the recent contracts concerned a 15 metre lifeboat for the Royal National Lifeboat Institution, with a design speed of 25 knots. Eight hull versiOns were tested at a scale of

1:10, some with twin water jets and others with twin propellers in tunnels. The boat was

intended for slipway launching, and the practical constraints on the bilge keel and central keel arrangement had significant implications for the manoeuvring. The other contract was for a 16 metre pilot boat to be built by Halmatic Ltd. It had a design speed of 32 knots, with propulsión by three water jets. The decision to incorporate a central jet precluded the use

of a conventional central skeg, and therefore had a large influence on the manoeuvring

characteristics. This boat was also tested at a scale of 1:10.

The operational requirements of both craft include good manoeuvring at high speed and in severe seastates, so the study of handlmg was given a high pnonty With a large number

of test configurations to compare however, a simple and low cost modelling and test procedure was required, which would address the concerns of the designers and help to

defme the boundaries within which the design could develop.

Manoeuvring qualities

There are a number of interrelated characteristics which are desirable attributes for most

craft:

Good directional stability. This quality will ensure that the boat will hold a steady course in the absence of external forces, and will respond positively to helm angle' changes. A boat with negative directional stability may continue to turn after the rudders have been put amidships,

Good manoeuvrabiity at speed.

The boat should be able to manoeuvre within a

reasonable turnmg circle Unfortunately manoeuvrability and directional stabthty are opposing characteristics, and an improvement in one is usually at the cost of the other.

Good manoeuvrabiity when idling. This is usually achieved through independent port and starboard throttle controls, but may be influenced .by design.

Good response to the helm. The boat should respond with a significant beading change following' the application of helm. This should not be confused with positive directional

stability, for example a boat with high directional stability and small rudders will be

difficult to turn.

.5. Moderate heel in a turn. Excessive heel in a turn, particularly if it is outward, gives the crew a feeling of insecurity, and frequently leads to other control problems.

Good control in following seas, with little tendency to broach.

Moderate influence of heel on control.

'Excessive changes to the manoeuvring

characteristics may occur when a boat is heeled, so that operation in waves or under the influence of other heeling forces becomes degraded.

Influence of Hull Design on Manoeuvring

There are a number of aspects of design which have significant influence on manoeuvring, but they are interdependent and changing one to improve a particular characteristic may have some detrimental effect on another characteristic. The following are considered to be the most important parameters.

Round bilge or chine configuration. There are many examples of both types of hull form

with control problems, and examples where changes to the chine configuration bas

improved some aspect of poor behaviour.

'Hull deadrise aft. In general, lower deadrise hulls, or those with excessive warp, are

more prone to control problems.

Hull deadrise forward. High deadrise forward, resulting in deep forward sections, tends to reduce directional stability.

Distribution of lateral area. The longitudinal distributión, or location of the, centre of

lateral resistance (CLR), affects the directional stability, with increased area aft being

benefrial.

Skegs and bilge keels influence the heel angle in a turn. They generate a side force directed into the turn and, being centred low down, this produces an outward heeling

moment.,

Longitudinal centre of gravity. This has two conflicting effects, a forward LCG promotes good directional stability by maintaining the centre of gravity forward of the CLR, in the same way as a dart with its weight forward and flights aft A forward LCG will also trim

the craft however, and this will reduce the directional stability by moving the CLR

forward'. The net effèct will depend, upon which of the two moves furthest.

Vertical centre of gravity. A high 'centre of gravity will increase the outward heel in a turn, and this may have a strong influence on the directional stability.

Trim control devices. By controlling trim with transom wedges,. trim tabs, or variable

drive angles, the directional stability will also be adjusted. Running with a low trim

generally reduces the directional stability.

Model Test 'Objectives

Most aspects of the hull design will be fixed or guided 'by other design requirements, the owner's specification, or practical constraints, and the objectives of a model test programme probably will be to ensure that the proposed design has no adverse characteristics, or to set limits for those parameters which may have a practical range within the design envelope.

the handling, and theft effects are incorporated in a model test progrpmme.

Depending upon the particular requirements of the project, the tests may be required to

confirm or quantify some or all of the qualities listed above.

Modeffing Requirements

Conventional ship model manoeuvring test techniques are described fully in Principles of Naval. Architecture, published by SNAME. The ideal requirements for modelling include correct representation of the displacement and centre of gravity, yaw and roll moments of inertia, rudder deflection angle and rate, propeller slip ratio, and response of the propulsion motor to the increased resistance when turning. The type of test to be conducted, and the

implications of the effects, should be considered before deciding whether each of these requirements needs to be met. For example, if the tests are only to: include steady turns,

there will be no need to model the moments of inertia or rudder rate.

For high speed craft, when the requirements include a study of the dynamic phases of the

manoeuvres, it is important to match the displacement, LCG, VCG, yaw inertia and röll

inertia. Lt is common for fast craft to trim down by the bow when heim is applied, so it is

worth noting that,

since for most monohulls the yaw inertia and pitch inertia are

approximately the same, any dynamic effect of a manoeuvre on the trim will also be

modelled correctly.

The maximum rudder deflection should be modelléd if manoeuvres at full helm are to be conducted, but frequently it is the effect of small or moderate helm angles which are of most interest. Usually it is not difficult to model the correct angle however, and this gives to

ability to make some assessment of the low speed manoeuvring qualities.

The rudder rate can be modelled by suitable choice of a constant speed servo and tiller

length. In thetwo recent projects concerning boats at an early stage of their design, the helm rate was not known and so no attempt was made to fix the model rate. The rudder rates for small craft are relatively high in comparison with those for large ships and it is more likely that a model servo will approximately match the scale rate for a small fast craft.

To match the propeller slip ratio is rather problematic, since the viscous scale effects result in a relatively high model resistance in comparison with the full scale vessel. To overcome this difference an air screw may be fitted to the model to provide a small thrust equal to the difference between the model resistance and the ship resistance at model scale. A carefully

engineered model of the full scale propeller is required, of a sufficient size that propeller scale effects are negligible. To satisfy this requirement requires 'large models and

correspondingly large facilities 'in which to test them, with accurately modelled propellers typically costing Several thousand pounds. The alternative approach frequently taken at the Wolfson Unit is to use a very simple model propeller which acts as a thruster rather than a

model of the full size propeller. The propeller diameter is modelled correctly, and so the flow upstream and downstream of the propeller is representative. Speed of rotation is

more power than at full scale. This enables the model size to be chosen on the basis of other practical considerations, and propeller costs are minimal.

The response of the propulsion motor to the increased resistance when turning will not be modelled in these simplified model tests, and so the reduction of speed in the turn will not be represented accurately. This may have other implications because heel angle in the turn,

and turning circle diameter may be dependant on speed, but the effects are likely to be

second order.

The high power to weight ratio of high speed craft presents a problem in scale modelling. The choice between' electric propulsion and internal combustion engine must take account the short range or high battery weight of the former, and the complications and poor reliability of the latter. There .is no dóubt that electric motors offer substantial advantages in terms of

controllability, ease of outfit, cleanliness, noise, and lack of pollution which may be a

consideration if testing indoors. They have been used with sealed 6 volt lead acid batteries distributed to obtain the appropriate ballast requirements. The high power consumptionmay

require water cooled speed controllers' or motor jackets, but these are readily 'available,

modelling components.

Care must be taken in model design and outfitting, to ensure that all variable parameters can be measured, and that all controls are repeatable and calibrated. There may be hysteresis in a steering linkage for example, which will result in significant differences between rudder angles when the indicated angle is approached from port or starboard. Thesteering nozzles 'or vanes fitted to model water jet drives are calibrated with the jets developing the same

thrust as at the test speed, but with the model tethered, because the measured anglesare

dependent on the flow speed.

The difficulty of calibrating standard radio control systems, and eliminating such problems as hysteresis, means that it is often preferable to use manual adjustment for as many variables as possible. Trim tabs are a typical example for, whilst it is essential to have control of them

on the boat, it is not necessary to vary them düriñg a single manoeuvring test. A simple

manual adjustment is used to fix the tabs accurately to a known angle, with tests conducted

at a range of angles if required. With tabs controlled via the radio, it is difficult to be

certain of their angle during the test, particularly if their deflection is affected by the force on them at high speed.

To minimise cost, all measurements of speed, heading, heel, trim, and rate of turnare made from observations aided by video recording. To install measurement transducers and data logging systems into the models moves the complexity and cost into another category which generally .is beyond the available budget and timescale.

Standard turbulence stimulation studs are fitted to ensure that the model boundary layer is representative of that at full scalè.

One of the lifeboat models, complete with propellers in tunnels, rudders, trim tabs and bilge keels, is illustrated in Figure 1.

Figure 1

Techniques for Testing in Cahn Water

Similar techniques were used in both propeller and jet driven models, but when conducting turns with jets, the vectored thrust remains relatively constant while the speed drops in the turn. As the speed reduces the turning circle diameter reduces, with further loss of speed. A steady rate of turn was achieved with jet propulsion, but a substantial settling period was

required.

Speed and Trim Calibration

Speed is measured to calibrate the throttle settings by timing the model over a measured distance marked with transit posts. Trim can be estimated from the video image, but all of

the designs tested at the Wolfson Unit have been tested in a towing tank prior

to

manoeuvnng tests, and the trim characteristics have been well documented.

If the

relationship between speed, trim and the intersection of the running waterline with the bow profile is known, the latter provides a reliable indication of the trim and speed combination during the tests, often being the first indicator of a drop in battery power for example.

Dieudonne Spiral Test

The standard ship manoeuvre for detennining the directional stability is the Dieudonne spiral. This involves timing the steady rate of turn at a range of rudder angles, and in particular at small rudder angles, with angles incrementing progressively from port to starboard through midships, and returning back to port. With an unstable vessel the turn to port will continue

be repeated for opposing rudder increments to eliminate any asymmetry in modelling or propulsion, or external effects such as the wind.

The importance of obtaining reliable data at small helm angles requires a large manoeuvring pond, and substantial endurance from the power source. If electric motors are used the short battery life may prevent constant speed turns of large diameter, although typically these small

craft settle into a steady rate of turn very quickly, and the timing of a 180 degree turn

following a 90 degree approach turn may be sufficient. To eliminate parallax errors, the rate of turn is best timed between the nearest and furthest points on the turning circle, by aligning two masts positioned in line athwartsbips.

Video freeze frame, and frame by frame time coding facilities give this methoda high level of accuracy. Figure 2 shows a model which has just turned through the mast alignment position close to the camera.

Figure 2

The view of the model from directly ahead or astern enables the hèel angle in the turn to be measured, again with the aid of the masts, as in Figure 3.

Small Angle Zig-Zag Test

An alternative technique for determining whether the model has positive directional stability is to conduct a low angle zig-zag manoeuvre. The model is driven away from the controller who applies small helm angles alternately to port and starboard, applying starboard helm when a port turn has become established, and vice-versa. A stable model will respond to

small helm changes by altering course accordingly, but an unstable one will require a

substantial helm angle to pull out of a turn. The heading changes can be seen clearly if the model is heading away from or towards the controller and the result, albeit qualitative, is obtained quickly with minimal battery consumption. This technique is very useful in tests

where a large number of model configurations need to be tested, and to conduct full

-

-*_

-;C

Figure 3

Large Angle Zig-Zag Test

The zig-zag test is a standard ship manoeuvre which provides information on the hull

characteristics, rudder performance and control system. In a 20-20 zig-zag manoeuvre the

helm is put over to 20 degrees to port, and when the heading alters to 20 degrees off the original the helm ¡s put over 20 degrees to starboard. This cycle is repeated a number of times and measurements made of the ship response in terms of the times of events, the

maximum heading changes and the maximum distances off the original track. Such trialsare difficult to conduct with fast models under manual control, but less precise large angle zig-zag tests provide a useful opportunity to observe the control qualities and heeling response to the helm. Tests with 20 degrees or full helm indicate the heeling response to the helm, the heel angle in the established turn, and quickly highlight any asymmetry in the model or

control system. See Figure 4.

Dynamic Heeling Test

With some ballast moved to one side of the model to achieve a static heel angle of about 5

degrees, the model is driven away from the operator, and helm applied as required to

maintain a straight course. An increase in the heel angle will indicate a reduction in stability with speed. When heeled,, most boats generate a yawing moment which turns the boat in the opposite direction to the heel, and this test alSo provides a measure of this tendency by the helm required to correct it.

Techniques for Testing in Waves

The self propelled model can be operated in a manoeuvring basin with wave making

facilities, and this provides an opportunity to observe its behaviour running at any heading in a seastate. One of the most important requirements for fast craft is the ability to operate in stern and quartering seas with sufficient control to avoid or pullout of a broach. At other headings, however, the waves are unlikely to cause problems with the control which are not exhibited in calm water. A large manoeuvring basin tends to be an expensive facility to hire and, if used for runs in stern and quartering seas only, may 'be inefficient,

The' towing tank offers an attractive alternative if it bas sufficient width to enable zig-zag

manoeuvres in stern seas. The self propelled model can be controlled and' filmed by

personnel on the carriage, which can follow the model 'along the tank. This procedure gives

the controller considerably more feel for the responses of the model' than he bas when

controlling at a distance from the side of the manoeuvring basin, because he is close to, and

above the model, looking in its direction of travel. He is better able to judge the position

and speed of the model relative to the waves, and take appropriate actions to initiate and pull out of potential broaching situations. Similarly, the video record is taken from a more useful viewing position than is possible in' the manoeuvring basin, and enables better observatiòn of model heel and yaw responses.

In either case the procedure adopted' has been to run the model in steep regular waves which have a celerity (wave speed) similar to the operational speed of interest. The model can then

be positioned as required on a wave, and run obliquely or in a zig-zag manoeuvre. A

particularly useful test is to run the model obliquely on the face of a wave, and then attempt to turn down the wave to take it directly astern. This simulates the broaching scenario and the level of control which is available.

Representative irregular sea spectra might be expected to provide more realistic conditions,

'but because there are a low number of wave encounters in a given run, many runs are

required to provide çncounters with sufficiently steep waves to challenge the capabilities of the model.

Sample Results

three different VCG values. Tests were conducted for a range of rudder angles, and the measured rates of turn were nondimensionalised with respect to speed. With the 'lowest

VCG the linear nature of the data indicate high directional stability. With the increasing VCG height the data become less linear as directiOnal staliity decreases,, and with the high VCG the discontinuous curve reveals that the model was directionally unstable.

I

z

-0.2

,

/

-03

-30 -20 -10

Rudder Angle - degrees

Figure 5: The Effect of VCG on Rìte of Turn

Figure 6 gives a chie as to the reason for this high sensitivity to VCG, showing the effect of VCG on the heel' angle m a turn for the sanie model This model heeled outwards durmg turns unless: the VCG was very low. As VCG was increased, the heel outwards increased, and the model's rate. of turn at low and moderate rudder angles increased as a result of the heel angle.

Table i presents further results for this model, to include the effects of :LCG on the

directional stability. 'In the case ofthe mid VCG height, the modèl was stable until the LCG was moved well forward of the design location It suggests that the VCG dominated the behaviour, but when the stability was marginal the effects of LCG also became important.

0.3 0.2

co.!

- VCG = 1.5m

= 1.8m = 2.Om VCG

vcG

o 10 20 30

Table 1. The Effect of Centre of Gravity on Directional Stability

io

8

-2

Rudder Angle 30 degrees Rudder Angle 15 degrees

/

/

-4

1.5 Ï.6 1.7 1.8

VCG - metres

Figure 6: The Effect of VCG on Heel Angle in a Turn

VCG LCG - metres forward of

design

metres -0.2 Design 0.2

1.5 _Stable Stable Stable

1.8 Stable Stable Poor

2.0 Unstable Unstable Unstabi

e

A model, of a different design always heeled into the turn, and proved relátively insensitive to VCG. When heeling into the turn, the induced yaw moment tends to reduce the rate of

turn, and a larger helm angle is required to maintain a, continuous rate of turn. For this design the handling was found to be more dependent on the trim, and at excessive trim tab

angles the model was notably less responsive. Figure 7 shows data obtained from the

dynamic heeling test fOr this model with a range of LCG values. When heeled to starboard

the model tended' to turn to port, and the starboard helm required to maintain a straight

course increased with increasing trim 'by the bow.

4

2

J

o

-o." -0.2 0.0

2

0.4

06

LCG Location - metes fwd of design location

Figure 7: HeIm' Required to Manintahi a Straight Course When Heeled, and its' Variation with LCG L cation

This is a normal result, and helps ils: to 'understand the problem ofbroaching. When a wave approaches from the port quarter, the boat accelerates down the face of the wave, and heels tO starboard. The heel induces a turn to port, along the face of the wave, and probably against the wishes of the helmsman who will attempt to turn to starboard to take the wave astern. If the trim tabs are down, the normal bow up trim will reduce as speed increases, and the yaw moment topo induced by heeling may be so great that a large starboard helm angle is required to counteract it. This may leave little additIonal helm angle with which to

turn the boat to starboard. This scenario was exhibited on several occasions when testing particular model configurations in quartering seas in the towing tank or manoeuvring basin

and collisions with the tank side were the result.

Acknowledgements

The author is grateful to the RNLI and Halinatic Ltd. for their permission to use data and photographs to illustrate this paper,.

The Development of Scantling Requireménts in

Support of the European Boat Directive

Fritz Hartz, Dipl. -ing., Naval Architect,

Technical COnsultant to ICOMIA

Abstract

This paper descrbes the lengthy procedure of developing scantling standards needed for the implementation of the European Boat Directive (Dir. 94/25/EG) by a working group within ¡SOIT C 188 SMALL CRAFT.

The paper deals with the difficulties of unifying national aspect, differing experience, the multitude, of existing rules, the involvement of mies making bodies, and the hopefully final

solution. The working group had to find a feasible way between a simple "cookbook"

approach and "'first principles". The result is a rules proposal which takes into 'account as many parameters as regarded necessary to take care of the vast spectrum of modem' boat proportions and configurations.

The draft proposai has been developed to a stage where comparative calculations have

evidenced that the proposal may go out for international comment. At the same time another set of calculations will check existing dimensions of boat structures 'in comparison to the' new approach.

1.

_Introduction

The necessity to develop a set of scantling rules for recreational craft became reality when the EU Commission decided that scantlïng determination should to be' incorporated in the

"Essential 'Safety Requirements" of the Directive on Recreational Craft. This was done

'against the industry's advice at that time. There was 'good reason for that negative reaction,

as a scantlings standard cannot be done on a few pages and within a normal time span available for the development of an International Standard (ISO Standard). Scanfling

determination has evolved with increasing knowledge in naval architecture and experience

from yards. It has been a "living" development unsuitable for the stiff format of an ISO Standard.

Furthermore there has been hardly any evidence of structural failures which might justify to regulate this part of a boat.

Despite this the decision was taken in Brussels, and ISO/TC 188 SMALL CRAFT was entitled' to do the' work. As only EN Standards. may be used to implement a EU Directive,,

the ISO Standard will be transferred into an EN Standard, which is planned to be

a

"Harmonised Standard" in the end. By this it represents the "state of the art", i.e. the

Alter more than nine years the ISO working group is now able to present a "Committee

Draft" on part 5, the first stage of an ISO Standard distributed world wide for comments. Parallel to this a validation program will start to check existing and well proven designs if

they fit into the new standard. The two-stage commenting procedure will allow for further adjustments.

2.

Basic information

The ideal basis for an international standard is a single natiônal' one which can simply be

transferred into an ISO draft - but this hardly ever happens. In many cases there is no

standard at all. Even this is a good starting point. Most complications are envisaged when there are many national standards, as they all have been developed with the best intention in mind and usually work pretty well - at least people have got used to them.

It is the abundance of existing Rules and Recommended Practices which made the matter so controversial and time consuming Almost any of the Classification Societies in the world have developed scantling rules for recreational craft in addition to their shipbuilding rules.

These rules have great benefits: They can be updated, changed, can take care of tests, investigätions, publications, can be withdrawn and be interpreted according to the latest knowledge. And the "surveyor's satisfaction" is another key element. They are or can be

"living rules". On thç other band: An ISO or EN Standard specifies binding requirements, there is hardly any interpretation possible, a revision will take years. It is a rather "frozen" paper.

The working group had to take into account six Class Rules and additionally two others

(Nordic Rules and Japanese Rules). When closely looking into these rules it becomes obvious that the origin of most of them is practical experience put into a mathematical format. The

boats were first (scantlings developed from experience) the rules were drafted then. The layout of these rules differ quite a lot, more even the parameters used. The simplest are

working with length and speed as parameters, the more sophisticated ones have been refined to incorporate many more parameters. But still then the very basics, i.e. pressures, contain a lot of experience and less results from systematic investigations.

The shortcomings of these Rules are evident:

- Rules based on experience tend to be conservative, especially if they are based on few

parameters, optimising needs a different approach,

- Accumulating experience takes time. Rules will always be a bit behind experience, - Rules cannot cover all the details, and they are more often responsible for failures thah

weakness of the overall or prime structure.

In addition to the above mentioned Rules there have been papers presented over the last

decades which deal with the subject of scantling determination; there have been measure-ments on accelerations and pressures. The majority of them are of U.S.., some of U.K., some

of Finnish origin. There does exist more information from tests and measurements, but pnvately financed as well as military investigations usually remam undisclosed for some

time. We can understand but we have to regret it.

3.

Composition of the ISO Working Group

Members of an ISO WG are national delegates of their respective Standards Institute. ISO1TC 188 SMALL CRAFT has 20 participating members, i.e. national Standards

Institutes, and a few observers. The present list of potential participants for Part 5

(Engineering) of the ISO Standard counts 33 members, which can be grouped as follows:

1 1 Naval architects; representing the industry

2 Boat manufacturers

1 Independent expert

1 User representative (naval architect)

2 Representatives from Boat Industry Associations

9 Representatives from Classification Societies or other Rule Making Bodies (among them Notified Bodies)

2 Representatives from Standards Institutes 2 Representatives from Government Agencies 2 Representatives from Notified Bodies

i CEN Consultant

Normal attendance at the WG meetings is about 20. The "industry" has a representation of 13 out of 33. Bearmg in mind that ISO Standards should be self regulated industry standards, the significant membership from rule making or rule applymg bodies (15) is exceptional The reasons, however, are obvious:, Once the draft has been adopted as an EN Standärd (and at least partly as a harmonised standard) it reflects the state of the art, the expected minimum. This means that other rules are obsolete or have to incorporate the findings of the WG.

4.

Format of the Standard

As the subject is complicated and cannot be dealt with by a .few-page draft but will result in a standard which will have the size of a booklet (see Class Rules),, the WG decided to split it up into six parts. This should allow fôr parallel develOpment and publication at least of parts of it when the work gets stuck at specific questions.

Part 1,: Small craft - Materials - Thermosetting resins, glass fibre reinforcements, reference laminate

Part 2: Small craft - Materials - Core materials for sandwich construction, embedded

materials