ADVANCED MARINE MATERIALS & COATINGS

RINA INTERNATIONAL CONFERENCE

ADVANCED MARINE MATERIALS & COATINGS

22 – 23 February 2006

© 2006: The Royal Institution of Naval Architects The Institution is not, as a body, responsible for the opinions expressed by the individual authors or speakers THE ROYAL INSTITUTION OF NAVAL ARCHITECTS 10 Upper Belgrave Street London SW1X 8BQ Telephone: 020 7235 4622 Fax: 020 7259 5912 ISBN No: 1-905040-22-9

Investigation into the use of Geopolymers for fire resistant marine composites A.C.J Flowerday, P.N.H Wright, R.O. Ledger, A.G. Gibson University of Newcastle upon Tyne

Main Participants This work was conducted as part of the individual research projects of:

Arran Flowerday

School of Marine Science and Technology

Rhiannon Ledger

Formerly School of Mechanical and Systems Engineering - Now working for Faber Maunsell

Background

Why are composites used?

Why is fire protection required?

Current options

Why are composites used?

Material properties

Weight

Cosmetics

QuickTime™ and a TIFF (Uncompressed) decompressor are needed to see this picture.

Why is fire protection needed?

The night before

•Rapid surface spread of flame •High heat release •High smoke emissions

Current options

Types of protection

Active

Intumescent coatings

Passive

Fabric materials

Rockwool

Sheet materials

Firebarrier

Properties defining fire performance

Fire Resistance

Thermal properties Mechanical properties post v. pre exposure

Properties defining fire performance

Fire Reaction

Time to ignition Surface spread of flame Heat release Emissions

SOLAS

SOLAS Chapter II provides specifications for fire resistant partitions

H - Capable of preventing the passage of smoke or flames for 1h against the hydrocarbon curve

A - Capable of preventing the passage of smoke or flames for 1h against the standard curve

Emissions

Geopolymer - Meyeb

Meyeb is a potassium alumino-silicate

Two component resin system Easy to work with No emissions to the workplace

Test Rig

Heat Flux Meter

Analysis The heat flow into a heat sink is given experimentally by

Q(t) =

msCPS (TS − T0 ) aS

This can be graphically equated to the theoretical values found using the thermal conductivity, diffusivity and the front face temperature n  ∞   n 2 π 2αt   −1) 1 t X 2 (  Q(t) = k(T1 − T0 ) −  + 2 ∑ 2 exp−  2  X α 6 π X n   n=1   

6 Glass - Meyeb 4.E+04

3.E+04

3.E+04

2.E+04

2.E+04

1.E+04

Q(t) Measure

Q(t) Theoreti 5.E+03

0.E+00 0

200

400

600

800 Time, t (secon

1000

1200

1400

1600

Analysis Verification

There is no other source of published thermal properties for this material.

The manufacturers were unable to supply any values.

How to check the validity of the results?

k

Q (Measured)

0

100

200 t0

300

Time, t

400

500

600

700

Valid results? t0 (sec) Sample

Calculated

Graphical

4 Glass Meyeb

45

45

6 Glass Š Meyeb (rpt)

110

107

6 Glass Meyeb

145

150

3 Carbon Meyeb

82

80

Results Sample 4 Glass Š Meyeb 6 Glass Š Meyeb 6 Glass Š Meyeb (rpt) 3 Carbon Š Meyeb Average k

kMEYEB 0.144 0.147 0.155 0.164 0.153

Results Average alpha

4.75E-08

Density of composite

2269.11

Density of fibre

2580

Density of Meyeb

2150

Vf

0.277

Vm

0.723

Cp G-M

1416.48

Cp fibre

810

Cpm

1695

QuickTime™ and a DV/DVCPRO - NTSC decompressor are needed to see this picture.

QuickTime™ and a DV/DVCPRO - NTSC decompressor are needed to see this picture.

Any Questions?

Advanced Marine Materials & Coatings, London, UK

CONTENTS Managing Coatings Through Life Raouf Kattan and Rodney Towers, Safinah Ltd., UK Improved Corrosion Resistance and Durability with Single Component Moisture Cure Urea Morten Sorensen, MC Technology, Belgium A New Approach for Ballast & Cargo Tank Coating: a Solvent-free and Humidity Tolerant Epoxy System with Edge-Retentive Properties Joao Azevedo, Euronavy, Portugal The Effect of a Foul Release Coating on Propeller Noise R. Mutton, M. Atlar, M. Downie, University of Newcastle upon Tyne, UK C.D. Anderson, International Paint, UK Environmentally Friendly Marine Anti-Fouling Additive Guy Seabrook, Magellan Companies Inc., USA Alocit Delta T and Delta dB Brian Glover, Alocit Systems Ltd Investigation into the use of Geopolymers for Fire Resistant Marine Composites Arran Flowerday, Rhiannon Ledger, Dr Peter Wright and Prof Gibson, University of Newcastle Upon Tyne, UK Vacuum Consolidation of Commingled Thermoplastic Matrix Composites for Marine Applications M Ijaz, Peter Wright, M Robinson and Geoff Gibson, University of Newcastle, UK Enviropeel Systems: Setting New Standards Tim Davison, Enviropeel Systems Ltd , UK A Non Chromate Conversion Coating Process for Corrosion Protection of Al2024 Aluminium Alloy in a Marine Environment. Wayne C. Tucker and Maria G. Medeiros, Naval Under Sea Warfare Center, USA Richard Brown and Dharma Maddala, University of Rhode Island, USA. Fatigue Crack Growth in Anodised Aluminium Alloys A M .Cree, Britannia Royal Naval College, UK G W .Weidmann, The Open University, UK Composite Overlay for Fatigue Improvement of a Ship Structure Gáspár Guzsvány and Ivan Grabovac, Defence Science and Technology Organisation, Australia Authors’ Contact Details

© 2006: The Royal Institution of Naval Architects


MANAGING COATINGS THROUGH LIFE M R Kattan and R H Towers, Safinah Ltd. UK SUMMARY This paper provides an overview of the issues related to the coating of vessels and the management of the coating system to prevent corrosion through the life of the vessel. The paper reviews how different factors throughout the ships life influence the potential for corrosion. Decisions made at each stage of the vessels life from design through to in service life and ultimately to its disposal will be considered and their impact discussed. Throughout the paper recommendations are made as to alternative approaches that owners may be consider to control costs. 1.

PLANNED MAINTENANCE

Planned maintenance has really grown out of the regulatory requirements of the Classification Societies, and as a function of ship management, it has become mandatory that all ships under Class must adopt some form of planned maintenance system. The main elements of any planned maintenance system will cover the main propulsion, navigational machinery and equipment, electrical power generation, and the hull structure. In fact planned maintenance systems have been developed to cover almost every part and function of the ship. The majority of shipowners/managers therefore already operate planned maintenance systems. Some owners have developed their own systems whilst others have contracted for proprietary software, the purpose of which is to provide planned maintenance information to the crew and the manager to ensure the safe running of the ship. Information from these systems will be monitored by Classification Societies as required, and they will provide other information for regulatory authorities such as for Flag or Port State control inspections. Planned maintenance systems are therefore established in the operation of the vessel.

well

These days there is a rising trend for Classification Societies to be offering hull management services that enable owners to record and retrieve a variety of information about their vessels and fleet, while at the same time allowing the Classification Societies to take a broader view on the performance of differing designs and alternative technologies in ships under their registration. Despite therefore all the progress and development in planned maintenance, and its wide acceptance as a vital function of ship management, there remains one on board system, which is often not subjected to a proper planned maintenance scheme, and that is the coatings system. This observation upon industry practice may be surprising for some, but perhaps not to others.


There seem to be two main reasons for this. One is about the perception of the role, which paint plays in the operational life of the ship. The other is to do with the cost of paint in the initial investment package of a new ship. 2.

PAINTING OR PRESERVATION?

In M&R, it seems that the traditional view of ship painting still generally prevails. The underwater area and the need to antifoul is a prima facie case for taking the ship out of service for drydocking, and has therefore always been treated as an important and specialist function. The direct link between fuel consumption and the performance of the antifouling coating has, of course, long been established, but despite this, many owners still seek to compromise on the quality of the work carried out on the underwater hull during dry-docking. Operators of certain special ship types, have also had the additional need to monitor and maintain particular locations. For chemical tanker operators, it was tank internal coating. Operators of LNG carriers came to recognise the direct relationship between preserving the WB tanks and achieving an extended trading life for such ships. More recently, there has been an entirely new industry focus on cargo hold painting in bulk carriers driven by ship safety issues. However the painting of other locations has often been seen more in terms of a ‘stop the corrosion spots and clean it up’ activity to deal with unsightly rusting and staining damage, which need covering up, rather than the need to maintain a long term preservation system, which can be important to sustain asset value and ensure the longevity of the vessel. Some owners do run their own systems for managing coatings. This often takes the form of a “Partnering Agreement” with a leading paint company, and there is trend amongst marine coating manufacturers to offer these types of services for the purpose of differentiating themselves, and as a way of providing added value to clients.


3.

PAINT IN THE INITIAL INVESTMENT PACKAGE OF A NEW SHIP

With the exception of chemical and products tanker ship types, coatings are generally seen as a low cost item, and in relative terms, this is correct. For example, and to place the paint supply and application in perspective, the following information may be of interest. The current market price (Feb 2006) for a 300K dwt VLCC newbuilding in Korea, as the benchmark, is in the range of $118-125 million dependant upon specification. Steel prices in Korea are currently around $ 700 per ton with some considerable tonnaqe imported from China. For such double hull design, the steel weight will be ± 35,000 tons This leads on to a quite logical question. What is the cost of protecting my asset within such a major capital investment? What therefore is the relationship between the overall cost to construct my asset and the input cost of preserving it? 300K VLCC ± 35,000 tons steel Full ship

Man hours Korea ± 500,000

≅ 14

Paint application

35,000 – 40,000

≅

/

Man hours per ton steel

1

The current cost for the supply of all coating materials, except shop primer, for one 300 K VLCC will be in the region of $ 2 million which currently equates to 1.6 – 1.7 % of total ship cost For standard ship types therefore, the cost of the paint package, that is the coating materials and the application work, will together represent something of the order of 9 -10% of the total ship cost, possibly higher. This is therefore a larger cost item than is perhaps generally understood by the shipowner. The shipyard will only seek the owners approval about the proposed paint supplier, whose supply items only represent between 1.5 – 2 % of the ship cost. For this reason there will usually be limited scope for the purchaser to influence the shipbuilder on preservation issues during the final negotiation.

work, and they will contract this work out in parcels of varying manhour content, often at a lump sum price. For this kind of work, shipyards will probably maintain better records of total cost, including consumables such as blasting abrasive where used, but figures for this are difficult to ascertain. When manhour figures can be obtained, these may be difficult to reconcile between yards. When viewed on their own, manhours may also understate the total cost, and so they should be used cautiously. In the case of chemical tankers and Aframax tankers with fully coated tanks, the paint package is likely to represent 15% of the ship cost or higher, and so the specification detail for a new tanker of this type becomes even more important, with approval and the decision on the buyers side often reaching Director level. It is worth stating that this cost differential will be reflected in the detail of the ship specification; for example, the painting of cargo tanks will be developed from a functional specification; the selection of an antifouling will have to meet a performance requirement, whereas the rest of the hull coatings will be a specification based on generic products. The starting point for painting specifications is therefore not consistent. Both the functional and performance approach, in conjunction with the acceptance of the manufacturers recommendations for maintenance, have shown that it is possible to achieve the functional and performance objectives in service. If therefore a shipowner is going to invest $100+ million in the steel structure of a new ship, it must surely be increasingly logical to ask the question – should I be taking a more modern approach to the preservation of my asset through its life ? 4.

ROLE OF COATINGS

Coatings can serve a variety of functions on board a vessel and these include: -

Protection from corrosion Prevention of fouling of the hull Cosmetic appearance and house colours Protection from corrosive effects of cargoes Protection from corrosive effects of having to carry non earning liquids, e.g. ballast, freshwater, grey water etc.

This paper will focus on how to improve the prevention element of corrosion.

With regard to the man hours for paint application, this is an area, which, in practice, is wide open for variance between shipyards in the same country, and between countries, where best practice painting methods can differ greatly; eg as between Korea and China. These days there is a widespread practice amongst shipyards to subcontract surface preparation and paint application



5.

COATING PERFORMANCE

7.

SELECTION OF COATINGS

Like any system the correct performance is dependent on a number of factors: • Coating selection • Specification of the coatings to perform the tasks required • Correct installation of the system • Subsequent maintenance at sea by crew or riding squad • Reinstatement of system at major maintenance events (Dry-dock)

The current approach to coating selection therefore may not be the best possible approach. The lack of confidence in coatings not being able to perform satisfactorily through life is reflected in the results of the Joint Tanker and Joint Bulk Carrier studies. Both studies have concluded that the only real way to ensure that ship strength will be maintained through life, is to design the hull structure with additional steel rather than by placing undue reliance on coatings systems being able to perform as planned.

6.

Whereas this increase in scantlings is not a substitute for the corrosion protection afforded by good coating systems, it is an indication of the importance given to coatings by Classification societies when thinking about the structural integrity of a vessel.

COATING SELECTION

For many fleet owners and operators, there is a well established pattern of treating the new build phase of the vessel as a separate project with finance and budgets set by non technical departments, and within which limits the new building project team must endeavour to deliver the vessel. In practice, this financial approach tends to override the technical requirement, and often results in scant regard being paid to the through life needs. As a consequence, this approach often leads to the selection of coatings, based wholly on cost without any evaluation of through life performance requirements, and cost benefits. Of course coating selection at new building can be influenced by a number of things: • the shipyard will purchase the paint and so may have a preferred supplier or a limited makers list. • the shipowner may have a fleet paint supply agreement with one or two paint companies, and may therefore wish to limit selection to these companies to sustain consistency across the fleet. • the paint supplier may be offering incentives or rebates to particular clients to help secure the business and keep out competition. • overall first cost and payment terms can also be issues Two key factors are often missing from this approach. They are • How will the coatings selected perform during the new build phase and the overall building schedule of the shipyard. • How the coatings selected will match the intended operational pattern of the vessel; one particular aspect of this is the nature of and the elapsed time between final coating acceptance and loading the first cargo. If the needs of these two criteria are not recognised, then the ongoing penalty to the shipowner will emerge in the form of increased costs through life. Once their performance has been compromised, there will result a commitment to increased costs of maintenance through life.


Thus a step change is required in the selection process; a change in which the following factors are taken into account: • Where the vessel will be built and the duration of the build as well as the season. • How the vessel will be operated and the anticipated operational environment to which the ship will be exposed. • The planned maintenance system to be adopted for maintaining the condition of the coating systems, whilst at sea and during scheduled drydocking periods. • The employment of suitable methods to assess coating condition, and enable preventative maintenance to be carried out • The resources available to carry out the maintenance on board. All these factors will have a considerable impact on the life expectancy of the coating and can only be properly taken into account through the use and development of a functional paint specification. This also has the added approach of very clearly distinguishing the true and objective technical merit of competitive bids from various paint companies as well as ensuring that the shipowner devotes adequate thought to the benefits of different functional requirements. 8.

THE MEANING SPECIFICATIONS

OF

FUNCTIONAL

A simple example of a traditional coating specification would be as follows for a deck: Coat in Scheme Coat 1 Coat 2

Type of coating Anticorrosive Finish

DFT

Colour

80 microns 50 microns

Grey Green

Generic type Epoxy type Epoxy type


All paint companies that respond will meet these technical criteria and therefore limit the choice down to price.

9.

Requirement 3 – year life with 1% spot rust acceptable Initial gloss retention and gloss retention after 24 months Taber value

Gloss

Abrasion resistance

OF

THE

We can summarise a number of problems found: -

If this was a new build deck, then the following shipyard functionality could be added: Function Max time to Overcoat Drying time to walk on at 10 centigrade

INSTALLATION

Picking the right system is fine but work with which the authors have become involved during the last 8 years has clearly shown that the majority of paint failures are caused by the paint being asked to perform out-with its defined envelope of capability. Poor selection has been in part responsible for this, but poor surface preparation and application (Installation) of the system has usually contributed. This leads to an analysis of paint failures carried out by Safinah Limited:

If however a functional specification were developed, then the following questions would need to be addressed by the intending paint supplier: Function Anti-corrosion

CORRECT SYSTEM

-

Requirement 6 months 12 hours -

Other factors could be considered for the deck but the above serves to illustrate the example. In this way the technical merit of each paint company solution can be quantified against specific functional needs and performance gauged.

-

Additional items can be added to meet specific vessel needs in operation, maintenance and also taking into account HS&E issues.

-

This makes performance measurable against defined and objective technical requirements.

Poor design of the vessel had resulted in inadequate access and ventilation for initial application work to be carried out. The installation work for these activities was found to have been poorly scheduled and planned or integrated into the build or repair processes. It was recognised that such instances had usually been treated as an after thought. Supervision of the installation was found to have been relatively poor when compared to other systems on the vessel (superintendents tend to have a marine engineering or deck background, with varying knowledge of coatings systems and their installation). Standards for assessing work tend to be subjective and open to broad interpretation. Coating system installation often takes a very low priority with owners superintendants, when compared to other systems. Weather and other environmental factors can adversely affect installation, and it was not possible to overcome these in instances where vessel or yard had to meet the defined delivery schedule or other operational constraints (e.g. tides, charters etc.)

45 40

Percentage

35 30 25 20 15 10 5 0 Design

Spec

Applic

Chem

Operation

Other

Cause Figure1: Cause of failure analysis © 2006: The Royal Institution of Naval Architects


Not withstanding all of the above, considerable time is still spent inspecting and approving the installation, and yet there is still often recourse in the event of a failure. Poor records of the installation or repair is another problem which complicates the issue. During installation the owner’s representative will usually rely heavily on the Paint Supplier’s technical supervisor, and it is a common misnomer that this supervision is there to protect the owner’s interest, when in truth at new build, the makers’s supervision is primarily to protect their own interests, and secondly to protect those of their paymaster, which, for any new construction, is the shipbuilder. When ownership passes from the builder to the owner, and after the 12 month shipbuilders warranty period has also expired, the paint supplier becomes responsible to the owner but this cannot undo what has been done at the shipyard during the construction. . At repair, the attendance of the paint company’s technical representative is primarily to protect their own interests, although it must be said that the owners representative will often impose upon the paint supplier to take general responsibility for control of surface pretreatment standards, and the inspection of the various applications by the shipyard workforce. The extent of the owners reliance upon the paint manufacturers representative will depend upon the degree of mutual trust and confidence, which can be established between the two parties at site. The specification is frequently changed or adjusted at the time of the indock inspection, and consequently much paintwork repair activity tends to be at best ad hoc, with few examples of pre-planned work packages linked to any planned maintenance system. By contrast, the offshore industry has been adopting planned maintenance for structural painting for some considerable time, probably driven by difficulties of confinement and safety constraints, when undertaking even basic maintenance painting, whilst platforms or drilling rigs are operational. 10.

WARRANTY INCOMPATIBILITY WITH LONGER TERM PRESERVATION OBJECTIVES

Every owner will get a 12 months shipbuilders warranty on their new ship. However, because of advances in antifouling technology, it may be anything up to 5 years before a VLCC will drydock for the first time. The consequence of this has been to prompt shipowner’s to search for the security of some longer term warranty from the paint supplier in respect of the antifouling system.


In the case of coated cargo tanks, the extended warranty concept is now fairly standard but initially this was promoted by paint manufacturers as a means of persuading shipowners to accept and specify advanced technology coatings, which would enable them to carry a much wider and more aggressive range of chemical cargoes with beneficial higher freight income. This practice of split warranty responsibility has become an industry standard but, shipbuilders, in stark contract to car manufacturers, have still not changed their warranty offer to meet that of the clients operational market requirement. At the end of the day, this is all about shipbuilders being prepared to embrace functional specifications within their standard offers, adjusting their working practices to achieve the necessary technical control over application, and to consider ceasing the practice of treating all paint products as commodities. This would open up the concept of shipbuilders entering into longer term arrangements with one or two appropriate paint suppliers, who have both the product assortment and technical capability to fully support the shipbuilder. 11.

SUBSEQUENT MAINTENANCE AT SEA BY CREW OR RIDING SQUAD

‘On board ‘ maintenance by crew is often poorly planned and poorly executed. A simple example of this would be to consider a small area touch up of a 2 coat deck system. The crew would detect a rust bloom or a coating failure; they would generally surface prepare the area back to bare steel using a mechanical method, and with some inevitability, feather the edges by grinding. Normal practice will probably be to apply the anticorrosive by either brush or roller or both, and then apply a finish coat. However, what is often not so well understood, is that for an original 2 coat system as outlined above, that is say 80 microns + 50 microns, this will require a 5-coat manual application to re-instate. A typical brush and roller application will achieve only 30 microns per coat. From our own observations, it is very rare for the crew to know this or be allowed the time to full re-instate the system in this way. The problem is that if the proper DFT’s are not reached, this invites an early repeat of corrosion activity at the same location. Repairs by riding squad tend to be better planned and better executed because they are seen as significant operational costs, and the work will usually have been specified in more detail to assist the contractor price the work. Nonetheless to ensure that the owner gets value for money these also need to be properly supervised and planned.


It is felt that the proper planning of coating maintenance should provide the crew with appropriate work packages and timings for their execution to ensure cost effective maintenance of the vessel to a “good” standard. Such standards exist for ballast tanks, but owners should establish their own standards to meet their own operational needs in particular locations, based on corrosion, cosmetics etc.

It is the opinion of the authors that if planned maintenance systems were applied to all the coating locations on board ship, then the payback would come from the reduction in costly steel replacement later in the ships life. However, there is again often a lack of consideration of the importance of planned maintenance to avoid even more and to ensure that the Anti-fouling is well applied to assure cheaper running/fuel costs.

12.

13.

REINSTATEMENT OF SYSTEMS AT MAJOR MAINTENANCE EVENTS (DRYDOCK)

At major scheduled repairs the quality of the work carried out is often simply dictated by the time available, and the weather conditions, as well as any unforeseen underwater hull work required either for steel or machinery/equipment. The net result is that, in practice, the proposed paint plan/programme is often compromised to meet operational schedules, and hence the practical result is often less than satisfactory. This can apply even to antifoulings. Assuming long term preservation is a continuing objective, the owner must apply the same rigorous approach to the reinstatement of full coating systems as they do to other critical systems. To achieve this, proper work packages need to be developed and a proper schedule of work defined with agreed standards and reporting. Such approach is clearly best suited to computerisation, and the use of existing software solutions can readily assist in accomplishing this. Supervision of coating activities during these major works is often given a very low priority by the owners’ personnel who often have a primary focus on steel and machinery repairs (perhaps more familiar ground for them). Reliance is again placed on the paint supplier. It is also an unfortunate fact, and very necessary to understand, that the costs of recoating are often disproportionately higher than during newbuild. The authors repeatedly hear about estimates for reinstating coating systems reaching up to $50 per square metre, which only seems to emphasise the importance of getting the initial selection right and the application done well.

OTHER BENEFITS

Planned maintenance systems allow data to be collected over time. The benefits of this are that the performance of particular solutions can be evaluated on a continuous basis, and feedback from lessons learnt in similar ships across a fleet are likely to prevent the same mistakes being made again. This in itself is likely to result in some future cost savings. Ship owners need to develop appropriate systems to monitor and plan the coating work, so that over time they can eliminate the most common causes of failure by better selection, and control of the installation. The effect of this will lead to the better overall management of coatings through life. Some of the existing planned maintenance systems allow Internet driven access, and provide remote expertise to deliver solutions working together with the crew and the managers of the vessels. To gain the benefits from such an approach, owners will need to consider how to develop in house expertise or alternatively to outsource the appropriate resources to manage such systems across a fleet. 14.

CONCLUSIONS

With the exception of antifoulings, and the attention given to paint in respect of some specialist ship types, coatings often appear to fall into a no-mans land in terms of the resource allocation of many ship owners. Standard specifications need to be carefully reviewed, not just for functionality but also to take into account the age of the vessel. Having a standard fleet wide specification might increase costs by using too high a quality of coating or too low a quality of coating on both new ships and older ships. Better to specify a coating that fits the need. Even a simple manual planned maintenance system could generate cost savings for owners whilst, at the same time, increasing the life of the vessel and reducing the incidence of coating failures. More comprehensive computer/internet based solutions are likely to result in additional benefits arising from the better management of paper work, and the automated. generation of reports and work packages.



The information generated will provide a real method of monitoring coating performance; will provide a system of early warning for locations likely to need maintenance; and will help to highlight common trends arising across a fleet of ships. Lessons learnt may lead to improvements in the detailed design of future ships, and possible fleet wide savings in operational costs. Improved knowledge in-house will lead to supervision of painting work both on board and repair yard, and it is felt that this can be achieved use of modern systems with a minimum of additional internal or external resources.

better in the by the either

If therefore planned maintenance systems, which are now being applied to so many shipboard functions, are in fact generating their own cost benefits for operational ship management, then it must make sense to consider extending this principle to embrace coatings systems also. The first steps therefore could be to just trial one or two specific locations.



IMPROVED CORROSION RESISTANCE COMPONENT MOISTURE CURE UREA

AND

DURABILITY

WITH

SINGLE

M Sorensen, MC Technology, Belgium SUMMARY In the 1970’s the offshore industry made a request to the coating industries to develop a coating that could be applied in high humidity conditions. This was due to the numerous pre-mature coatings failures for offshore marine maintenance. We will discuss here one of the most viable new technology developments the moisture cure urethane, more aptly called polyurea and developments over the past 30 years. The most viable result was a single component pure urethane product that goes through a rather complicated production procedure where it is pre-reacted. The final cure is through the exposure to minimal amounts of moisture forming a polyurea. This is quite a different product to the two component polyurethane coatings in many ways. In comparison to two-component urethane, it is a much safer product for the applicator in primarily due to the lack of isocyanate (carcinogenic) free monomer. The polymer technology for this technology is used for production of many common products today including artificial heart valves, tennis shoes, automobile parts and caulking compounds for example. Independent third party testing has proven this to be more effective for corrosion resistance and long term monitoring of projects have demonstrated the MCU’s to be more durable and have a longer lifespan, when compared to plural component technology. 1.

EARLY MCU DEVELOPMENTS:

Initial products produced associated with them;

had

inherent

problems

•

Over application of the then low recommended dft’s produced gas entrapment or film blistering,

•

The intercoat adhesion was not very good

•

They were unstable – often curing in the can prior to opening Developments in laboratories often do not succeed in the real world. In the real world a coating must have as wide tolerances as possible as it is next to impossible to complete an application exactly as per manufacture’s recommendations. Time is limited, the overcoat needs to be as short as possible with subsequent coats applied too soon or beyond the overcoat limitations, the coating application is often too thick, and the climatic conditions are often beyond the manufacturers recommendations, too humid, too cold or beyond the dew point. Several coating manufacturers attempted very lengthy and complicated production methods of MCU’s in order to try and stabilize the material often involving “nitrogen blankets” over the product in the production tanks and in pails after canning. This process was time consuming, not cost effective and in the end offered no guarantee of stability and the other problems still have not been dealt with. All the major coating suppliers became discouraged and simply put this down as another good idea that does not work and could not justify continuing with the developments. Over the initial years several smaller firms maintained a research into developments of these coatings. However the initial three key problems still exist with most


commercially available products even today. Despite these inherent problems numerous projects were completed. 2.

WASSER, BACKGROUND

In 1980 William J. Brinton made significant discoveries and developed his own proprietary resins and formulations. Mr. Brinton’s new formulations were the beginnings of a major turn-around for the MCU technology. He was able to solve the three key problems and manufacture a product line with; •

Better than average intercoat adhesion,

•

No maximum recoat time,

•

Products capable of being applied at 2 – 3 times their recommended DFT, without gas entrapment.

• Long-term stability in the can. The new firm, Wasser High-Tech Coatings, was established in the USA, with the sole intent to manufacture and market MCU’s. In the early stages Wasser focused on the bridge and Dam business. Within six years Wasser MCU became the single largest supplier of bridge and dam maintenance coatings in the USA [1]. This was quite an achievement for a brand new, previously unknown firm, producing a technology product, which the industry’s majority did not believe in. Naturally to gain approvals for these major government projects came only after third party testing required generally by each state and provincial’s approved third party testing laboratory. This involved numerous testing and the results were very positive. All testing showed this material to be either the top performer or of the top, depending on the test criteria.


The Army Corps of Engineers (ACE), a premier American group, dealing with marine structures, have completed a series of tests for steel structures in harsh corrosion environments. In total Wasser has been tested by ACE for over a 12-year period with various tests. This has included Panama Canal project, 3.5 years testing (Wasser became the exclusive supply for maintenance for 15 years), also testing for over-coating of existing coatings (Wasser rated best), and recent test comparisons of various MCU firms Wasser rated the best. [2]

labour time and also the revenue loss for down time, which can be considerable, the owners and superintendents need to keep the vessel in dock a little as possible. Therefore time is of the essence and time is money, usually big money. Many ship owners are adopting riding crew work completed at sea during voyage. This reduces the loss of revenue associated with downtime at key-side or in drydock. However the time factor of completing a job is still a significant cost, with the labour costs, equipment required, mobilization, air travel, etc., are still over 90% of the project cost and the coatings are generally 5% 10% of the cost. The MCU coatings non-restrictions can save and reduce project time and associated costs by as much as approximately 20% - 30%. 3.1

TYPICAL PROBLEMS ASSOCIATED WITH COATING PROJECTS

The problems start out simply in three parts, coatings choice, surface preparation and the application. Figure 1: The Astoria-Megler Bridge The Astoria Bridge was the one of the first major projects Wasser was entirely used for. This structure is on the Pacific Oregon coast and subjected to constant salt fog and condensate. It was also part of a 6-year joint Federal Highways Agency and Oregon Department of Transportation coatings evaluation program [3]. This test report included 10 of the top-performing technologies available including; various zinc systems, various epoxies, waterborn’s, Wasser mcu and even rust converters. It is interesting to note that in this test the zinc silicates as a stand-alone coating, out-performed the same primers as compared to being over-coated with epoxy and a polyurethane finish coat. This test concluded that the Wasser system was the only coating that was rated SSPC SP10 (less than 0.01% corrosion) after 6 years exposure to marine salt fog environment. After the last full inspection 14 years after this Wasser coating project was completed, the structure was still rated SSPC SP10. This was previously unheard of for bridge structures in or outside of a marine environment. 3.

COATING OF CONDITIONS

STEEL

IN

ADVERSE

The ship and platform owners and their contractors are often faced with these adverse environment conditions, potentially and frequently causing delays for the completion of a project, due to the application restrictions of most all coatings. However delays are usually unacceptable and the applications have to be carried out in conditions beyond the manufacturer’s recommendations, which can cause potential paint failure down the road. Due to the expense of a dry-docking, the

The choice of the correct coating for the job is the first task. The coating generally must be as capable and flexible in its characteristics as possible. When deciding the suitability for the intended use some considerations are; surface tolerance, overcoat & cure time, adhesion, flexibility, moisture tolerance, abrasion and erosion resistance and life span. The life expectancy of the vessel and of the project should be considered in order to determine when this may have to be repaired or recoated again. Surface preparation problems: The tight time frame for vessel repairs, leave little time tolerance for proper surface preparation. It is often inadequate and at times when proper the substrate flash rusts. This should be reblasted to accept most coatings, however time or budgets do not allow. Coating requirements for these situations need to be surface tolerant to flash rust, exhibit good wetting out characteristics as well as having good corrosion protection properties. Coating’s intolerance to flexing varies. Flexibility in coatings helps a great deal as vessels are subjected to flex, in their weld seams, corners and along the longtitudinals. The external hull is subject to reverse impact and decks, cargo hatches and coamings are subjected to impact and abrasion. If coatings do not have enough flexibility or become brittle over time, the coatings can crack. In turn these cracks open the film and allow moisture to penetrate under the coatings starting at the crack interphase and in many cases start the premature failing and corrosion process.



• • • • •

Remain flexible, inert after full cure and n/c chemically or physically. Better resistance over salt contamination. Better UV stability. Better abrasion resistance. 2 – 3 times longevity

These MCU coatings offer many applications as well as performance advantages. They are all single component and cure with minute traces of moisture. They can be applied in humidities up to 99% and without dew point restriction and in temperatures from -15°C to +50°C. Figure 2: This paint cracking caused by hull flexing (likely a reverse impact) and a coating choice that was too rigid. The top coating adhesion was not good and when the membrane broke the result is intercoat adhesion failure causing the paint to peel. A contributing factor to this could be that the surface was either damp when coated or the surface may have reached the dew point before it had a chance to cure.

Figure 3: This coating cracked along an inside corner and as this coating system had no corrosion protection other than the film itself, corrosion started from moisture creeping in through the interphase and undercutting the coating. 4.

THE MCU SOLUTION

Initially these moisture cure urea coatings were intended to be a solution for cold/hot and damp conditions. However during the past 25-years of the varied applications of these products and numerous third-party [3] testing have shown them to be a superior coating in many ways: • They have outperformed most all coatings available on the market in corrosion testing. Passed long-term salt spray 20,000+ hrs. [4] • They have excellent adhesion and require less blast profile (20-35µm), primers are surface tolerant to flash rust and magnetite. • Better wetting out properties, necessary to penetrate into deep pittings. • Excellent adhesion to and often rejuvenating old coatings. • Better edge retention. • They can be applied without dew point restrictions and in humidities to 99%.


4.1

THE BENEFITS OF MICACEOUS IRON OXIDE

Figure 4: This illustration shows the two of the key workings properties of micacious iron oxide (mio). The upper part of the diagram demonstrates the layers of mio flakes shielding the medium form the degrading effects of uv radiation. The lower diagram shows how the interleafing particles also reinforce and strengthen the coating film, by impeding the penetration of moisture and pollutants. This overall structure also avoids accumulation of moisture and gas entrapment by allowing micro permeable dissipation[5]. This phenomena is similar to the workings of gortex material. The key for a MIO pigmented coating is the quality of MIO used and the quantity used by weight in the volume of the coating. MIO is a mined material (later developments of synthetic types do not perform), and some mines may have more impurities than others, therefore it is critical to use a high quality. Two other key attributes of MIO is the resistance to erosion and longevity of the system and the edge retention of the film. The degree of MIO hardness is considerably more than any coating material and as the top layers of coatings erode the “MIO” becomes exposed and then retards this erosion further. The laminar aspects improve the edge retention, findings by improving the film strength, reduce polymer swelling, and form a tough laminar seal [6]. The pigment characteristics are much improved over glass flake epoxy and aluminium, without the negative aspects of cohesion problems. The MIO actually improves the inter-coat adhesion significantly.


The Wasser coatings use MIO in most of their coating products and are used in all of their marine and structural steel coating systems. 5.

MCU PRODUCT HIGHLIGHTS

The two key primers are both surface tolerant to dampness and also flash rust. They exhibit excellent wetting our properties that allow the coating to penetrate into pittings and into poor weld seams and inside corners. They offer excellent adhesion to steel, iron, aluminium, alloys, stainless steel, galvanized steel, Metalized and corten steel. They will also both adhere to most all, existing coatings. The over-coat time is generally 3 – 4 hours and with the ® PURQuik additive a 3-coat system can be applied in as low as 3 hrs. Most coatings in these systems, (including the surface tolerant zinc primer), have no maximum overcoat time and can be over-coated (on a clean surface) in many months later and have ideal intercoat adhesion. These coatings can be subjected to rain, condensate or even immersion within 30 minutes. There will not be any affect as to the cure and will not cause an amine blush. There are 3 key coating systems that can be applied and used on the entire vessel. One of two primers, one of two intermediate coats and one of three finish coats. Two coat systems can also be used. 5.1 • • • • • • • 5.2 • • • • • • • • • • •

MC PREPBOND Surface tolerant penetrating primer/sealer Highly abrasion resistance Designed initially poor surface preparation All metal, GRP and concrete surfaces Will penetrate loose rust, recommend to remove scale & apply mechanically Overcoat within 3 – 5 days Passes 5,500 hrs. salt spray, NORSOK approved MC MIOZINC st

The industry’s 1 surface tolerant zinc rich primer. Can be applied to both ferrous and non-ferrous substrates. Zinc and Mio filled, excellent edge retention. Compatible with zinc anodes. Recommended for immersion Surface preparation from ST 2 to SA 2.5. Excellent adhesion to existing coatings. Capable of high builds to 300 µm without bubbling or cracking. Infinitely re-coatable. Potable water approved [7] Passes 10,000 hrs. salt spray, NORSOK approved

Figure 5: Upper – Macro photo of a proper blasted corrosion spot. Small omega pittings are visible. The pittings should be cleaned as much as possible. Lower – Good filling properties of the Wasser MC Miozinc into the fine cavities. [8] 5.3 • • • • • • 5.4 • • • • • 5.5 • • • • • • •

MC BALLASTCOAT / MC CRPW Light coloured for ease of tank inspections Suitable for; ballast, drinking water, grey-water, black-water, drilling mud, cargo and fuel tanks Use as an intermediate coat for white finish Potable water approved [7] High abrasion resistance Applied in a one or two coat over primer MC LUSTER True aliphatic pure urea Excellent gloss and colour retention Capable of exposure to condensate, dew, rain fog or immersion within 30 minutes after application. Will not amine blush. High abrasion resistance. MC FERROGUARD Environmental friendlier coal-tar epoxy replacement Manufactured with further refined pharmaceutical grade coal tar. Mio pigmented, uv stable, resists cracking Adheres well to existing coal tar without abrading VOC compliant Excellent moisture, chemical, Passed 20,000 hrs salt fog test [4]

There are 13 coatings in the full range. All are VOC compliant worldwide.



6.

CASE HISTORIES

Figure 6: MV Hual Trubador, Ballast Tank, Hoegh Fleet Services AS System: MC Miozinc – MC BallastCoat

Figure 6: The Stena Discovery is a large aluminium fast ferry catamaran traveling at 40 knots capable of carrying 400 cars or combinations of cars and commercial trucks. Having a design life of 40 years. The much of the internal facing aluminium was not coated. In order to reduce the corrosion of these some of these areas for example ballast tanks, are flushed with sweet water and together with dry voids use a complicated and costly system of air-drying. However the air intake chambers presented a problem area.

Completed by riding crew in 2000, water jetting surface preparation. After 3.5 years, the coating is in 100%, with no coating breakdown and no signs of corrosion, even on edges and scuppers.

Figure 7: MV Spaarneborg, RoRo vessel deck, Wagonborg Shipping. System: MC Miozinc – MC Prepbond RoRo vessels experience pre-mature coating failures often and typically within 6 – 12 months after application either new or after recoat. Wagenborg conducted a detailed test study on one of their vessels with various coating systems including ceramic filled epoxy, glass flake epoxy, high-build epoxy and Wasser’s MCU. After several months and destructive testing by a consulting firm [8], Wasser was chosen as the replacement for their deck coatings. These vessels load containers weighing 60 to 90 tons each.


Figure 7: Aluminium in the turbine air intake chambers, developed 4 – 6 mm omega pittings in the 8 mm plate. Many areas required already to be replaced. These pitting vary from steel as they have extremely sharp edges and can develop inwards at an angle. Initially an epoxy system was applied to a portion of the air intake chamber. After several months it was discovered that this system failed to solve the problems, and the electrolysis in the pittings under the coating were still active. A test panel was prepared by UHP and the MC Prepbond was applied. After approx. 60 days in service an inspection was carried out and the system appeared to be performing well and the problem had been solved. The penetration was evident deep into the pittings and the adhesion value was between 7 – 9.2 Mpa. In all 14 readings the failure was in the glue attachment of the dolley, not in the coating or the aluminium – coating interphase.


PC 30% % µm M2/Ltr M2/Ltr

PRODUCTS SV DFT TC MC BallastCoat

Figure 9: The photo of the affected areas after a test MC Prepbond. The project was completed in January 2005, a full detailed reporting of this project will appear in an article in Shiprepair and Conversion Technology Magazine in March 2006. [8]

62 75

8,27

5,79

AREA M2

Est. Qty Ltr

10000 1728

MC BallastCoat 62 75 8,27 5,79 10000 1728 2 coat appl. total ltr 3456 MC BallastCoat 85 150 5,67 3,97 10000 2521 Figure 10: This illustrates the comparison from a twocoat application of the MC BallastCoat product. The first two rows are calculated at 62% VS, applied in two applications. The last row is calculated at 85% VS applied in one coat. The new HS material does not sacrifice any of its well know characteristics such as surface tolerance, wet out penetration, flexibility, shore hardness or moisture impedance.

Figure 8 These areas are in a critical area and no abrasives of any kind could be used, that could be sucked into the turbine. The surface preparation was completed by UHP Systems BV with their ultra high-pressure equipment. 7.

NEW MCU BREAKTHROUGH – HIGH SOLIDS MOISTURE CURE

Another factor, which scared away most coatings manufacturers, is the cost factor. The common mcu technology is capable of producing a 53% volume solids (VS) product. Mr. Brinton was able with new technology to develop a 62% volume solids product. However the raw materials to produce these technology materials is significantly higher than many coatings on the market. 7.1

INCREASED TOLERANCES

The latest new development improves the MCU through minor modifications MCU HS coatings improves the film integrity. These coatings can be manufactured at a high volume solids of 75% to 85%. These new developments allow for film thickness of 400 – 500 microns without gas entrapment or bubbling. The typical system for tank internals such as ballast, drinking water, grey and back water and fuel tanks, can now be applied in two coats instead of three, with fewer errors still and at a considerable further application cost savings. The square meter material cost of the new high solids will be reduced by an approximate 20% to 30% in comparison to the present material.

The volatile organic components (VOC), or solvents are reduced from 320 grams per liter to approximately 100, thereby also making the products more environmentally friendly. This is primarily a concern for shipyards in Europe and the United States where certain maximum allowances may be enforced or may become enforced in the future. 8.

CONCLUSION

The Belgium – Norwegian group of MC Technology acquired the Wasser technology a few short years ago, purchasing it from Mr. Brinton, who remains on their board as the technical director. During this brief period the group has become known as a quality producer of this unique coating. The coating has gained acceptance from some key ship owners and management firms. The coatings have proven themselves in tough marine environments such as ballast tanks, cargo tanks, RoRo vessels and offshore platform work, world-wide. This technology has become successful due to the fact that the time and labour savings is considerably more than the cost of the coatings in most cases. Projects could be completed with savings of 15% - 30% of the project cost as a whole and savings exceeding the coating costs in many instances. In addition the longevity is perhaps triple of what of typical current technology products. The products are now being specified on rehabilitation work as well as new build projects.



9.

REFERENCES

[1] Pittsburgh – 1986 Pittsburgh Engineering ReviewNace Conf. [2] Army Corps of Engineers [3] ODOT/USA FEDERAL HIGHWAYS Agency, 6year test [4] BP Labs Ltd., Independent laboratory testing [5] JPCL February 1981, by D.M. Bishop - The Mystery of the Magic of MIO, by Malcolm Hendry - ASTM Designation D 5532 –94 mio [6] JPCL November 1995, Comparison of natural and Synthetic MIO, by S. Wiktorek [7] Complies with ANSI/NSF Standard 61 potable water. [8]Wink Inspections BV 10.

AUTHOR’S BIOGRAPHY

Morten Sorensen, holds the current position of Managing Director, at MC Technology group of companies. He has been with Wasser High-Tech Coatings for 15 years as technical sales director and also marine manager, prior to establishing MC Technology to develop the marine and European / Middle East markets.



A NEW APPROACH FOR BALLAST & CARGO TANK COATING: A SOLVENT-FREE AND HUMIDITY TOLERANT EPOXY SYSTEM WITH EDGE-RETENTIVE PROPERTIES J Azevedo, Euronavy SA, Portugal SUMMARY The needs for anticorrosive protection are changing. New performance concerns (e.g. TSCF15, IMO Directive), emergent environmental issues, careful cost analysis and increased safety concerns are creating new trends in the market for anticorrosive marine coatings. This article describes the new needs and presents a novel tolerant epoxy technology incorporating the best trends to answer it (Low-VOC, humidity & surface tolerance, edge-retention). Results of the testing and application of this technology, both in ballast & cargo (crude) tanks, are mentioned. Some aspects of MILPRF 23236C standard are highlighted as possible contributions for the coating selection processes. A conclusion is made about how the particulars of the presented coating system may introduce important savings for both refurbishment and new building projects when trying to meet the most recent IMO / TSCF Guidelines.

NOMENCLATURE International Maritime Organization (IMO), International Association of Classification Societies (IACS), International Association of Independent Tanker Owners (INTERTANKO), International Convention for the Safety of Life at Sea (SOLAS), Tanker Structures Cooperative Forum (TSCF), European Maritime Safety Agency (EMSA), Naval Sea System Command - US Department of Defence (NAVSEA), Volatile Organic Compounds content on coatings (VOC). 1.

INTRODUCTION

The needs for anticorrosive protection are changing regardless of the field of application, leading to new trends regarding both surface preparation and coatings type. Environmental pressures, new cost factors, performance needs driven by new aggressive operation conditions and safety concerns can be answered by some emerging trends, summarised in Figure 1.

Figure 1: New trends in surface preparation and coating methods driven by changing needs


These needs (detailed below) and corresponding trends have been recognized, partially or totally, by numerous authors [4, 7, 8, 9, 10, 11, 15, 16]. The changing needs are acutely felt by the shipping industry, maybe more than by any other economic activity. Since the 90’s, increasing attention is being paid to the corrosion of ballast and cargo tanks within the tanker fleet. The first trigger for this was the EXXON VALDEZ accident, Alaska, in March 1989, followed by ERIKA in 1999, CASTOR (2000) and PRESTIGE more recently. Several organizations like IMO, IACS, INTERTANKO and SOLAS reacted to those events [14] promoting the study of corrosion causes and introducing marked changes in the regulatory environment regarding corrosion assessment and control obligations by the ship owners. The regulatory changes regarding the use of coatings for corrosion control began with IMO Resolution A798 (1994, guidelines for coating prevention systems in ballast tanks), followed by SOLAS Amendment Chapter II-1 Reg. 3.2 (after July 1st 1998 tank and bulk ships should provide hard coatings at ballast tanks). More recently IMO has started preparing Performance Standards for Protective Coatings [18] based mostly upon the TSCF Guidelines for Ballast Tank Coatings Systems and Surface Preparation published in 2002 [17], which is already being widely used for newly built oil tankers. EMSA is also addressing this issue, reporting a number of recommendations in a recent report [20]. Meanwhile, an important parallel effort driven by the same basic concerns (corrosion problems onboard ships due to poor coating systems performance) was developed by NAVSEA [7][8], resulting in upgraded coating selection and surface preparation procedures. The MILPRF Standard 23236C (2003) is the ultimate result of this work [18], implementing stringent rules for the acceptance of coating systems to be used onboard US Navy ships, specially for ballast tank protection. These rules apply not only to the performance of the coating (evaluated by standard tests) but also to specific features


such as edge-retention and VOC content of the coating and health, safety & environmental compliance of its components. One can conclude that the problem is identified and the tools to solve it are available. But the solution may be difficult to implement due to the present “business” environment. The lack of such implementation is well felt by the oil tanker owners: the estimated worldwide average corrosion cost per year due to maintenance, repairs and downtime for oil tankers only, is 2353 million US$, based on 6920 registered tankers as per Joshua [13]. This equates to an average cost per ship of 200.000 US$/year for corrosion repair and 140.000 US$/year due to downtime! Failure to assure long-lasting coating systems for ballast tanks at new building stage, means the owners are facing the prohibitive cost of refurbishment of the ballast tank coating during the service life (cost 3~17 times the cost of coating it at new building). Eliasson [15] identifies the traditional antagonistic relationship and conflicting interest between the parties involved (ship owners, ship builders and coating manufacturers) as the main problem making the correct implementation of the right solutions for the corrosion problem difficult. Despite the new rules, one can not see their reflection in the standard guarantee traditionally given by the shipyard from the delivery date: just one year. This means that the shipyard does not feel the pressure to supply an effective corrosion protection system if its application means extra production costs and delays. Result: the shipyard claims exorbitant extra costs to supply efficient and TSCF compliant ballast tank coating systems for 15 or 25 years service life and the owner fails to correctly evaluate the gains of such extra costs on future avoidable costs of refurbishment and downtime, often accepting a standard-priced ship with a sub-standard corrosion protective coating system. Lodhi [10] departed from this assessment of the current situation to further evaluate typical cost factors of a VLCC new building regarding coating. The average cost of such new ship building increased from 100 M US$ in 1995 to 130 M US$ in 2005, the total painting cost being steady at 7% of this total (around 9 M US$ in 2005). A huge slice of this painting cost (50%) is for ballast tank coating, divided between material cost (35% = 1,59 M US$) and labour cost (65% = 2,96 M US$). The bad news is the labour costs for blasting and coating, which have doubled from 1995 to 2005. Conclusions: the cost of the coating application of ballast tanks that the yard is wiling to assume (and the owner wiling to pay…) increased 30% during last 10 years but the man-hours cost increased by 100%, meaning that the total labour resources available for the coating application in ballast tanks decreased by 30% (from 22.700 man hours in 1995 to 14.800 man.hours in 2005)… The same author goes further, extrapolating for 2007 for the case that shipbuilding expected overcapacity in the near future will

depress the VLCC prices down to 94 M US$. Maintaining the same scenario (frozen 3,5% of total costs dedicated to ballast tank corrosion protection), the result will be a further reduction in resources availability, down to 8.800 man.hours, 60% less that the resources available for ballast tank coating application in 1995! The author would like to add two extra factors that are also contributing to the bad coating performance after new building: the lack of influence of the ship owner on the coating supplier selection and application control by the ship yard at new building stage. This means that specifications with too many degrees of freedom are being used by the shipyards, who may select the materials from a broad spectrum of traditional coating suppliers with me-too products. To compete in this environment, coating suppliers may be tempted to downgrade product quality to cope with depressed prices or (even worse) to be more “tolerant” (more than the coating itself...) to surface preparation and other prerequisites in order to cut delays, as a means of pleasing the yard and helping future business. In such a difficult scenario, the solution depends on all parties giving up a little of their interest, by understanding the broad picture and the un-sustainability of the present situation. A fourth party not yet mentioned (the classification societies) should also play a more active role in such an evolution. The intention of this paper is not to advise on such desirable moves, but to present an additional contribution from a particular coating system. The coating system presented below can help the adoption of the desirable new building coatings specifications for tanks such as the TSCF ones for 15 years service life with a lower impact on the costs of surface preparation and application and with diminished extra delays on the production progress. This coating system may be used for both ballast and crude cargo tanks at new building stage, and incorporates all the new above-mentioned trends. Moreover, due to its surface and humidity tolerance features, the system enables the ship owners to choose voyage repair for ballast tank refurbishment with good durability expectations, thus saving the huge downtime associated if the same job is done during dry-docking. 2.

CHANGING NEEDS

2.1

GENERAL CHANGES

The needs driving the new trends in surface preparation and coating methods can be divided as follows: •

Performance needs: - Longer service lives required. - Attention is given to previous neglected causes of failure as salt level at steel surface or coating thickness at the edges, welding seams.



•

•

•

2.2

Cost needs: - Emergence of cost-benefit analysis; - Collateral costs as environmental control costs (e.g. abrasive disposal), downtime in case of failure and delays are evaluated; - The cost of the coating itself takes a decreasing slice of overall cost. - Surface preparation costs are taking an increase slice of overall cost. Environmental Needs - New regulations regarding VOC content of coatings and abrasive use and waste disposal. - Media and social pressures towards the adoption of “cleaner” technologies. Safety Needs - New regulations regarding health & safety compliance of coatings. - Increased constraints in the use of technologies that introduce additional risk (for equipment and people) within industrial environments. SHIPPING INDUSTRY PARTICULAR CHANGES

The main consequences of the above-mentioned regulatory changes within the ship industry are: • • • •

•

Industry shift from single hull to double hull tankers. Hard coatings mandatory for ballast tank (SOLAS Reg. II-I/3.2) Mandatory coating of crude cargo tanks (bottom and top) is expected to be settled in a near future. Increasingly stringent rules regarding tank condition control (as per IACS classification). The economic advantage for the ship owners of keeping the ballast tank in GOOD condition (corrosion assessment) is becoming increasingly important (if classified as FAIR or POOR, important restrictions may apply for the vessel operation and the imposition of increased frequency of inspections will impact downtime costs). TSCF Guidelines for ballast tank coating: prequalification of coating systems, stringent rules for surface preparation (primary and secondary), salt level, film thickness, stripe coating, etc.

The shift to double hull and the augmented need for good corrosion control are contributing to an increase in importance of this issue for the owners. The factors are: • •

Double-hull VLCC tanker ballast tank areas range from 240,000 to 280,000 m2. This is about double the ballast tank area on an equivalent size single hull. Due to economical and technical reasons, the double-hull vessels are built using 80 – 90% of High Tensile Strength steel, to allow a diminution of the thickness of steel plates. A double-hull ship has only 11% weight increase when compared to an


•

3.

equivalent size single hull. The result of this new building strategy is that any steel thickness loss is more critical and the steel flexibility is higher. Another effect with consequences for corrosion due to coating failure is the dimensional changes during sailing: the thinner steel plates are much more flexible. Single-hulled tanker’s crude tanks were not coated in the past. But double-hull cargo tanks are experiencing heavy pitting problems only 4 – 5 years after new building. The reason is connected with the “thermos flask” effect of the double bottom ships and the flexibility of the steel. Warm crude cargo from Middle East do not cool as quickly as used to happen in single hull ships, due to the direct contact with colder seawater, only one steel plate away. Now, the void space between acts as an insulation that keeps the cargo warm, at optimal temperatures for microbial growth where the water is present (in the bottom and in the top of the tanks, in this case because of condensation). The microbial induced corrosion (MIC) is playing an important role. EMERGENT TRENDS

Numerous examples of authors presenting their solutions for changing needs can be cited. In 2000 Thomas and Webb [7] presented the US NAVY’s answers to face service life problems in ships, namely through the use of solvent-free edge-retentive coatings. The present version of the US military standard MIL-PRF 23236C [19], states the performance features required for painting systems to be used in ships and submarines, and represents a demanding challenge and an important upgrade of former performance requests. Meunier presented (1998) [16] the SNCF (Société National de Chemin de Fer) opinion about the best way to give economic answers to environmental control needs during steel bridge painting, using Ultra High Pressure (UHP) hydroblasting. Quintela, Silva e Leite [9], in 2002, presented the PETROBRAS vision about the needs of environmental preservation while saving costs, through a reformulation of coatings characteristics (low VOC, humidity tolerance) and the use of hydroblasting. Lodhi (2005) [10], claims that if the steel surface quality can not be improved at the time of new build due to yard reluctance, the use of a surface + moisture tolerant and edge-retentive coating instead of conventional solvent borne coating is worth the extra cost. Many other references could be shown, in agreement with the conclusions and results of the examples listed above. In summary, to cope with the new needs, some key new trends on both the coating and surface preparation activities have been observed: the increasing adoption of Ultra-High Pressure water jetting (as defined by SSPC SP12 standard) as surface treatment method, the increasing use of low VOC coatings (e.g. solvent-free or Ultra-High Solids epoxies) and the attempts to increase the edge-retentive properties of the coatings.


4.

CHALLENGES

Answering the new needs in the way outlined by Figure 1 may lead to new problems. Each option carries advantages but also some drawbacks that should be correctly identified and solved. 4.1

HYDROBLASTING

The use of hydroblasting has a number of advantages, namely a lower salt level content on the prepared surfaces, a lower environmental impact, decreased safety risks (compared with abrasive blasting) and less countereffects on the original surface profile (see Figure 2) or on adjacent coated areas. Due to safety concerns, hydroblasting is much more compatible with ongoing industrial activities, when compared with abrasive blasting. Nevertheless, some drawbacks can be associated with this surface preparation, mainly the dry surface requirements vs. flash rust dilemma and the reduced performance of the protection compared with traditional coatings over Sa 2 ½ abrasive blast standard [11]. Conventional coatings are the main adversaries for a larger adoption of this preferable surface preparation technology. There are numerous reports of maintenance jobs using hydroblasting, under high humidity, where the absence of moisture tolerance of the coating can lead to increased waiting times and often to excess flash rust and dangerous salt levels when the surface finally dries. Reblasting is then necessary and the cycle may repeat forever... A possible outcome is giving up the job. Or carrying out the application anyway, therefore crossing the red line, delivering reduced coating performance.

Figure 2: Hydroblasted surface without salts and conserving the original profile. The salt level issue should be looked at carefully. In fact, immediately after hydroblasting, a SC-1 condition (surface free of detectable salts, as per SSPC SP12 standard regarding non-visual contaminants) can be achieved. But under marine conditions, salt contamination of the surface will occur quickly, risking surpassing the SC-2 condition (7 micrograms/cm2 chlorides) after drying. A preferable chloride level below

3 or 4 micrograms /cm2 is much easier to obtain if the coating application is possible a short while after blasting or washing, without need for extensive drying. 4.2

EDGE-RETENTION

Coatings with increased edge retentive properties allow a better protection of critical areas, such as welding seams, stiffener edges, etc [1]. The advantages of such coating’s ability are especially important on ballast tanks and complex steel structures (e.g. steel bridges). A typical solvent-based coating, applied on a 90º edge by airless spray, will experience an after curing dry thickness reduction on the edge, dropping to only 20 to 30%, compared with the adjacent flat surfaces. A good edgeretentive coating should present a ratio superior to 70% (see Figure 3), as per US Navy / NAVSEA requirements for long service life ballast tank coatings (20 years) [7][19]. With such coatings, a trade-off is possible between extra performance and stripe coating needs (which can be reduced). On the new building side, edge grinding is mandatory as per TSCF Guidelines and other steel preparation specifications. This costly operation may also be reduced without critical impact on the performance, if an edge-retentive coating system is used. Normally, edge-retentive coatings have very high viscosities, low pot-life and may cause airless application problems. These drawbacks may imply the need for plural-airless equipment to apply the product. 4.3

SOLVENT-FREE EPOXIES

Epoxies are one of the chemical families of coatings more adaptable to solvent-free formulations (meaning Ultra High Solids with VOC ranging from 0 to 150 g/L). Solvent-free epoxies are environmentally preferable (reduced VOC emissions). A number of technical advantages when compared with solvent-based epoxies are generally observed. Solvent-free epoxies reduce the problems arising from solvent retention in applied coating films, are more tolerant to over-thickness situations, allow a reduction in the number of coats of the coating system, present better cohesive and tensile strength resistances and better appearance (glossy surface, easier to clean). Today’s typical solvent-free epoxies on the market also have some drawbacks: reduced pot life and extremely high viscosities are the main ones affecting easy application. Solvent-free epoxies tend to be applied with excess thickness, given the application difficulties, with negative impact on costs (paint consumption) and curing. A majority of solvent-free epoxies on the market are not surface tolerant and are dew-point restricted, thus making the option for hydroblasting problematic.



4.4

INTEGRATING NEW TRENDS

The ideal solution to answer the new needs, in order to avoid the above drawbacks of the individual approaches, would be a coating system incorporating ALL the new trends. This system should be solvent-free, tar-free, tolerant to humidity and to surface preparation / flash rust, user friendly (pot-life, viscosity, curing profile) and have good performance (allowing long life protection over hydroblasted surfaces, similar or superior to that achieved using traditional technologies over abrasive blasted Sa 2 ½ surfaces). Moreover, it should present good edge-retention properties and be formulated without any dangerous or undesirable component such as heavy metals, chromates, zinc, etc. EURONAVY have developed a solvent-free epoxy technology that combines all these aspects. 5.

THE SOLUTION

5.1

THE ES301 COATING SYSTEM

In the beginning of the 90’s Euronavy created the first version of a novel epoxy system (branded ES301) that combines a majority of the new trends (including the solvent-free and humidity tolerance features). The system was further optimised during a decade and as been used by respected customers such as US NAVY, SNCF, PETROBRAS, TRANSPETRO and NITC, some of the companies that were more aware of the need for change. The edge-retentive feature was added in 2000 and the result (ES301S version) was presented by M. Paiva and J. Martins during SSPC 2000 Conference [1]. The main features of this product family and the respective system characteristics are presented below. The system is now being increasingly used, namely for offshore equipment (maintenance, FPSO conversions and new building), marine industry (mostly for ballast and cargo tank protection, both for new building and refurbishment at dry dock or voyage repair), industrial sites and steel bridges. 5.2 • • • • • • • • • • • •

5.3 •

• • • • •

SYSTEM CHARACTERISTICS Surface preparation standards: hydroblasting to SSPC SP12 WJ2 (WJ2-M visual standard as per SSPC Vis4 – NACE Nº7, MEDIUM flash rust); SSPC SP10/ISO 8501 Sa 2 ½, SSPC SP2/SP3/SP11. Typical DFT per coat: 100 – 150 microns Same product type for priming / intermediate / finish If extra UV resistance is required, the system may include a classic polyurethane topcoat 2 or 3 coat systems, ranging from 200 to 450 microns total DFT, depending on the areas to be protected, corrosiveness and service life target. No need for dehumidification in enclosed spaces such as tanks and voids.

The use of the novel system itself allows extended protection, as can be seen below (see Results). The high adhesion values, the solvent-free feature without application drawbacks, and the edge-retention abilities are the basis of this superior performance. Example: US Navy / Navsea [21] expects a service life of 20 years for ballast tanks coated with this system, over SSPC SP10 / ISO 8501 Sa 2 ½ abrasive blasting, with a total DFT (two coats) of 10-14 mils (250-350 microns), with one stripe coat only.

Figure 3: Edge-retentive performance of ES301

PRODUCT FEATURES Modified epoxy, polyamine cured Solvent free Tolerant to damp surfaces No dew point restrictions Tolerance to adherent iron oxides High compatibility with old coatings or shop primer. Strong adhesion to steel (> 12 MPa pull-off). Chemical interaction with steel, oxides and moisture Steel profile is not a critical factor. Friendly pot-life (up to 3 hours at 25ºC, depending on version) Applicable by brush, roller or airless (60:1) “edge retentive” version (ES301 S) available.


Figure 4: Humidity and surface tolerance: ES301 priming over UHP hydroblasted surface, flash rust M.


Table 1 lists relevant performance indicators concerning ES301 system. The increasing adoption of the ES301 system is driving developments towards added capabilities, given the extension of its use on new building. One of these added capabilities is the certified compatibility of the system with welding operations. A primer coat of 75 microns of ES301 was certified has having no influence on the welding process (SGS Cert. 1201/10202). Another recent additional feature is the possibility of supplying ES301 for tank coating with UV-sensitive pigmentation to allow dark-light inspection. Figure 5: High adhesion (pull-off 20 MPa, during FPSO conversion works in Singapore) 6.

RESULTS

The system has been used by a number of respected customers over the last years. Significant track record can now be shown as result of customer-based certification or approval processes and inspections during service life. The system is approved by NAVSEA / US Navy, and is qualified per MIL-PRF-23236C standard [19]. Under this standard, ES301 is qualified as Type VII, Classes 7 (seawater ballast tanks), 15b (can be applied over wet surfaces prepared to bare metal) and 17 (bilges). The qualification as Type VII means that ES301 is recognized as (quote) a formulation with no solvent added, VOC < 150 g/L, absence of pigments that are hazardous or create hazardous waste above trace levels and the dry coating is not a hazardous waste under USEPA regulations (unquote). Up to now (December 2005), the ES301 system is the only one being approved for 15b class (tolerance to humidity).

Figure 6: ES301 primer and stripe coat during crude cargo tank hydroblasting and coating job. A two-coat system with 300 microns DFT is used for this purpose.

PETROBRAS, the Brazilian oil company, approved ES301 system as main coating system for off-shore new building or conversion projects. TRANSPETRO, Petrobras’ owned oil tanker company, is also using ES301 system for both ballast tanks and crude tanks refurbishment, normally over UHP hydroblasting. Regarding Health & Safety, other relevant features should be mentioned. ES301 system is found suitable for safe use by the Navy Environmental Health Center (NEHC) as per MIL-PRF 23236C rules [19] (meaning absence of heavy metals or other hazardous pigments above trace levels). On the safety side, ES301 two-coat systems were tested at independent NGC Labs (Fire Testing Laboratory). The flame spread index of ES301 systems ranged from 15 to 20 (0 – 100 scale) and the smoke developed index from 300 to 450. Accordingly, the ES301 systems was rated Class “A” (the most demanding) as per National Fire Protection Agency (NFPA) standard Nº101, meaning a coating with a good resistance to flame spread and limited smoke development while burning.

Figure 7: edge-retention is a key feature for the durability of tank protection.



TEST

RESULT

Salt fog ASTM B117

Rating 10 (0-10, ASTM D1654) after 1000 h.

(A)

> 2000 h no defects

(B)

Rating 10 (0-10, ASTM D1654) after 1000 h.

(A)

> 2000 h no defects

(B)

Combined weathering test (NACE TM0184)

4000 h no defects

(D)

Adhesion (pull-off, ASTM D4541 or equivalent)

System applied over Sa 2 ½ or WJ2 standards

After application and curing

12,0- 24,0 MPa

(B)

After 1000 h salt fog

9,3 – 10,8 MPa (ES301K+ES301S)

(A)

After 700 h salt fog

10,0 MPa (ES301L+ES301S+PU) NF EN 24624

(C)

After 1000 h condensation

11,5 – 13,8 MPa (ES301K+ES301S)

(A)

Atmospheric exposure (2,5 years)

Ratings 0-10 accordingly to ASTM D1654

(A)

Condensation ASTM D4585

SOURCE

Rust: 10; Blistering: 10; Scribe undercut 0,5 mm. Cathodic disbondement (MIL P24647, ES301K+ES301S)

No defects (90 days, “pass”).

(A)

Choc resistance (falling weight) (NF EN ISO 6272)

6,4 – 8,3 J (fall from 65 to 85 cm) (ES301L+301S+PU)

(C)

Fire resistance ASTM E84-01

Rating A (NFPA Nº101, evaluating flame spread and smoke liberation)

(E)

Edge-retention (procedure of MILPRF 23236 C standard)

% retention 74% - 101%, for edge radius between 0,1 mm and 2,4 mm, respectively. (ES301K+ES301S system)

(A)

(A) (B) (C) (D) (E)

Naval Research Laboratory, Center for Corrosion Science and Engineering, US Navy. CENPES/PETROBRAS, Centro de Pesquisas e Desenvolvimento Leopoldo A.M. Mello SNCF, Société National de Chemin de Fer (Eurailtest Laboratoire de Vitry) DNV Preliminary Report on Type Approval NGC Testing Services, Fire Testing Laboratory

Table 1: Test Results of ES301 system. An important complement to the system is an organic epoxy shop-primer (branded PE31) using the same resincuring agent system as ES301. This shop primer is Type Approved by DNV (certificate Nº K2751) as compatible with welding. The combination of PE31 + ES301 is being used for some years now at off-shore new building and conversion projects, using UHP hydroblasting (damaged spots) and HP (700 bar) washing of sound shop primed surfaces as secondary surface preparation. The durability is similar to that of ES301 system applied directly to bare steel, given the ES301 surface and humidity tolerance abilities, the full compatibility of the resin types and the consequent very high adhesion of complete PE31+ES301 full system over steel. 5.

USING ES301 TO COPE WITH TSCF GUIDELINES

ES301 coating system joins performance and tolerance in an unique way. The specific features of this system may now be well perceived by the market and can be proved


by the numerous success cases within the marine market (for maintenance and new building) and by extensive approval and test report data. How can both owners and shipyards use these features to meet emergent rules for tank coating in an economic and successful way? This document presents the first tentative answer: a draft specification for a typical 15 years service life using TSCF15 Guidelines as departure point and adding some additional options to fully utilise the ES301 system potential. It is the author’s opinion that the data presented above gives enough indication about ES301 ability to be applied cost-effectively, with important savings regarding steel (edge grinding) and secondary surface preparation, without affecting the 15 years service life targeted by TSCF Guidelines. This ability is possible thanks to the added surface and humidity tolerance and edge-retentive features that enabled the MIL-PRF 23236C approval.


Table 2 presents a suggestion about how a specification for 15 years service life, following basic TSCF15 requirements, may be adapted, to include more economic and less time consuming secondary surface preparation, edge grinding and stripe coats when selecting ES301 as coating system, together with PE31 organic shop-primer. ITEM

Table 3 details the differences, justifies the adoption of alternative methods with the specific ES301 features and summarises the advantages of each adaptation made, compared with typical TSCF15 procedures when using conventional products.

REQUIREMENT

COMMENTS

PRIMARY SURFACE PREPARATION BLASTING

Abrasive, Sa 2 ½

ISO8501, 4.1-4.6

PROFILE

30 – 75 microns

ISO8503-1/3

SOLUBLE SALTS

< 30 mg/m2 (chlorides)

ISO 8502-9

PRE-CONSTRUCTION PRIMER

EURO-shop PE31, 25 microns

Organic epoxy using the same resin-curing agent system than ES301

SECONDARY SURFACE PREPARATION STEEL CONDITION

P1 Grade (one pass edge grinding)

SURFACE PREPARATION

UHP Hydroblasting to WJ2 on damaged areas. Moderate flash-rust level accepted.

ISO 8501-3 SSPC SP12 / SSPC Vis4

HP washing (> 700 bar) of other areas with intact shop primer. SOLUBLE SALTS

< 30 mg/m2 (chlorides)

ISO 8502-9 (NOTE : using water as secondary surface preparation means that the salt level is more easily achieved)

DUST

“1”

ABRASIVE INCLUSIONS

None

ISO 8502-3 Using water as secondary surface preparation means that this control is not needed any more.

COATING 1st coat

ES301K62 (red oxide colour, with luminescent additive as option)

150 microns DFT

Stripe coat

ES301S brush applied

150 microns DFT

2nd coat

ES301S00 (light color)

150 microns DFT

Table 2: ES301 system specification for 15 years service life as per “adapted” TSCF15 Guidelines



ITEM

ES301 System

Typical TCSF 15 approach

Observations & references

Shop Primer

May use the organic epoxy PE31 shop primer.

Zinc ethyl silicate

Using PE31 means cost savings (shop primer cost per sq m.) and lower failure risk (inorganic zinc is very sensitive to application). The conjugation PE31/ES301 is already being used in several conversion and new building projects.

Secondary surface preparation

Allow UHP hydroblasting

Demands Sa 2 on all areas (Sa 2 ½ damaged spots)

OUTPUT OF CHOOSING ES301 PRODUCT COST SAVINGS SAFER PRODUCT APPLICATION

PE31 shop primed surfaces can be over HUGE SAVINGS ON coated with tolerant ES301 system without 2nd SURFACE the need of removing it by abrasive PREPARATION blasting. EURONAVY procedure accepts COST UHP WJ2 (SSPC SP12) on damaged spots and intact areas preparation by HP water SAVE WASTE jetting (> 700 bar). This option is possible DISPOSAL COSTS due to the humidity and surface tolerance (abrasive) features of ES301, and has been used in several shipyards in Singapore and Brazil. ENVIRONMENTALLY FRIENDLIER

EASIER SALT LEVEL CONTROL Steel preparation (edges)

Accepting P1 (one pass grinding)

Demands P2 grade (three pass grinding)

ES301 system is approved by MIL-PRF 23236C from US Navy as Type VII (no solvent added) and by classes 7 (ballast tanks), 17 (bilges) e 15b (no dew-point restricted, may be applied over wet bare steel). This approval means that the Edge Retention Ratio is > 70% with a 1 mm radius edge: P1 is quite enough to assure good thickness over edges and welding seems.

SAVING TIME & COST SPENT FOR EDGE GRINDING

Stripe coats

One stripe coat, between the 2 coats, is enough to assure edge thickness.

Two stripe coats mandatory for 2 coat systems.

ES301 advantage is a result of the edgeretentive behaviour (approved by MILPRF 23236C). Among the few systems with such approval, ES301 was accepted with minimum 300 microns DFT (as TSCF15 requirements). All others have to be applied to a minimum 350 microns DFT. Moreover, among the 23236C Type VII approved systems, ES301 is the only one tolerant to humidity (Class 15b).

COST SAVINGS ON MAN POWER

Painting

Without dewpoint restrictions.

Dew-point restrictions.

MIL-PRF 23236C Class 15b approval means that ES301 may be applied over wet surfaces without dew-point restrictions. As solvent-free product (Type VII of MIL 23236C, no-solvent added, VOC < 150 g/L), can be considered environmentally friendlier and with higher performance

NO DELAYS DUE TO HUMIDITY CONSTRAINTS

Low-VOC, friendly pot-life

Durability

20 years as per US Navy PPIs

15 years

ES301 system is specified by US Navy / NAVSEA for ballast tanks as 20 years service life system. Petrobras choose ES301 system for 25 years dry-dock-less offshore projects.

SAVES DEHUMIDIFICATION

COSTS SUPERIOR PERFORMANCE

Table 3: Advantages and justification of the ES301 approach to TSCF15 Guidelines



6.

CONCLUSIONS

Ship owner’s and shipyards perception of the value of top quality tank coatings for long service life is changing. The perceived added value of such coatings, compared with traditional non-tolerant solvent-based solutions, is increasing. If such a coating system, together with the performance added value, can be applied with reduced impact on the cost of steel and surface preparation and application delays, it may constitute a reasonable “agreement opportunity” for both owners and yards interested in answering the current challenges and avoid the potential dangers of conflicting interests. Despite the focus of this article on the coating specifications for tanks at new building stage, a straightforward additional conclusion should be highlighted: the high-performance + tolerance features joined together in ES301 is also helping the adoption of this coating system as cost and performance effective tool for ballast and cargo tank refurbishment. The advantages are especially significant for dry-docking repairs using hydroblasting or for voyage repair when a fair extension of service life of the ship (> 10 years) is targeted. This particular field of ES301 application, including data from the biggest ever voyage repair ballast tank refurbishment project (being carried out since 2003 using this system) may contribute for a separate article to be published in the future. A final mention should be made regarding the cost of a solvent-free, edge-retentive and surface + moisture tolerant system as ES301. Looking at the specific case of a 300 microns system for 15 years service on ballast tanks, such system (material cost only) will cost about 10 US$/m2. A conventional good quality solvent-borne modified epoxy system for the same application costs aprox. 6 US$/m2 [10]. Taking into consideration the prohibitive cost (as explained above) of coping with TSCF Guidelines using a conventional coating system (extra edge grinding, secondary surface preparation to Sa2, extra stripe coat and the correspondent additional delays), the gap of 4 US$/m2 may look quite attractive if it allows the adoption of the “modified” TSCF15 specification as per Table 2. The other option is NOT to cope with TSCF Guidelines or other reliable specification to assure 15 years service life. In this case, the expected result in terms of coating repair and downtime costs during the ship’s service life is incomparably higher than a mere 4 US$/m2 on the CAPEX expenditure at new building stage. 7.

8.

REFERENCES

1.

M P PAIVA, J MARTINS, ‘An Edge-Retentive Coating Solution Based on a Tolerant Solvent-Free Epoxy System’, SSPC2000 Industrial Protective Coatings Seminar, Nashville, Tennessee, November 2000.

2.

M P PAIVA, J MARTINS, ‘An Edge-Retentive Coating Solution Based on a Tolerant Solvent-Free Epoxy System’, Protective Coatings Europe, Vol.7, Number 6, June 2002.

3.

J AZEVEDO, ‘Protecting Offshore Investments against Corrosion with Innovative Epoxy Technology’, OVERFLATEDAGENE 2003 The Surface Protection Conference & PCE Marine and Off-Shore Conference, Stavanger, Norway, November 2003.

4.

J AZEVEDO,‘A new approach for steel structures protection using UHP hydroblasting and a solventfree and humidity tolerant epoxy system with edgeretentive properties’, EUROCORR 2005 The European Federation of Corrosion Conference, Lisbon, September 2005.

5.

SSPC/NACE Joint Surface Preparation Standard SSPC SP12/NACE Nº5 ‘Surface Preparation and Cleaning of Steel and other Hard Materials by Highand Ultrahigh- Pressure Water Jetting Prior to Recoating’.

6.

MTTC, ‘Project Book Fleet Maintenance Reduction Program, Project A8 – Epoxy Bilge Paints’. Issued by McConnel Technology & Training Center, Spring 2005.

7.

E. D. THOMAS, A. A. WEBB, ‘World Class Tank Coating Materials, Practices, Procedures’, in: SSPC2000 Industrial Protective Coatings Seminar, Nashville, Tennessee, November 2000.

8.

A. A. WEBB, B BRINCKERHOFF, ‘Reducing Navy Fleet Maintenance Costs with High Solids Coating and Plural Component Spray Equipment’, JPCL

9.

J. P. QUINTELA, A. T. M. SILVA, P. B. LEITE, ‘Ecological Paint Systems – The Petrobras View’, Corrosao e Proteccao de Materiais, Vol. 21 (3) Jul/Aug/Sept 2002

ACKNOWLEDGEMENTS

The author would like to thank Brian Goldie and Anwar Lodhi for the important help given reviewing this paper and for the helpful suggestions, both from the editorial and technical point of view.

10. A LODHI, ‘Double Hull Tankers Ballast Tank Maintenance – Experience Feed-Back’, The 22nd DNV Technical Committee Meeting, Dubai, 7th December 2005. 11. P. LE CALVÉ, P. MEUNIER, J. M. LACAM, ‘Evaluation of Reference Paint Systems after UHP



Water Jetting’, Protective Coatings Europe, Vol.8, Number 1, January 2003. 12. K.B. TATOR, ‘Risk Assessment and Economic Considerations when coating Ballast Tanks’, April 2004 13. J T JOSHUA, ‘Cost of Corrosion – Appendix O – Ships’, US Department of Transportation, Federal Highway Administration Report, 2001 14. INTERTANKO, ‘Crude Oil Tanker Cargo Tank Corrosion – Update’, Cardiff, 24th October 2002. 15. J ELIASSON, ‘Coating Ballast and Cargo Tanks on Ships: a Status Report’, Journal of Protective Coatings and Linings, July 2005 16. P MEUNIER, ‘The Experience of SNCF in Preparing Previously Painted Metal Surfaces by UHP Water Jetting’, Protective Coatings Europe, Vol.3, Number 9, September 1998. 17. TSCF, ‘Guidelines for Ballast Tank Coatings Systems and Surface Preparation’, Witherbys Publishing, 2002 18. IMO, ‘Performance Standards for Protective Coatings – draft’, 48th Session of sub-committee on Ship Design and Equipment, 19th November 2004. 19. NAVSEA, ‘Performance Specification Coating Systems for Ship Structures’, US Military standard MIL-PRF 23236C from 12 August 2003. 20. EMSA, ‘Double Hull Tankers: High Level Panel of Experts – Report’, June 3rd 2005 21. NAVSEA “Preservation Process Instruction for Ballast Tanks, CHT Tanks, Compensating fuel tanks, high traffic interior decks (abrasive blasting)”, NAVSEA document PPI 63101-001H Rev 06, March 10 2004 9.

AUTHOR BIOGRAPHY

Joao Azevedo holds the current position of Sales & Marketing Director at EURONAVY SA. He is responsible for the technical assessment of market needs and for the communication of EURONAVY developed solutions to cope with those needs.



THE EFFECT OF A FOUL RELEASE COATING ON PROPELLER NOISE AND CAVITATION R Mutton, M Atlar and M Downie, University of Newcastle upon Tyne, UK C Anderson, International Paint Ltd., UK SUMMARY In the search for increasing efficiency and reduced costs a number of ship operators have turned to coating their ships propeller to prevent fouling build-up and reduce cleaning costs. In support of this the University of Newcastle upon Tyne and International Paint have an ongoing investigation into the effects of foul release type antifouling coatings on marine propellers. To date over 180 full-scale marine propellers have been coated using foul release coatings. The results have shown that propeller coatings can give fouling free performance for in excess of 36 months. In addition a major feature of feedback from the ship operators has been reports of reduced noise and vibration after the application of the coating. To investigate if these claims could be substantiated a series of tests using the Emerson Cavitation Tunnel at the University of Newcastle upon Tyne was conducted for the first time. This paper reports on the early results of an experimental investigation into the effects of a foul release coating upon the noise and cavitation generated by a scale model of a commercial marine propeller with the aim of quantifying the extent of these effects once the coating has been applied. The propeller model was selected to be representative of an actual full-scale coated propeller that is typical of coating applications to date. NOMENCLATURE D = Propeller diameter g = Acceleration due to gravity H = Propeller shaft immersion J = Advance Coefficient = V/ (ND) n = propeller rotational speed P0 = Atmospheric pressure PV = Vapour pressure of the water r = Distance between the propeller shaft and hydrophone R = Propeller radius SPL = Sound pressure level V = Propeller inflow velocity ∆f= Bandwidth for each 1/3 octave band filter ρ = Density of the tunnel solution =1026kgm-3 σ = Cavitation number 1.

INTRODUCTION

1.1

PROPELLER COATINGS

With the recent increases in fuel costs worldwide, ship operators are looking for increasing methods to improve the efficiency of their ships. In this endeavour, most of the effort is rightly placed upon keeping the hull free from fouling. However more and more attention is now being placed upon maintaining the propeller condition. In terms of energy lost per unit area the propeller is of greater importance than the hull [1]. This means that a large saving can be made for little investment. The traditional method for protecting marine propellers is a regular polishing [2]. Polishing however, is both expensive and time consuming. It also has the disadvantage that the initial stages of fouling (a biological conditioning film) will start to recolonise the blade surface within 24 hours of it being polished [3].


From this the condition of the propeller rapidly deteriorates. With this in mind a collaborative research programme was launched between the University of Newcastle upon Tyne and International Paints Ltd. to investigate if the use of a modern antifouling coating system would be beneficial in helping the prevention of fouling on the propeller and the associated increases in performance. The antifouling system used is a modern non-toxic antifouling system referred to as a Foul Release system. The system top coat works by being principally composed of a PDMS (Polydimethylsiloxane) elastomer. The PDMS molecule has a long flexible backbone that has been shown to give a very low surface energy [3]. The surface energy is the property that principally controls the strength by which fouling organisms are able to bond to a surface. The PDMS was found to have a surface energy that permitted fouling organisms to attach with a strength that is an order of magnitude less than on other surfaces [4]. In addition to their proven antifouling ability, silicone based foul release systems have been shown to provide a reduction in frictional drag of between 2% and 23% depending upon the quality of application, when compared to a Tin-free SPC (Self Polishing Copolymer) system [5]. The roughness of the silicone Foul Release systems is considerably different when compared to the Tin-free SPC (Figs. 1 and 2). The roughness amplitude of the silicone systems is reduced and in addition the wavelength (texture) of the roughness is much longer. The silicon based Foul Release system has what is termed an ‘open’ texture whilst the Tin-free SPC system has a ‘closed’ texture. Detailed boundary layer measurements conducted at the Emerson Cavitation


Tunnel at Newcastle and the CEHIPAR tunnel in Spain, using LDA measurements showed that the friction velocity for Foul Release surfaces is significantly lower than for Tin-free SPC surfaces and that at the same streamwise Reynolds number the ratio of the inner layer to the outer layer of the boundary layer is smaller for the Foul Release surfaces. The inner layer is the part of the boundary layer where major turbulence (and hence drag) production occurs. This lead the Foul Release surfaces to have significantly lower roughness functions when compared to the Tin-free SPC coatings [6].

by hand between the propeller being assessed and the standard surfaces on the gauge. The surfaces range from an average roughness amplitude Ra= 0.65m to an amplitude of Ra=29.9m. A (smoothest) and B represent the surface roughness of new or reconditioned propeller blades while the rest (C, D, E and F) are from propellers eroded after increasing periods of service. The computer simulations [7] showed that for a propeller from a medium sized tanker (the basis vessel for the model tests reported in section 3 of this paper), the gains in efficiency in moving from a rough propeller vary from 3% (Rubert D) to 6.5% (Rubert F) at the design operating point, in this case an advance coefficient, J =0.48 (figure 3). This increase in efficiency was due to a reduction in the frictional resistance of the blades with no corresponding change in thrust.

Figure 1: A roughness profile of the Foul Release system.

Figure 2: A roughness profile of the SPC system.

1.2

EFFECT OF THE PROPELLER COATING

The application of a Foul Release coating to a marine propeller is perceived to have two distinct methods for improving the performance as described in the following. The first is a short term benefit, similar to propeller polishing, from moving from a roughened and fouled propeller to a clean and well coated propeller. Initial computer studies have shown that the application of a foul release coating to a propeller would have the equivalent drag of a new or well polished propeller [7]. The increased efficiency in moving from a roughened propeller to a coated one depends on the roughness of the propeller prior to coating. A quick and simple method for assessing the roughness of a propeller is to use a Rubert gauge. The gauge consists of six examples of actual propeller surface finishes (A – F). A comparison is made

Figure 3: The gain in efficiency in going from a specified Rubert surface to the coated propeller for a medium sized tanker. The second benefit is longer term from the prevention of fouling building back up once the propeller is back in service. This benefit does not occur with propeller polishing. In June 2003 the propeller of the Newcastle University, School of Marine Science and Technology research vessel, Bernicia, was coated in preparation for a series of sea trials [9]. Although the sea trials proved inconclusive, as shown in Fig 5 due to poor weather, in measuring the short term increase in performance due to the application of the coating, in the first two years since the trials took place, the propeller has been inspected at both 12 and 24 months [10]. The coating was found to be in good condition, 95% intact, except for slight removal of the



coating at the edges and tip of the blades. The results have shown that despite the vessel operating in a heavy fouling, coastal and estuarine environment, little fouling has returned to the propeller. What fouling has returned is a light ‘slime’ layer that is easily removed with a damp cloth. This is very different for the uncoated propeller where after 14 months in service after a polish, barnacles were present to about half the blade radius. Roughness measurements that were conducted on the propeller both when the coating was freshly applied and at each inspection showed little change in either the roughness amplitude or the roughness texture over the period of service. Figure 6: The propeller of Bernicia after 14 months in service before coating. Hard shell fouling is present to half the blade radius.

Figure 4: The Research Vessel Bernicia Overall Length Metres 16.2 Beam Metres 4.72 Draft Metres 2.59 Gross Tonnage Tons 46.25 Service speed Knots 8.0 Deck Space Sq. Metres 20 Endurance at sea Days 3 Table I: The general particulars of RV Bernicia

Figure 7: The propeller of Bernicia after 12 months in service after coating. 95% of the coating is intact, except some detachment of the blade edges. Light slime fouling is present on the inner half of the blades (grey material on the red coating is the dried biofilm).

Diameter 1.14m Blade Area Ratio 0.466 Number of Blades 4 Maximum Rpm 440 Table II: Details of the Bernicia propeller

Figure 5: The final Results of the Bernicia sea trials show no statistical difference between the two curves. The trials were particularly effected by the weather leading to large error estimates and making the results inconclusive.


Figure 8: The propeller of Bernicia after 24 months in service after coating. Again 95% of the coating is intact, little difference can be seen between 12months and 24 months. Light slime fouling is still present on the inner half of the blades (grey material on the red coating is the dried biofilm).


To date coatings have been applied and in service for over 37 months. Figures 9 and 10 show the full scale coating on the propeller that is the basis for the model tests described later in this paper. They show that about 90% of the coating is still attached to the propeller despite some losses round the tip and the leading and trailing edges. It can also be seen that little or no fouling is present on the blades (the prop had not been cleaned during the 37 months in service). The coating has been able to prevent the build up of fouling and its associated increases in drag for extended periods.

1.3

PROPELLER NOISE

In addition to the savings gained by the prevention of fouling drag, further saving are achieved by the reduction in the number of occasions that a diver has to be used to clean the propeller. In high fouling environments it has been seen that some slime fouling can attach to propeller on the inner half. To maintain the propeller at maximum efficiency it is still recommended that the propeller is still periodically cleaned by a diver. This however would be at a reduced frequency and would not involve any hard abrasive cleaning (a damp cloth is generally sufficient to remove slime fouling from the Foul Release Coatings).

There are four principal mechanisms by which a propeller can generate sound pressures in water [11]:

In addition to the benefits described above, a number of operators have reported that after the application of the Foul Release coatings to their ship’s propellers, a reduction in propeller generated noise was observed. In order to substantiate these claims a series of noise measurements were taken using the Emerson Cavitation Tunnel propeller test facility in the School of Marine Science and Technology at the University of Newcastle upon Tyne (see section 2).

• • • •

The displacement of the water by the blade profiles. Immigration of flow from the pressure to the suction side of the blades in developing thrust. Fluctuating volume of cavitation on the blades when cavitation develops on the blades of propeller operating in non-uniform wake flow. Collapse of cavitating bubble and/or bursting of a cavitating vortex.

Of these four mechanisms for generating propeller noise the first two are associated with “non-cavitating” propeller flow while the latter two with the “cavitating” flow.

Figure 9: The full-scale propeller after a period of 37 months in service.

Figure 10: the back of the full-scale propeller after a period of 37 months in service.

The non-cavitating component of sound pressures will have distinct tones – known as the blade rate noiseassociated with discrete (lower) blade frequencies together with a broad-band noise at higher frequencies. The blade rate noise is closely associated with the unsteadiness caused by circumferentially varying wake field in which the propeller operates. This causes a fluctuation in the angle of attack of the propeller blade sections and hence sound pressure. However this can hardly be affected by the presence of the coating. On the other hand the broad-band noise is mostly affected by the level of turbulence in the incident flow and its interaction with the wall boundary layer which will be affected by the coating. One of the important mechanisms contributing to the broad-band noise is the trailing edge noise, which is perhaps the least well understood mechanism. The role of the turbulence in the boundary layer is a crucial parameter, which will be affected by the presence of coating, while this noise component would suffer from the effect of possible fouling with uncoated propeller as well as from hydro-elastic effects. The collapse of cavitation bubbles creates shock waves and hence cavitation noise. This is manifested as mostly ‘white noise’ in a frequency band up to around 1MHz. It is thought that the coating will mostly affect the trailing edge noise and may even act as a damper by absorbing the energy of cavitation noise due to its flexible nature. The effect of the coating on cavitation will be discussed in more detail in section 5.



Figure 11: An idealised non-cavitating noise spectrum. Taken from [11] 2.

THE TEST FACILITY

The tests have been conducted using the Emerson Cavitation Tunnel within the School of Marine Science and Technology at the University of Newcastle upon Tyne. The facility consists of a 60 tonne water capacity, vertical plane, enclosed, re-circulating water tunnel and the associated control and measurement systems. The measuring section of the tunnel is 3m x 1.2m x 0.8m. The measurement system to be used for these tests consists of a Kempf and Remmers H45 dynamometer mounted to the lid of the measuring section. This enables measurements to be taken up to 3000N Thrust and 150Nm Torque. Speed of rotation for the propeller is up to 3000rpm. Water speed is generated by a 4 bladed impeller mounted in the lower arm of the tunnel generating tunnel speeds up to 8ms-1. A vacuum can be applied to the tunnel allowing it to test at scaled pressure conditions. The layout of the tunnel can be seen in Fig. 12. Other testing facilities available include digital noise and pressure measurements, high speed photography and laser velocity measurement systems (LDA/PDA and PIV). A complete description of the facility can be found in [12]. Noise measurements have been recorded using a Bruel and Kjaer type 8103 miniature hydrophone mounted in a water filled, thick walled, steel cylinder placed on a 30mm thick plexiglass window above the propeller at a vertical distance of 0.405m above the shaft centreline of the dynamometer. The signals from the hydrophone were collected by further Bruel and Kjaer equipment, in this case a PC based Pulse digital acquisition and analysis system. This system uses a sample length of 1 second and can produce both pure FFT spectrum data and 1/3 octave Constant Percentage Bandwidth (CPB) averaged data in real time and up to a frequency of 25KHz. The data is recorded against a reference noise level of 1Pa. 3.

details of the basis vessel and its propeller are shown in Table III. This model is to be used as it is representative of typical modern merchant propellers, which are both the most numerous and seen as the most likely candidates for coating application at full-scale. The model has been constructed from aluminium to a scale of 1:19.57 (this unusual ratio was used to provide a model with the maximum diameter feasible for the tunnel) so that multiple sets of blades, manufactured with great accuracy can be installed or replaced quickly and easily. This allows rapid and reversible changes between the coated and uncoated condition. The blades used in these tests were previously tested before coating and were shown to be statistically similar. The particulars of the model can be seen in Table IV.

THE MODEL PROPELLER

The model to be used for the tests is a scale model of the propeller of an existing medium sized tanker of about 100,000dwt. The vessel has been under surveillance to investigate the performance of its coated propeller. The


Figure 12: The Layout of the Emerson Cavitation Tunnel at the University of Newcastle upon.

One set of blades was coated with the Foul Release system. This is a three layer system consisting of an epoxy base coat, a silicon polymer topcoat and a tie coat between these two in order to facilitate a good bond between the epoxy and the silicone. The whole system dries to a film thickness of between 320 and 360m. The uncoated and coated propeller model can be seen in Fig 13 and 14 respectively. Vessel data Ship type Deadweight Length Overall Max Draught Speed Power (installed) Built

Medium Tanker 96920 tonnes 243.28 metres 13.616 metres 14.86knots 9893kW 1992

Full-Scale Propeller Dimensions Diameter 6.85m Mean Face Pitch 4.789m Number of Blades 4 Expanded Blade Area Ratio 0.524 Design Advance Coefficient, J 0.48 Table III: The main Particulars of the Basis Vessel and Propeller


Model-Scale Propeller Dimensions Diameter = 0.35 m Blades = 4 (multiple sets) Expanded Area Ratio = 0.524 Pitch Ratio = 0.699 Material: Aluminium Alloy Table IV: The model scale propeller details

Figure 15: Open water curves for the uncoated and coated propeller. A slight reduction in torque at higher advance coefficient has led to an increase in efficiency for the coated propeller. The design operating condition for this propeller is J=0.48; no difference is detected in performance at this condition.

Figure 13: The model Propeller with uncoated Blades Results from these model tests have previously been published [13] show the effect of the coating on the propellers efficiency (fig. 15). Little difference between the coated and uncoated conditions was observed; a slight increase in performance at high values of advance coefficient, with the coated blades was seen. It is thought this is due to the decrease in the frictional resistance caused by the coating. It was also demonstrated that relatively large damage (approximately 20% coating removal) to the coating had very little effect on the propeller performance.

4.

NOISE MEASUREMENTS

4.1

ANALYSIS METHODOLOGY

Measurements were made of the model propeller noise at different rotational speeds in both the uncoated and coated case at reduced pressure corresponding to the fully loaded and ballast conditions of the full scale vessel. For this present study 10 one second samples were recorded at each condition and the results averaged before any further analysis occurred. The ITTC analysis method [14] requires that the measured sound pressure levels in each 1/3 Octave band be reduced to an equivalent 1Hz bandwidth by means of the following formula.

Where SPL1 is the sound pressure level reduced to 1Hz bandwidth in dB; re 1Pa. SPLM is the measured sound pressure level at each centre frequency in dB; re 1Pa These results are also corrected to a standard measuring distance of 1m using the following relationship.

where SPL is the equivalent 1Hz at 1m distance sound pressure level (in dB; re 1 Pa) and r is the distance between the propeller centreline and the hydrophone (in this case 0.405m).

Figure 14: The model propeller with coated blades.



Once all the data have been converted for both the uncoated, coated and background noise measurements, the net sound level of the propeller (SPLN) at each centre frequency was calculated using the following logarithmic subtraction formula given by Ross [15].

Where SPLT is the total sound pressure level measured at an equivalent 1Hz bandwidth and 1m (in dB; re 1Pa). In this case this has to be calculated for both the coated and uncoated propeller results. SPLB is the background sound pressure level measured at an equivalent 1Hz bandwidth and 1m (in dB; re 1Pa). Measurements were taken in two load conditions. The first is representative of the basis vessel in a “fully loaded” condition. The second is representative of the basis vessel in a “ballasted” condition. In order to fully replicate these conditions at model scale it is important to match the full-scale and model-scale cavitation numbers. These are calculated using the formula below for the full scale conditions and then the pressure in the cavitation tunnel reduced via a vacuum pump till the same cavitation number is obtained.

For the two conditions reported here the full scale conditions can be seen in Table V. LOADED CONDITION

BALLAST CONDITION

= 0.498

= 0.320

Propeller Immersion = 10.0m RPM = 100

Propeller Immersion = 4.66m RPM = 104

Although these are very preliminary observations the results do show a number of clear trends. The coating does appear to have an effect on the noise levels, particularly on the broadband frequencies (above about 500Hz). The coated propeller shows a relative reduction in noise level in this region for high advance coefficients. As the advance coefficients get smaller however the uncoated propeller begins to exhibit lower noise levels. At the lowest advance coefficients (where cavitation is most extensive) the difference between the noise levels of the coated and uncoated propellers is minimal across the entire frequency range. This is particularly evident in the results for the ballast condition. In the discrete frequency range (below 500Hz) little difference can be seen between the two propellers at all advance coefficients and in both fully loaded and ballasted cases. 4.3

EXTRAPOLATION TO FULL-SCALE

Accurate extrapolation of these results to full-scale would require a detailed knowledge of the influence of the proximity of the tunnel walls and many other factors which might effect the scaling from model to full scale. One such factor is the thickness of the coating. It is not possible to scale the thickness of a full-scale application onto the model. The coating applied to the model is the thinnest that is applicable and still maintain the film integrity of the coatings. If this thickness was scaled to full scale it would be substantially thicker than those applied in real life. Although scaling laws have been developed for scaling uncoated propellers, nothing exists for coated propellers. Without a large programme of model and full-scale tests to determine the correlation factors it was felt that no accurate method for extrapolation to full-scale could be applied and the results have remain as model test data. Despite this it is thought that they show general trends that should be present at full scale.

Advance Coefficient = 0.48

Advance Coefficient = 0.486 Table V: The full scale data for the fully loaded and ballast condition. During the tests the dissolved oxygen content of the tunnel water was kept between 30 and 40% 4.2

NOISE MEASUREMENT RESULTS

In the following series of figures (figs 23 to 31) the analysed SPL against centre frequencies for a series of advance coefficients are presented. It should be noted that the discontinuities that appear in some of the propeller sound curves are due to the recorded background noise being higher than the measured results.


Figure 16: Net Noise for the uncoated and coated propeller in the loaded condition, J = 0.75.




p




Figure 24: Net Noise for the uncoated and coated propeller in the ballast condition, J = 0.75.








Figure 31: Net Noise for the uncoated and coated propeller in the ballast condition, J = 0.40. 5

CAVITATION OBSERVATIONS

5.1

EFFECT OF THE COATING

In addition to the noise measurements high speed video digital photography was used to record the inception and extent of the developed cavitation patterns on the propeller in order to investigate the effect of the coating on these characteristics.




The effect of a coating on the phenomenon of propeller cavitation can be investigated in two stages: • “Inception” stage of cavitation • “Developed” stage Cavitation inception itself is a complex phenomenon which is far from being completely understood at the present time. The mechanisms underlying this phenomenon are threefold: • Water quality (mainly nuclei content and its statistics) • The growth of the boundary layer over the blade sections • Type of cavitation to be developed Amongst them, it is most likely that the growth of the boundary layer will be most affected by the presence of the coating while the latter mechanism may be of secondary importance. In the case of a “surface” cavitation, as oppose to a “vortex” type, inception occurs in the region of the boundary layer transition. In this respect, roughness stimulates the transition of the boundary layer from laminar to turbulence and hence inception. Within this framework, the Foul Release coatings studied here are made of the similar type visco-elastic material as “compliant coatings” which were first tested by Kramer [16] in the late 1950s with claims of up to 60% drag reduction and which are also capable of delaying transition from the laminar to the turbulent flow regime [17]. Furthermore, the ever clean and “open texture” of the Foul Release coatings, as stated earlier, display hydrodynamically smooth surfaces compared to uncoated propeller surfaces in service which are expected to have rough and jagged textures. Based upon these features it can be speculated that the delay in the boundary layer transition will delay cavitation inception. However, this effect may not be so important for full-scale propellers which operate in fully turbulent regime. Even if it is limited to the leading edge regions, full-scale propellers will experience laminar flow and this effect can be important for special propellers designed to avoid cavitation. Another interesting nature of the visco-elastic materials is their effect to alter the turbulence characteristics of the flow near the wall and even “re-laminarise” the turbulence flow. This will not only affect cavitation inception but also influence the characteristics of the developed cavitation. In contrast, the protuberances of uncoated and not well-maintained rough blade surfaces are expected to destabilise the vortices more quickly, thus hastening the eruption and bursting process increasing the turbulence production and ultimately causing early inception as well as higher friction.

In the case of “vortex” type cavitation, particularly the tip vortex, the nature of the vortex is strongly dependent upon the nature of the boundary layer over the blade in the tip region, which can be affected by the coating. If the boundary layer separates near the tip then the tip vortex will be attached to the blade while the preservation of a laminar flow near the tip can avoid the detachment of tip vortices. 1.2

CAVITATION INCEPTION

During the model tests careful observation was taken of the advance coefficient at which visible cavitation appeared. This consisted of a thin unattached tip vortex appearing behind blades as they passed through the 12o’clock position during each rotation. Note was also taken of the advance coefficient value at which dissidence occurred (cavitation visibly ceased). Table VI shows the recorded values at which inception or dissidence of cavitation occurred. For the loaded condition they are at a lower advance coefficient, meaning a delay in the occurrence of cavitation. In the ballast condition the advance coefficient for cavitation inception was slightly increased, however while this can be attributed to the changing flow conditions over the blade due to the coating it could easily be due to slight areas of damage observed on the coating. LOADED CONDITION (J) Uncoated Inception 0.517 Dissidence 0.513

Coated Inception 0.505 Dissidence 0.510

% Change Inception -2.32 Dissidence -0.58

BALLAST CONDITION (J) Uncoated Coated % Change Inception Inception Inception 0.542 0.590 8.86 Dissidence Dissidence Dissidence 0.540 0.557 3.15 Table VI: Advance coefficient (J) values at which visible cavitation inception and dissidence occurred for both the loaded and ballast conditions. 1.3

DEVELOPED CAVITATION

The following figures (figs. 32 to 43) show the developed cavitation patterns captured from the video films for the loaded and ballast conditions.



ENVIRONMENTALLY FRIENDLY MARINE ANTI-FOULING ADDITIVE S G Seabrook, Jr., Magellan Companies, Inc., United States of America. SUMMARY Magellan has developed environmentally friendly anti-fouling additives that have demonstrated efficacy in controlling hard and soft fouling when used in marine paints for boat hulls and in latex coatings for aquaculture systems. Test panels after 12 months immersion in Florida and India remained free of fouling. This safe, economical technology combines several phytochemicals that are found in the human diet. The phytochemical formula has also been successfully combined with cuprous oxide thereby enabling a reduction in the environmentally harmful levels of biocides leached from existing marine paints. This is achieved by integrating our environmentally friendly formula with existing methods in phases to progressively reduce the environmental impact of anti-fouling agents, such as leached copper. Our goal is to reach a 100% green alternative. We meet a need, which is becoming a worldwide demand, for the availability of non-toxic marine paint anti-fouling agents for pleasure craft. We are developing time-release formulas for long-term ship hull applications and the treatment of ship ballast before discharge. 1.

PHYTOCHEMICALS

The term “phytochemical” refers to a compound or combination of compounds isolated from or found in botanical sources. These phytochemicals may be incorporated in marine paints and coatings as biocides. The following table lists some uses of phytochemicals against certain target organisms as they relate to marine fouling: Anti-Bacterial (Bacterial biofilms facilitate attachment of organisms) Plant Extract Organism or Disease Soursop B. subtilis; E. coli Custard Apple B. subtilis Ginseng E. coli; P. aeruginosa; salmonella Caper E. coli Calendula B. subtilis; S. lutea; staph Globe E. coli artichoke Pumpkin staph *** Lemon grass B. subtilis; B. mycoides; staph; E. coli Spearmint E. coli Basil anthrax; food poisoning; staph; B. subtilis; P. aeruginosa; strep; Actinomycetes Rosemary B. subtilis; P. aeruginosa; typhoid; staph Licorice B. subtilis; staph; tooth decay Garlic B. subtilis;*** H-17(rec+); C. perfringens; E. amylovora; E. coli; pneumonia; X. campestris; P. aeruginosa; salmonella; Staph spp; S. sanguis; E. carotovora


other

Aloe vera Tangerine Pomegranate

Black currant Tea Turmeric Ginger

Anti-Crustacean Red pepper

B. subtilis; C. xerosis; P. vulgaris E. coli; P. vulgaris; P. aeruginosa anthrax; B. subtilis; E. coli; *** pneumonia; P. aeruginosa; staph E. coli Actinomycete spp; E. coli; P. aeruginosa; staph; cholera B. subtilis; L. acidophilus; H-17(rec+) B. subtilis; anthrax; E. coli; L. acidophilus; staph

barnacles

(Capsaicin) Mint (menthol)

barnacles

Algicidal Pomegranate

algae

Grapefruit

algae

Molluscicidal Red pepper

Zebra mussels

(capsaicin) Mint (menthol)

Zebra mussels


2.

CHARACTERISTICS OF PHYTOCHEMICALS IN PAINTS AND COATINGS

Certain phytochemicals when combined and added to marine paints and coatings have been shown to have a synergistic effect, having greater biocidal efficacy than when used alone. For example, the combination of capsicum and menthol have been shown to be effective in controlling the attachment and colonization of barnacles. The combination of other phytochemicals such as, grapefruit seed extract and pomegranate have been shown to be effective against soft fouling. Phytochemicals that leach out of paint dilute and disperse extremely well into water. Their biodegradability will ensure they do not accumulate to unacceptable levels in the environment. The substitution of our phytochemical formulations in whole or in part for current anti-fouling agents, such as the replacement of copper-based agents, will significantly reduce harmful environmental exposure to such toxic metal ions. The anti-fouling effect particularly resides at the external surface layer in immediate contact with the surrounding water. This characteristic makes the compositions comprising phytochemicals compatible with the following; leaching coatings, ablative coatings and self-polishing coatings (polymers), latex for coating aquaculture or other systems in constant contact with an aqueous environment, and water-soluble polymer delivery systems for ship ballasts. The phytochemical compositions are effective against marine and/or freshwater organisms capable of attaching to and colonizing the submerged hull surfaces of ships and boats. Some of these organisms include the following; parazoans, coelenterates such as polychaete and oligochaete worms, mollusks such as, zebra mussels, arthropods including crustaceans such as, acorn and goose barnacles.. The compounds or their mixtures are effective in inhibiting the attachment and/or subsequent development of the adult or larval forms of the targeted organisms. The phytochemical-based compositions are also effective against marine and freshwater plants including algae and higher plants that can attach to a ship hull or other submerged surface. The compositions are also applicable in coating water intake and discharges from power stations or other submerged pipes.

These often become obstructed by marine growth such as zebra mussels. Costly mechanical cleaning may be delayed or avoided by restricting biogrowth by passive means, providing significant economic benefits. The phytochemicals may be added in any combination to leaching paints (hard and soft), ablative paints (coatings), self-polishing coatings (polymers), longlife anti-fouling coatings and fouling-release coatings, at levels from about .01% to about 60% by volume or weight to weight.

A concentrate (liquid or powder) comprising about 4% capsicum may be used in the paint compositions of the invention at a concentration of between about .01% and about 60% v/v or weight to weight of a liquid paint base. Four percent capsicum concentrate (liquid or powder) may be between about .01% to about 15% v/v or weight to weight of a liquid paint base. A concentrate (liquid or powder) comprising about 14% capsicum may be used at a concentration of between about .01% to about 60% v/v or weight to weight of a liquid paint base, preferably between about .01% to about 15% v/v or weight to weight of a liquid paint base. In one example, a concentrate (liquid or powder) comprising about 87% capsicum may be used at a concentration of between about .01% to about 60% v/v or weight to weight of a liquid paint base, preferably between about .01% to about 15% v/v or weight to weight of a liquid paint base. Menthol (liquid, crystals or powder) may be included at a concentration of between about .01% and about 60% v/v or weight to weight of a liquid paint base. The compositions may also comprise a mixture combining grapefruit seed extract and menthol (liquid, crystals or powder) in a weight-to-weight ratio range of from about 0.5:1 to about 2:1, the final concentration of the mixture in the paint composition being between about .01% to about 60% v/v or weight to weight of a liquid paint base. The following photographs are examples of the phytochemical formulas at work against a variety of marine organisms that may attach to boat and ship hulls, aquaculture systems and water intake systems.

Figure 1: Tuticorin, India -Sacred Heart Marine Laboratory, 12 Months Total Immersion Figure 1 shows test panels after 12 months of immersion in salt water at Sacred Heart Marine test laboratories, Tuticorin, India. The panel on the left is the control panel coated with a standard ablative marine paint. It does not contain a biocide or anti-fouling phytochemical formula. The panel on the right was coated with Magellan’s phytochemical formula. The barnacles at the top of the panel are due to the drill hole placed in the panel after the panel was coated exposing an unpainted surface and therefore permitting subsequent barnacle colonization. These panels were subjected to static and dynamic



testing to simulate a ship both moored and underway. They were alternated between static and dynamic every 30 days. The dynamic speeds varied from 7 knots to 26 knots.

NOTES: • At time of inspection all panels sprayed with • Street pressure water, then kept wet with a sea water spray • Biofilm ratings performed before water rinse. Scale 8= trace 2= severe. • Panel Edges and mounting holes are not considered during the fouling ratings

Figure 2: Battelle Labs, Florida, 12 Months Total Immersion In Figure 2, the full panel shown (listed as Panel E in the table below) is a test panel containing Magellan’s phytochemical formula after 12 months immersion under static conditions. The fouling around this panel is the untreated frame on which the test panel is mounted. Battelle Marine Research Labs in Florida, USA that conducted this test gives a 10 as its highest performance rating as a measure of fouling resistance. The key to the table below: Tr = trace, which means less than 1% fouling, numbers for each line item are reflected as a performance rating between one and ten with ten being the best performance. Exceptions are the last two line items, which are percentages. 2.1

Figure 3: Poseidon Labs, Lake Erie Zebra Mussel Test Panels The photograph above illustrates a result from the Poseidon Zebra Mussel Exposure Testing Program and shows panels placed in Lake Erie before the Zebra Mussel growing season and removed after the five-month growing season. The panel on the left is the test panel and the panel on the right is the control panel.

BATTELLE REPORT

Test Panel #

A

B

C

D

E

F

General Performance

10-

3

9

10-

10-

10-

Barnacles

1

+3

+9

Tr

Tr

Tr

Mollusks

Tr

Tr

Annelids

Tr

Hydroids

Tr

Tr Tr

Tr

Bryozoa Encrusting

Tr

Tr

Tr

Tr

Tr

Tr

Tr

Tr

Tr

Tr

Tr

Tr

Bryozoa Filamentous

Tr

Bio Film

6

8

8

8

6

6

Algae%

4

1

2

1

2

4

Amphipod Tubes %

5

2

5

2

5

5

LEGEND: 10=Perfect (0% fouling) Tr=Trace (1% fouling) Trs=Traces(2-5% fouling) 10- = (<10% fouling) 9= (10% fouling 8= 20% fouling 0= Complete fouling(100%)


Photo A: Control Sample

Photo B: Test Sample Figure 4: Four Months Aquaculture Rope Test


The aquaculture photographs are of polyethylene rope that has been coated with a solvent-based butyl latex incorporating Magellan’s marine anti-fouling formula. The photographs show results after 4 months immersion. Photograph A is the control sample. Photograph B is the test sample. Aquaculture is a 56.3 billion dollar per year business globally that spends over a billion dollars per year in the maintenance and cleaning of its systems due to fouling. Over 10% of fish production can be lost due to fouling. Currently, there are no safe alternatives for fouling control of these systems that would not introduce toxic compounds to the fish stocks that could render them unfit for consumption. Magellan has been pioneering the use of environmentally friendly anti-fouling additives for marine and aquatic systems for over 6 years. We are committed to finding safe, economically viable approaches to improving our oceans and waterways through the use of phytochemical technology as a replacement to use of toxic chemicals. We have built a team of experts to pursue the advancement of this technology and we welcome the opportunity to partner with like-minded companies, organizations and philanthropists associated with the marine industry. 3.

AUTHORS’ BIOGRAPHY

1988 and 1989 he was recognized as a NATO Postdoctoral Fellow at the Ecole Superieure de Physique et Chimie in Paris. Dr. Freeman has advised Magellan in developing its technology for the last seven years and serves as a board member. Stephen Beckstrom-Sternberg, Ph.D., is a Senior Scientist in the Pathogen Genomics Division of the Translational Genomics Research Institute, Phoenix. He is also an Associate Professor of Biology at Northern Arizona University, Flagstaff. He was past co-director of Intramural Sequencing at the National Institutes of Health. Washington, DC. He is also an expert in the science of Phytochemicals and has been an advisor to Magellan with the permission of the US Government for the last nine years. He is a Magellan board of directors’ member and serves on Magellan’s board of scientific advisors. Jonathan Matias, is President of Poseidon Ocean Sciences, NY, NY and Poseidon’s Sacred Heart Marine Research Centre, Tuticorn, India. He has been conducting anti-fouling tests utilizing Magellan’s technology for the past five years. Glenn Stockum, Ph.D., is a retired chemist formerly with Johnson & Johnson, Dallas, Texas. In 2005 he developed a latex coating containing Magellan’s technology for aquaculture materials and oceanographic equipment. He serves on Magellan’s board of scientific advisors.

Samuel G. Seabrook, Jr., is the Founder, President, and CEO of Magellan Companies, Inc. He is the named inventor of two patents and five patents pending. He is a graduate of the University of Georgia.

4.

ASSOCIATES

James A. Duke, Ph.D., (Botany, University of North Carolina, 1961) is retired from the USDA (1995) as an Economic Botanist. He has published over 30 books on phytochemicals and is internationally acclaimed. He is refining his phytomedicinal database (http://www.arsgrin.gov/duke/) still maintained at the USDA (Beltsville, MD). In 2001, 40 or more years after receiving his PhD from the University of North Carolina (Chapel Hill, NC), he was awarded a Distinguished Alumnus Award. He serves on Magellan’s board of scientific advisors. Benny Freeman, Ph.D., is a Senior Chemist and Polymer Expert with the University of Texas in Austin, Texas. Dr. Freeman is a polymer diffusionist (the science of placing additives into different plastics). Before assuming his current position Dr. Freeman was a Professor and Associate Department Head of Chemical Engineering at North Carolina State University. He received his Ph.D. from the University of California, Berkeley, after writing his thesis on The Effect of Hydrostatic Pressure on Mutual Diffusion Coefficients in Polymer Solutions. In



DELTA T AND DELTA DB B Glover, Alocit Systems Ltd., UK SUMMARY When it comes to the designing and building of boats and ships, one of the key areas for consideration is “HOW TO INSULATE IT?” There are many types of insulation products available on the market today, most really on the basic foundation of using an insulating foam or insulation lagging cover to wrap a substrate. Until recently the issues these types of insulations create have been “lived with” as the benefits from insulation have outweighed the problems, such as corrosion under insulation (CUI), promotion of fire, extensive maintenance and the intense labour required for application. But there is hope!! The new Delta range of products from the Alocit Group is a fresh approach to both thermal and sound insulation, using technology originally designed for the space shuttle. The Delta range has been adapted from standard ceramic coatings, such as tile paints and infused with beaded technology to create a new ceramics insulation coating (CIC). The entire range is water-based, solvent-free and non-toxic. 1.

INTRODUCTION

To understand CICs we need to first look at the history of their design and the characteristic differences they show from that of standard insulation methods. Ceramic insulation coating produces temperature differentials across the surface of a substrate. This dramatically reduces substrate surface temperatures, increasing efficiency and creating a personnel protection barrier where required.

Conventional insulation works by trapping air in cavities. We all insulate our houses using conventional insulation in the roof and similar insulation to this is the norm when it comes to insulating shipping. This type of insulation is called loose fibre insulation and works by trapping air between the fibre layers. On boats and shipping the insulation is applied by cutting the insulation and then wrapping or fastening it to the substrate using shooting pins, capping pins or a tape and foil around piping. This method of application is highly labour intensive (approx 10 sq ft per hose) and can also create additional issues with its use.

CICs are not to be confused with Ceramic coatings, which are extremely hard coatings designed to protect a substrate and look good, i.e. tile paint, protective piston coatings etc. A third type of ceramic coating is the ceramic roof-top coating which uses glass spheres to reflect UV and insulates externally by trapping air within the coating. Again, while adopting a similar approach to CICs, they do not have the same performance characteristics.

Figure 2: Corrosion Under Insulation (CUI)

Figure 1: Loose Fibre Insulation


The most obvious issue is that of corrosion under insulation or CUI. When insulation is applied moisture may be trapped, or form through condensation, at the substrate level where it can quickly corrode the substrate surface. This is often hard to detect as the insulation itself provides a barrier to inspection, and removal is often time intensive and costly. In a marine environment, corrosion problems associated with condensation and tend to be more severe and hidden problems under bulky insulation batts could result in safety as well as cost


implications. Moisture, too, may get into the fibres of the insulation, reducing effectiveness, as can the effects of constant vibration. Using Delta T ceramic insulation coating eliminates all of these issues whilst at the same time creating added energy efficiencies when used on hot substrates. Firstly, it is spray applied, giving the user complete control over the material and its application. No longer do you need to cut and pin insulation to fit or create complex wraps for piping. The Delta range is easy to use by just taping off the area and spraying. Because of the ease of application, it can be applied to a substrate surface at up to 1000 sq ft per hour making installation a hundred times more efficient. As the Delta range bonds directly to the substrate, it eliminates CUI, prevents condensation and there is no wear due to vibration. Delta T is totally inspectable without removal as any substrate issues are easily detected on the coating surface and, if maintenance needs to be carried out, it can easily be restored with just a coating “touch up” applied when the work has been finished. Delta T has been designed to minimise or eliminate all the problems currently faced by builders, architects and repairers of ships when dealing with insulation issues. 2.

SO HOW DOES IT WORK?

Heat transfer can take place in three ways: Convection, Conduction and Radiation. In a standard insulation project, the impact of convection is not immediately relevant – although, of course, the ability of an object’s environment to transfer energy away from the surface may have a cooling effect. For insulation being applied to a solid surface, both conductive and radiant heat are relevant, yet standard insulation relies solely on its ability to reduce conductive heat. Delta T, however, employs a dual approach, reflecting radiant heat back into the source and reducing conductivity. Delta T is made using 5 different types of glass and silica particles encased in a high grade acrylic binder. The structure contains both an innovative popcorn structure and standard sphere particles, providing incredible insulation properties. The particles have entrapped air voids to slow heat transfer and heat-reflective surfaces to eliminate the reflective light-wave transfer. The result is that up to 85% of the radiant energy is prevented from escaping. The coating also has very low emissivity, reducing heat transfer further. Coupled with the low conductivity produced by the microbeads, heat flow transmission is radically reduced compared to conventional insulation systems.

Figure 3 All this means that if Delta T is used inside a vessel and applied directly to a shell, wall stiffener or overhead, an average 2 mm (2000 micron) coating achieves an RVE value of 9-13. In short, using Delta T in conjunction with wall board provides better insulation than if 2-3 inch conventional insulation had been used. With any new product though, whether being launched into the marine or industrial areas, environmental precautions must be taken. Testing must be carried out. Delta T Marine carries full Lloyds certification and has a Class A fire rating under IMO A653.16. It has achieved the highest grades available in the ASTM EB4-87 flame spread and smoke development tests and is completely non-toxic, including the effects of smoke. The entire Delta range is environmentally friendly and contains no harmful VOCs. Also, by using Delta T, you gain a major environmental advantage by eliminating the need to dispose of worn or damaged insulation fibres. All this sounds very impressive, but if the product doesn’t live up to its advertising, having Lloyds certification, IMO testing and being environmentally friendly counts for nothing. Prior to its launch, major test studies were carried out, including energy efficiency, comparing conventional insulation against Delta T. The results showed Delta T provided a 33% reduction in energy usage. Tests by the Dutch Navy have also shown similar energy savings with internal pipeline temperatures showing only a 4°C reduction in temperature when insulated with a 3mm Delta T coating, compared with a 38°C loss using conventional 2” R13 pipe lagging. Tests also showed lower radiated temperature above the Delta T at given distances from the substrate, than those recorded above the R13 lagging. This is because the lagging actually builds up heat within the fibres which then is dissipated through radiation.



Bells were setup in a test that applied 4.5 kg of force to ring the bell on a weighted arm. Sound level meters recorded levels at identical distances (40mm) from the source. Temperature for the test was 22°C. Delta dB was applied with a small application sprayer in two coats with approx 4 hours dry time in between coats. The coating was allowed to cure for 7 days. The following graphs show the non-coated version of the bell vs. the coated bell. The relationship of the graph is decibels (dB) on the "Y" axes - vertical area) and time of the vibration or sound on the "X" axes (horizontal

Figure 4 During the testing process, it was also observed that the Delta T coating deadened sound vibrations in the substrate by approximately 3-5 dB (@ 60 dB). This was caused by the reduction of sound wave movement through ceramic particles in the coating. The discovery of this additional property led to the development of Delta dB. 3.

DELTA DB

The base formulation for Delta dB is the same as throughout the Delta range, the main difference being the increase of particle weight to provide sound dampening, rather than the thermal characteristics of Delta T. Delta dB is a sound-dampening coating designed to reduce structural and mechanical noise generated through substrates and surfaces. Environmental noise has two major vectors - through substrates and through the air. Eliminating noise-vibration in the substrate prevents its transference into the air. Delta dB has special materials which are encapsulated in a high grade acrylic based binder that provides damping across the various frequency ranges. In fact, Delta dB was tailored to control this transmission, turning the energy into low grade heat. This translation offers an efficient means to "killing sound" through the substrate or surface. Delta dB offers a very cost effective solution to sound damping problems in a flexible easy to use spray coating format. Like Delta T, it can also be applied quickly, without cutting or taping bulky materials or having to glue and pin to surfaces. Its direct adhesion to the surface with its high-grade resin system provides "no worry" attachment for years to come. Delta dB was tested to contrast the sound generation from a coated surface with that from a non-coated surface. The test was performed on brass bells.


Figure 5


INVESTIGATION INTO THE USE OF GEOPOLYMERS FOR FIRE RESISTANT MARINE COMPOSITES A C J Flowerday, P N H Wright, R O Ledger and A G Gibson, University of Newcastle upon Tyne, UK SUMMARY Rising demand to reduce weight in some components of warship and high speed vessels has forced designers to consider the use of lightweight materials such as aluminium and FRP composites in areas traditionally served by steels. These materials have, in general, a significantly lower resistance to fire, requiring significant extra protection to meet the required standards for fire safety. This paper explores the potential of an alumino-silicate geopolymer, already considered for use in aircraft and other transport applications, as an inorganic matrix in fire resistant marine composites. Thermal properties for the material are established following the procedure outlined by Gibson et al. [1]. Using a small scale, low cost technique for measuring the fire resistance of composites, a propane burner was used to deliver a constant heat flux to a variety of samples. The data collected was used to establish the values for the effective thermal conductivity, k, and thermal diffusivity, α, during exposure to fire. The thermal conductivity of the geopolymer was found to be 0.123 J/msK and the thermal diffusivity was 1.42E-08 m2/s. NOMENCLATURE k α ρ Cp v T T0 T1 t X Q(t)

thermal conductivity (Jm-1s-1K-1) thermal diffusivity (m2s-1) density (kg/m3) Specific heat (J/kgK) volume fraction temperature (K or °C) initial temp of laminate and heat sink hot face temp of laminate time (s) sample thickness (m) heat flux (kW/m2)

Subscripts: f fibre m matrix c composite 1.

INTRODUCTION

Demands for faster more revenue efficient vessels are increasingly forcing designers to consider the use of lightweight materials such as aluminium and FRP composites in areas traditionally served by steels. The reasons for adopting FRP components in larger ships of steel construction have been reported previously [2-4]. Such materials have excellent durability, resistance to corrosion and allow for potentially significant weight savings by reducing the fixed mass or lightship of the vessel. This is of particular benefit for weight sensitive designs. Modern high speed passenger vessels and warships are such weight sensitive designs due to the increasing demand for the ability to carry more fuel oil for increased endurance or more payload and, in the case of warships, more extensive combat weapon systems. Reducing weight in components higher up the ship, such as in superstructure and masts, has the further benefit of facilitating this extra weapons fit by providing a lightweight platform for mounting the associated system sensors to offset the influence on the vertical centre of


gravity of the ship and so allow acceptable transverse stability to be maintained. An example showing the importance of such weight savings is the French La Fayette class frigate that has a FRP helicopter hanger aft of an otherwise steel superstructure to exploit particularly this advantage of reducing weight above the main deck [5]. There are many issues associated with the use of these materials both on their own and in conjunction with metallic main structure but fire performance is a critical design factor. Although FRP composites have demonstrated a surprisingly good level of endurance to ‘thermal insult’ [6], the nature of organic matrices which contain oxygen within their chemical structures, means they are inherently combustible [7]. Composites with resins based on solvent monomers such as styrene usually have short ignition times, high heat release rates and may produce smoke and toxic products. Various means are used to mitigate these effects [8], including the incorporation of halogens to suppress ignition, phosphorus additives to promote char formation, and antimony compounds, which interfere with the combustion reaction. Improvements in one property, however, such as ignitability, are usually accompanied by adverse effects on others, such as toxic product generation or mechanical properties [9]. A secondary insulation system is therefore required where there is a need to ensure adequate fire resistance for the structure to maintain structural integrity during and after fire [10] and to conform to necessary fire reaction characteristics with respect to heat release, surface spread of flame and toxicity. The secondary attachment of such systems to meet the required standards of fire protection detracts from the weight advantage of such materials and increases the cost and complexity of manufacturing the finished structure.


Alumino-silicated geoploymer has the potential for use as an inorganic matrix to produce inherently fire resistant structural composites, provided the mechanical properties of the laminate are suitable, or as the basis of a secondary protective laminate. The cost of geoploymer is significantly greater that of more conventional organic matrix materials but its inherently superior performance in fire means that it has been considered for specific applications where fire safety is a critical design requirement, such as for aircraft cabin materials [11, 12]. For its inclusion in marine structures the material cost would have to be acceptable for the applications considered and the processing of the material would have to be capable of being integrated into boat or ship production at reasonable cost. 1.1

REGULATORY REQUIREMENTS

It is not possible to carry out a single test to determine the fire performance of any one material, as this it subject to many different variables. Therefore, an extensive assortment of ‘testing techniques and standards’ have been developed, each being employed for different scenarios, in different countries or in different industries [13]. Most transport applications involve regulations on fire performance that vary between industries [14-19]. The most relevant to marine applications are the SOLAS requirements [18] and the IMO standard for qualifying marine materials for high speed craft as fire-restricting materials [19]. This standard is intended for passenger vessels but has become adopted as the standard when investigating new materials for warship applications. One requirement which provides a good indication of whether a material is suitable for use in fire protection systems is the ‘time to flashover.’ This phenomenon occurs in closed compartments and is the time taken for flammable gases released as a result of combustion to ignite, signifying the end of human survivability. Of significant additional concern with the use of organic composites is the potential for the release of toxic fumes and smoke. Products of both complete and incomplete combustion can often be more hazardous than the fire itself. This is of concern in all enclosed environments where there are large numbers of unprotected persons, typical of applications in transport. One such environment is the London Underground, who have particularly stringent regulations concerning emissions of smoke and toxic fumes in the event of a fire due to the highly enclosed nature of significant parts of the network. Their standard states that “toxic fumes are the most common cause of deaths in fires” and even at a non-lethal level act to impair escape [20].

The properties of geoploymers make them a viable candidate to meet such requirements. 1.2

GEOPOLYMER BASED COMPOSITES

A geopolymer is a ceramic matrix for use in fibre reinforced composites [21]. They are formed by geopolymerisation, a chemical reaction between various alumino-silicate oxides and silicates under highly alkaline conditions, yielding polymeric Si–O–Al–O bonds [22]. It is claimed that geopolymers have fire properties that are superior to all organic matrix composites obtainable at present [23]. Whilst many commonly used organic matrix materials typically undergo decomposition at temperatures approaching 300˚C and subsequent ignition, geopolymers have shown that they will not ignite throughout a range of heat fluxes spanning from 25 to 100 kW/m2 or at temperatures of up to 1000˚C [24]. Because of their inorganic nature, geopolymers do not burn and therefore do not produce any flammable gases from material combustion [23]. They also have mechanical properties comparable to their organic counterparts. The United States Federal Aviation Administration has been investigating geopolymers with a view to certifying their use in aircraft, where fire performance is a critical design consideration. The individual geoploymer that will be investigated in this study has the trade name Meyeb and is manufactured to order by CORDI-Géopolymère. The Meyeb resin is an alumino-silicate of the form shown in Figure 1. When used as a composite matrix it claims to be [25]: •

Fire-resistant.

•

Non-flammable.

•

Non-ignitable.

•

Non-toxic.

• Lightweight. Ceramic matrix composites typically have to be cured at temperatures above 1000°C, this leads to high fabrication costs and also limits the range of suitable fibre materials [26]. Meyeb geopolymer resin can be cured at ambient or higher processing temperature meaning that composites can be produced using existing manufacturing facilities. The alkaline nature of geopolymers has a corrosive effect on conventional glass reinforcement with a corresponding significant reduction in the mechanical properties of the fibres. However the use of alternatives, such as carbon fibre or basalt fibre reinforcement, overcomes this problem for structural applications.



2.

20:1 < Si:Al , 35:1

Figure 1: Chemical Structure of Meyeb Resin [21] 1.3

INITIAL FIRE-RESISTANCE TESTING

Preliminary testing was carried out as part of a broad overview of novel fire protection systems undertaken in collaboration with QinetiQ. One of the specified samples of a secondary insulation system was a thin face of Basalt-Meyeb composite backed with two layers of 8 mm basalt wool and a 5mm steel backing plate representative of the protected structure. Basalt-Meyeb facing

The method of fire resistance assessment employed followed the procedure outlined in Gibson et al. [1] for using a small scale, low cost technique for measuring the fire resistance of composite materials. Having recognised the potential that this resin has demonstrated for use in fire protection systems in the previous applications discussed, the objective was to use this technique to investigate the potential of Meyeb Geopolymer as a matrix for fire resistant composite laminates by determining the relevant, and otherwise unknown, thermal properties for future investigations. The aim was establish ‘effective’ values for the thermal conductivity, k, and thermal diffusivity, α, for different combinations of composite materials including the Meyeb resin, by subjecting the samples to a constant heat flux. 2.1

Steel back face


Resin

Fibres

No. of Plies

Meyeb Meyeb Meyeb Meyeb Meyeb Vinyl-ester Vinyl-ester

Glass Glass Glass Glass Carbon Glass Glass

2 4 6 6 3 4 6

Table 1: Samples prepared for testing

Figure 2: Initial Meyeb application investigated

The purpose of the work presented here was to formally investigate the fire resistance behaviour of geoploymer based composites and most importantly to quantify the thermal properties required for further assessment of its fire performance characteristics.

SAMPLE PREPARATION

The samples were laminated using a manual consolidation technique and were cured at ambient temperature for at least 36 hours. The samples prepared for testing are detailed in Table 1.

Basalt wool insulation

The system was tested using a propane burner rig, (as described later in the paper), under conditions approximating the SOLAS requirements [18] to find its fire resistance, i.e. the length of time taken for the cold face of the sample to reach 140°C above ambient temperature. Although the sample did not meet the required 60 minutes specified in the initial testing criteria, failing in 18.3 minutes, it was apparent that geopolymer had great potential for use in fire resistant systems as it did not ignite or give off any visible emissions during the test.

FIRE RESISTANCE ASSESSMENT

The glass fibre used was a low crimp, bi-axial glass fibre weave. The carbon fibre was a stitched, tri-axial (0°, 45° and 90°) material. The samples were manufactured to a larger size than required (170mm square) and then cut to the required specimen size of 150mm square to fit the experimental apparatus. Visual inspection of the cut edges showed that the samples appeared to be well consolidated with no voids in the laminate. The vinyl-ester samples were included to evaluate how the Meyeb resin would perform with respect to a more conventional organic matrix to provide a benchmark for comparative purposes. 2.2

PROPANE BURNER RIG

The 150mm square test sample was inserted into a steel frame, which was open at the front, to reveal an area of 100mm square to be exposed to the flame of the propane burner. The sample was thermally insulated from the frame using Kaowool to ensure that one dimensional heat transfer could be assumed. A copper heat flux meter,


which was also insulated, was then positioned in contact with the rear face of the sample and fastened in place. A layer of thermal conducting paste was applied to ensure a constant contact was achieved. Three k-type thermocouples were embedded into the copper heat sink as shown in Figure 3. Another was placed 10mm from the front face of the sample to measure the hot face temperature. These were all connected to a data acquisition system to record the temperature history from each thermocouple with respect to time.

2.3

EXPERIMENTAL PROCEDURE

In order to vary the heat flux received by the front face of the sample, two variables can be manipulated. These are the distant between the burner and the sample front-face or supplied gas pressure supplied to the burner. If required both of these can be altered to provide the required heat flux. For these tests, a pressure of 0.20bar was maintained at a constant burner to sample distance of 350mm to provide a constant incident heat flux of 32.1kW/m2 [27]. Ensuring that the burner was positioned to direct the flame at the centre of the sample, the tests were allowed to run until a linear relationship between the cold face temperature and the lapsed time was firmly established.

Frame Insulation

Copper Heat Flux Meter Propane burner

Thermocouples linked to data acquisition system

Sample

Figure 3: Diagram and photo of propane burner rig with heat flux meter 2.4

OBSERVATIONS

As is illustrated in Figure 4, the vinyl ester samples ignited during testing. It can be seen from Figure 5 that the samples fabricated with the vinyl-ester matrix suffered from complete decomposition of the resin with only the reinforcement remaining in the heat affected area. There is also extensive charring on the remaining exposed surface.

Figure 5: Comparison of samples after testing (anticlockwise from top left: 6 Glass/Vinyl-ester, 4 Glass/Vinyl-ester, 6 Basalt/Vinyl-ester (untested), 2 Glass/Meyeb, 4 Glass/Meyeb and 6 Glass/Meyeb) Figure 4: Vinyl-ester/Glass sample during testing



The samples produced using the Meyeb resin demonstrated hardly any sign of oxidation, did not ignite and remained largely unmarked with only slight evidence of charring to the exposed surface. The observations from these tests provided qualitative agreement with the performance reported for the material. 3.

CALCULATION PROPERTIES

OF

By extrapolating the linear relationship back so that it crosses an extension of the horizontal induction period, as illustrated in Figure 6, the value of the induction time can be determined, this can then be used to calculate the value of α given by

α=

THERMAL

(6)

and the value of k from

The data collected from the tests was used to calculate the ‘effective’ values of thermal conductivity, k, and thermal diffusivity, α, for the laminates tested. Making the assumption that these values remain constant with increasing temperature, i.e. that decomposition is negligible, k and α are related by the following relationship

k α= ρ Cp

dT ∂ 2T =α dt ∂X 2

.

(2)

The accumulated heat flux, Q(t), passing through a sample of material, subjected to a step change in temperature on one face, and in contact with a heat sink on the other, is given by the expression  t X 1 2 Q(t ) = k (T1 − T0 ) −  + 2 X α 6 π  

 X k =   T1 − T0

∞

∑ n =1

(− 1)n exp − n 2π 2α t    n2

 

 

X2

(3)

Q(t ) =

X  k (T1 − T0 )(t − t 0 )  t Q (t ) = k (T1 − T0 ) − = X  X 6α 

The induction time, t0, is given by .

(5)

Effective single point values of thermal diffusivity and thermal conductivity can be determined by fitting this model to the temperature rise data from the copper block.


(7)

m s C PS (TS − T0 )

(8)

aS

heat of heat sink, TS is the temperature of heat sink and aS is the area absorbing heat. The heat flux out of the laminate, into the heat sink is given by differentiating equation (3) to give q(t ) =

∞  n 2π 2α dQ(t ) k (T1 − T0 )  n = 1 + 2∑ (− 1) exp −  dt X X2 n =1  

t  .    

(9) By using the rule of mixtures and assuming that water and air content are negligible, it is possible to calculate the value of the thermal conductivity of the Meyeb resin, km, from the result obtained for the laminate, kc, if the value for the fibre, kf, is known along with the vf and vm values for fibre and matrix:

1 v f vm = ⋅ kc k f km

.

(10)

Similarly the specific heat of the resin, (Cp)m, can then also be calculated given the value obtained for the laminate and again knowing the fibre properties and also the density of the resin matrix:

. (4)

X2 6α

.

where ms is the mass of the heat sink, C P is the specific S

where T0 is the initial uniform temperature of the laminate and heat sink, T1 is the hot face temperature of the laminate and X is the laminate thickness. There is an induction period, during which no heat passes from the material sample into the heat sink. At long times the solution approaches the linear relationship

 dQ (t )  lim t →∞ dt 

The temperature rise in the heat sink can be used to determine Q(t), since

(1)

where ρ and C P are the density and specific heat, respectively. By following the procedure of Gibson et al. [1], the problem can then be treated using a solution of Laplace’s Equation

t0 =

X2 6t 0

(C p ) c =

v f (C p ) f ρ f + (1 − v f )(C p ) m ρ m v f ρ f + (1 − v f ) ρ m

. (11)


Sample

Heat flow, Q

k

2 Glass/Meyeb 4 Glass/Meyeb 6 Glass/Meyeb 6 Glass/Meyeb 3 Carbon/Meyeb

k

α

t0

J/msK

2

m /s

s

0.175 0.190 0.197 0.208 0.208

2.3E-8 5.9E-8 3.6E-7 5.0E-7 4.0E-8

37.0 134.2 145.2 104.5 77.0

Table 2: k, α and t0 values obtained

t0 Time, t Figure 6: Finding k and t0 from graphical data THERMAL PROPERTIES DERIVED

Property

Units

Glass/Meyeb Composite

Meyeb Resin

Effective values for the samples are shown in Table 2. The average of these values was then analyzed using the method previously discussed. The results of this analysis are shown in Table 3

ρ Cp k α

kg/m3 J/kgK J/msK m2/s

1588 2884 0.1924 4.20E-08

2150 4034 0.123 1.42E-08

3.1

Table 3: Calculated data 11500 Heat Flow, Q (kW/m2)

9500

7500

5500

3500

1500

-500 0.0

50.0

100.0

150.0

200.0

250.0

300.0

350.0

Time, t (s)

Q(t) Complex

Q (Measured)

Figure 7: Typical heat flow v time relationship

4.

DISCUSSION

The Meyeb Geopolymer showed highly promising fire resistant characteristics significantly out performing organic composite alternatives in the tests conducted. The samples remained intact with virtually no sign of thermal damage after testing except for a slight carbonaceous charring on the hot face. The organic matrix samples, however, ignited and suffered considerable damage with the resin decomposing leaving only the reinforcement and a considerable amount of char.

As can be seen from the example plot shown in Figure 7, the data calculated, using the modelling techniques previously outlined, shows an excellent level of correlation between the calculated heat values (Q(t) complex) and the corresponding experimental data (Q(t) measured). Values of thermal diffusivity and thermal conductivity were found using an iterative process to match the two curves as closely as possible.



The resulting thermal conductivities from the four glass/Meyeb samples show very similar values for the matrix-fibre combination, deviating only slightly from one another. The thermal diffusivity values however, show a larger variation. As this property can be highly dependent on factors, such as surface finish of the sample, this is to be expected with the hand lamination technique used for fabrication. With retrospective examination of the samples tested, it was proved that some showed a more marked curvature than others. Therefore, the area of contact with the copper heat sink could have been adversely affected leading to a source of inaccuracy within the results. A greater sample range would have acted to verify the data collected and a more sophisticated sample preparation technique would have eliminated this curvature and any discrepancies associated with it. It was determined that the thermal analysis carried out on the glass/Meyeb samples would not be suitable for the vinyl-ester samples as the level of decomposition was too great. It would be inaccurate to assume the effect of decomposition on the thermal properties to be negligible. In order to model the vinyl-ester samples a thermal model as described by Dodds et al. [6], Gibson et al. [28] and Looyeh et al. [29], could be used. This thermal model no longer assumes that the decomposition effects are negligible and instead models temperature and resin content with respect to ‘through-thickness’ by fitting to a modified form of the one-dimensional Laplace’s Equation. It assumes that matrix polymers decompose to mainly form volatiles and only a relatively small amount of carbonaceous char via an endothermic decomposition process.

5.

The analysis performed showed an excellent correlation with the experimental data collected. The value of thermal conductivity, k, was calculated to be 0.1924 J/msK for the glass/Meyeb combination composite and 0.123 J/msK for the Meyeb resin alone. The thermal diffusivity values were calculated to be 4.20E-08 m2/s and 1.42E-08 m2/s for the composite system and the resin respectively. In conclusion, this paper supports the opinion that the alumino-silicate Meyeb geopolymer has considerable potential for use as an inorganic matrix in fire resistant composite systems. It was found not to ignite or to give off any noticeable combustion gasses or smoke. Therefore, this material would not exhibit flashover characteristics as there are no gaseous products to ignite. It is an interesting material to consider for fire critical marine applications subject to cost and production issues.

6. 1.

2.

3. The assumption that water content of the samples was negligible was considered suitable and is supported by the work of Lyon et al. [12]. By carrying out a thermogravimetric analysis on a cured geopolymer resin sample, they determined the ‘weight loss history at elevated temperatures’. The sample proved to be thermally stable up to approximately 250˚C, after which, 7% by mass of was lost by what was assumed to be a dehydration reaction, yielding gaseous H2O. In the context of this paper, any fraction less than 10% by mass was assumed to be negligible. Acknowledging that the method employed within this paper is an approximation due to the assumption that the values of thermal properties remain constant with increasing temperature, it has shown to be valuable and practical method of establishing otherwise unknown values. It also has the advantage of characterising materials simply using a small-scale, low cost technique.


CONCLUSIONS

4.

5.

6.

7. 8.

REFERENCES Gibson, A.G., Wright, P.N.H., Wu, Y.-S., 2004, ‘A Small Scale Low Cost Technique for Fire Resistance of Composites’, SAMPE Conference, Society for the Advancement of Materials and Process Engineering, Long Beach CA, USA,16-19 May 2004. Wright, P. N. H., Wu, Y., and Gibson, A. G., 2000, ‘Fibre Reinforced Composite-Steel Connections for Transverse Ship Bulkheads’, Plastics, Rubber and Composites: Processing and Applications, PRC, The Institute of Materials, Vol.29, No.10, 2000, pp. 549-557. Wright, P. N. H., Wu, Y., and Gibson, A. G., 2002, ‘Factors Affecting Mechanical Strength of Steel to Composite Connections for Warship Structures’, FRC 2002, 9th International Conference on Fibre Reinforced Composites, 26-28 March 2002, University of Newcastle upon Tyne, UK. Wright, P. N. H., Gibson, A. G., and Wu, Y., 2004, ‘Fire Safe Composite to Metal Joints for Warships’, SAMPE Conference, Society for the Advancement of Materials and Process Engineering, Long Beach CA, USA,16-19 May 2004. J Y Le Lan, P Livory and P Parneix, Proc. ‘Colloquia Sandwich Construction 2’, Gainesville, USA, 1992. Dodds, N., Gibson, A.G., Dewhust, D., Davies, J.M., ‘Fire behaviour of composite laminates’, Composites: Part A, 31(2000), pp. 689-702. Anon, ‘Fire Safe Composites for Mass Transit Vehicles’, Reinforced Plastics, September 2002. Proceedings of ‘Composites in Fire’, 15-16 Sept. 1999 and ‘Composites in Fire - 2’, 12-13 Sept. 2001, and ‘Composites in Fire - 3’, Sept. 9-10 2003, University of Newcastle upon Tyne, ([email protected])


9.

10.

11.

12.

13.

14. 15.

16. 17. 18.

19.

20.

21. 22.

23.

24.

Sorathia, U., Ness, J., Blum, M. ‘Fire Safety of Composites in the US Navy’, Composites: Part A, 30(1999), 707-713. Gibson, A.G., Wright, P.N.H., Wu, Y.-S., Mouritz, A.P., Mathys, Z., Gardiner, C.P., 2004, ‘The Integrity of Polymer Composites During and After Fire’, Journal of Composite Materials, 38(15/2004), 1283-1307. Giancaspro, J., Balaguru, P., and Lyon, R., 2004, ‘Fire Protection of Flammable Materials Utilizing Geopolymer’, SAMPE Conference, Society for the Advancement of Materials and Process Engineering, Long Beach CA, USA, 16-19 May 2004. Lyon, R.E., Balaguru, P.N., Foden, A., Soratia, U., Davidovits, J., Davidovics, M., ‘Fire Resistant Aluminosilicate Composites’, Fire and Materials, 21(1997), pp. 67-73. Gibson, A.G., ‘Basic Mechanism of Fire Damage in Organic Matrix Composites’ 3rd International Conference on Composites in Fire, University of Newcastle-upon-Tyne, September, 2003. ISO 9705, Fire tests- full scale room test for surface products, 1991. BS 6853, Fire precautions in the design and construction of railway passenger rolling stock, 1987. Epiradiateur and NFF-16-101, French surface spread of flame and smoke standard. BS 476, Fire Tests on Building Materials and Structures. International Convention for the Safety of Life at Sea (SOLAS) 1974, Consolidated Edition 1992, IMO London (1992). Standard for qualifying marine materials for high speed craft as fire-restricting materials, Resolution MSC.40(64), International Marine Organisation, London, 1994. London Underground Ltd, ‘Code of Practice – Fire Safety of Materials Used in the Underground’, Manual of Good Practice (Standards and Audits), 1997, Issued January 2001. Géopolymère Institute, 2004, ‘Science: about Geopolymers’, www.geopolymer.org, 2004. Xu, H., Van Deventer, J.S.J., ‘The Geopolymerisation of Alumino-silicate Minerals”, International Journal of Mineral Processing, 59(2000), pp. 247-266. Hammell, J.A., Balaguru, P.N., Lyon, R.E., ‘Strength Retention of Fire Resistant Aluminosilicate-carbon Composites under Wet-dry conditions”, Composites Part B: Engineering, 31(2000), pp. 107-111. Sorathia, U., Perez, I., ‘Improving the Fire Safety of Composite Materials for Naval Applications’, SAMPE Conference, Society for the Advancement of Materials and Process Engineering, Long Beach CA, USA,16-19 May 2004.

25. Géopolymère Institute, ‘Technical Data – Meyeb™ Resin for Composite Materials’, www.geopolymer.org, 2004. 26. Papakonstantinou, C.G., Balaguru, P.N., Lyon, R.E., ‘Comparative Study of High Temperature Composites’ Composites Part B: Engineering, 32(2001), pp. 637-649. 27. Wu, Y.-S., Naas, A.L., ‘Calibration of Incident Heat Flux Emitted by Propane Flame’, unpublished, 2004. 28. Gibson, A.G., Wu, Y-S., Chandler, H.W., Wilcox, J.A.D. and Bettess, P., ‘A Model for the Thermal Performance of Thick Composite Laminates in Hydrocarbon Fires’, Revue de l’Institut Français du Pétrole (Special Issue), 50 1: 69-74, 1995. 29. Looyeh, M.R.E., Bettess, P. and Gibson, A.G., ‘A One-dimensional Finite Element Simulation for the Fire Performance of GRP Panels for Offshore Structures, Int. J. of Numerical Methods for Heat and Fluid Flow, 7 6: 609-625, 1997.



VACUUM CONSOLIDATION OF COMMINGLED THERMOPLASTIC MATRIX COMPOSITES FOR MARINE APPLICATIONS M Ijaz, P N H Wright, M Robinson and A G Gibson, University of Newcastle upon Tyne, UK SUMMARY An experimental and modelling study was conducted on the vacuum consolidation of commingled glass/thermoplastic composites as part of a larger project on manufacturing large monolithic structures from these precursors, such as those that could be used for marine structures. Two matrices were employed: semi-crystalline PET and an amorphous PET copolymer. Samples of commingled fabric were processed into consolidated composites by means of both a convective oven, as would be used in a real marine production environment, and a small scale experimental characterisation rig, designed to measure consolidation accurately. The samples were then cooled to room temperature. In this paper, the thermal and consolidation characterisation of these fabrics is reported. Thermally induced consolidation was observed to occur in two stages: a low temperature solid state de-bulking near to Tg, followed by full melt impregnation at a higher temperature. Both stages were modelled separately using an empirical model based on the Kamal equation. The measured consolidation vs. time profiles suggested a rapid impregnation and wetting of the fibres, occurring near to the melting point of the semi-crystalline polymer. The PET melting endotherm and crystallisation exotherm had little effect on the observed thermal profiles, suggesting that these effects could possibly be neglected when modelling the process.

NOMENCLATURE A B E H η k LVDT T Tg Tm L n&m P p S t

X x PET

first stage consolidation second stage consolidation Young’s modulus of fibres in bending activation energy viscosity of resin constant or Boltzmann’s constant linear variable differential transformer absolute temperature glass transition temperature (oC) melt temperature (oC) characteristic flow length exponents compaction pressure driving pressure permeability time (s) fibre volume fraction (no compaction pressure) fibre volume fraction (under pressure) degree of impregnation depth of melt penetration Polyethylene Terephthalate

1.

INTRODUCTION

φ0 φ

In the last two decades, thermoplastic matrix composites have established themselves as versatile, high performance and economically attractive materials in the automotive, aerospace, sports and, latterly, marine industries. There are potential benefits for using commingled thermoplastic matrix composites in the manufacture of large lightweight marine structures. They have very recently been applied in the manufacture of small craft, Figure 1, but there is considerable interest in expanding these applications to either components or hull structure of larger vessels previously based on thermoset composites. The largest of which are represented by the Sandown Class SRMH, Hunt Class MCMV and, the


largest, the Swedish Visby Class corvette. They share the same advantages of established thermoset marine composites such as good strength characteristics, light weight, corrosion resistance, durability and formability. The particular driver behind this interest is that there use involves different production processes to those of thermosetting composite technology and avoid increasingly important environmental issues. There are the not the same environmental concerns of VOC emissions, such as styrene, and the difficulty of recycling as the material can be milled and reformed into new secondary applications after the primary use. Commingled yarn and woven fabrics consisting of continuous glass fibres and various thermoplastic matrix systems have now been commercialised and are readily available for producing thermoplastic matrix composite structures for marine applications. The work presented here relates to commingled glass/thermoplastic materials employing PET homopolymer and copolymer. There has been extensive research to develop and improve consolidation and processing techniques for commingled materials, including the works cited here [114]. The effects of three critical processing parameters: pressure, temperature and time at temperature have been investigated, with the aim of devising cost-efficient manufacturing routes.


Figure 1: Halmatic Pacific 22 Mk II RIB and Assault Craft (courtesy of VT Group Plc) Press moulding of commingled fabrics has been widely studied because this route provides a rapid means of fabricating parts of moderate size, often for automotiverelated applications. For much larger parts, such as those encountered in marine structures, larger scale processes are required, the most promising being vacuum consolidation, also known as vacuum bag moulding. Of all the processing techniques for thermoplastic composites, vacuum consolidation has probably received the least attention, despite being a potentially viable route for large thermoplastic composite structures. The typical arrangement of this vacuum consolidation process is shown in Figure 2. Breather Bagging film

Release film

Mould tool

Sealant tape

Figure 2: Typical arrangement of vacuum consolidation process There are a number of key issues that need to be addressed before the production of thermoplastic matrix composites can be scaled up for larger applications: Understanding needs to be developed of the heat transfer required for the softening, impregnation and consolidation of such commingled thermoplastic reinforced composites. The key technical considerations become identifying the best means of heating to ensure the required quality of consolidation, the choice of bagging materials and tooling materials and the problems associated with achieving good secondary bonding. It the first fundamental one of these issues that is addressed here so that thermoplastic matrix composites can be applied with confidence in appropriate marine applications. This paper describes part of a larger project undertaken to characterize and model the vacuum-consolidation

process of a range of thermoplastic commingled composite systems under different heating and cooling modes. Experimental and analytical investigations were carried out, to study the transient heat transfer and consolidation behaviour of these systems [14]. In the present paper thermal and consolidation data obtained for Comfil® fabrics are discussed and a consolidation model is presented. Future publications will concentrate on thermal modelling of the heating and cooling cycles. 2.

PREVIOUS STUDIES

First, it is relevant to comment on previous literature. Hamada et al [1] observed that longer holding times and higher moulding pressures could be useful to increase interfacial strength of composites consisting of CF/PA6 commingled yarn. Shonaike et al [2] found that holding time had no significant effect on flexural properties of a commingled GF/PET composites system; and both the melting point and crystallinity of the matrix were insensitive to holding time. Lystrup and Andersen [3] suggested that consolidation pressure had a significant influence on the porosity of the laminates fabricated from commingled precursors, while McDonnell et al [4] found that processing temperature and time at temperature had more of an influence on the properties than the processing pressure during consolidation of a CF/PA12 commingled system. Bernhardsson and Shishoo [5] showed that the flexural properties of press-moulded warp-knitted fabrics of commingled GF/PP yarn were strongly dependent upon the processing temperature. Fatigue properties of GF/PP composites processed at 170 oC, were reported to be inferior to those processed at 180-190 oC [6]: hence a process temperature of 180 ºC was recommended for vacuum consolidation of these composites. Wakeman et al [6] suggested correlating the compression moulding parameters with mechanical properties and microstructural void content evolved during processing of GF/PP composites.



Consolidation of commingled fabrics differs from the processing of other material forms because of the geometry and distribution of reinforcing fibres and matrix in the unconsolidated preforms. Several investigators [7-14] studied impregnation and consolidation mechanisms and suggested a variety of impregnation models, usually employing Darcy’s law, to describe the flow of molten matrix into the tows, with the flow direction assumed to be perpendicular to the tow axis. Van West et al [7] developed one of the earliest and bestknown consolidation models for commingled fabrics and performed a theoretical parametric study on the consolidation of a commingled CF/PEEK system. Later, Ye et al [8, 9] showed that the void content predictions of another semi-empirical model were in good agreement with the experimental data of consolidated commingled GF/PP and CF/PEEK composites. A comparatively simple model based on Darcy’s law was derived by Klinkmuller et at [10] to calculate the impregnation time of different kinds of GF/PP commingled yarns. Cain et al [11] suggested integration of fibre deformation, heat transfer and consolidation sub-models into a single model to optimize processing of commingled preforms. Bernet et al [12] proposed a detailed and refined analysis of the laws governing fibre bundle impregnation and demonstrated the accuracy of this model in describing the consolidation of a commingled carbon polyamide/polyamide 12 (CF/PA12) composite system. In a previous investigation of the vacuum-driven consolidation process Gibson et al [13], observed that for this process the consolidation in Comfil® GF/CoPET composites occurred in two stages. Departing from the usual practice of using Darcy’s Law to model the impregnation, both consolidation stages were modelled separately using an entirely empirical model: a simplified version of the Kamal equation [15]. Good agreement with experimental measurements was found. The same approach is followed in the present study. 3.

EXPERIMENTAL

3.1

MATERIALS

The commingled, textured yarn products investigated here were Comfil® woven fabrics from Trevira Neckelmann (now Johns Manville, International). They comprised continuous glass fibres, with the matrix in the form of either PET homopolymer or PET copolymer fibres. The glass fibre surface treatment, coupling agents, and sizing were proprietary, as were the grades of polymer. As the commingled materials are proprietary products, melt viscosity data were not available for the matrix materials. However, it may be taken that these are in approximately the same range.


The copolymer was a variant of PET, modified to facilitate processing at a lower temperature. Unlike the homopolymer, which was semi-crystalline, the copolymer was amorphous. It is designated CoPET in this work. The glass transition temperature, Tg, of both matrix resins was around 67oC. The manufacturer’s recommended processing temperatures for PET and modified PET matrices were 280-300oC and 190-240oC, respectively. The hybrid yarn produced by air texturing had a basic glass fibre content of 60% by weight. This would correspond, for a void-free composite, to a glass fibre content of 45% by volume. The textile preforms used were 3x3 balanced twill weave fabrics with an area density of 800 g/m2. 3.2

PROCESSING

To characterize the consolidation and thermal behaviour of the commingled precursors, specimens of the woven fabrics were vacuum bag consolidated by forced convective heating in an oven, as in Figure 3. This is the technique that would be used for manufacturing large area mouldings in practice. This group will be referred to as ‘O-samples’ (oven samples). In addition, measurements were made in a specially designed characterization rig, Figure 4, in which the samples were heated primarily by conduction. This smaller rig was designed to facilitate more accurate and reproducible accurate measurements of consolidation during processing. The samples produced were denoted ‘R-samples’. 3.3

OVEN CONSOLIDATION

The tooling assembly for the O-samples is shown schematically in Fig. 1. The lay-up consisted of ten plies of the commingled fabric, separated from the flat aluminium tooling by release film. On top of this were layers of breather fabric and the bagging film, the latter being sealed to the tooling by high temperature tape. In a typical experimental run, ten plies of commingled fabric were stacked on a 6.3 mm thick aluminium tooling plate, 500 mm by 250 mm. A vacuum was established inside the laminate pack by a vacuum pump, attached by means of a high temperature flexible pipe. The pressure in the laminate pack was maintained at 850 mbar below ambient air pressure during consolidation. To record the temperature of the laminate during its process cycle, the lay-up was instrumented with a total of seven K-type thermocouples during its build up. Their arrangement is also illustrated schematically in Figure 3. Care was needed to ensure an adequate vacuum seal around the region where the thermocouples penetrated between the tooling and the vacuum bag. This was achieved using a high temperature sealing mastic.


LVDT

Vacuum

Hot air convective oven Aluminum tooling

Laminate pack with breather fabric and vacuum bag Data logger

Figure 3: Schematic of experimental set-up for the vacuum bag consolidation process for commingled fabric laminates

LV D T

R ubberr diap hragm

V acuum

ger P lun g Alum inu m rig bod y

Insulation In Therm ocoup les

Mo uld

Lam iin ate pack

heater Mo uld heat Figure 4: Schematic of the small scale consolidation rig



To measure the through-thickness contraction of the laminate pack in-situ during a processing cycle, a linear variable differential transformer (LVDT) transducer was employed. Because of temperature limitations it was necessary to mount the LVDT on a post outside the oven, with its probe extended to rest on top of the evacuated laminate pack, as illustrated in Fig. 1. This led to some problems with accuracy because of adventitious movements due to thermal movements between the different components of the rig. The magnitude of this background displacement was determined by running ‘dry’ tests on the tooling assembly, simulating the actual processing conditions as closely as possible. The displacement signal from the ‘dry’ run, where there was no consolidation, provided a small background correction, which was subtracted from the measurements made during consolidation. Finally, the consolidation data were normalized to give values from 0 to 100% corresponding to the beginning of an experimental run and the final consolidation at the maximum processing temperature. This neglects the effect of any voids present in the final laminate. Laminates manufactured by this route have been found to possess very low void contents, typically lower than 1%, so this was not felt to result in an appreciable error. The tooling assembly and laminate pack were pre-heated to 40 oC, in the oven until a stable uniform initial temperature distribution was established. The oven temperature was then set to the prescribed maximum processing temperature. After complete consolidation and holding the laminate at that temperature for 5-10 minutes, the assembly was cooled to room temperature, a constant vacuum being maintained. During the experimental run, the consolidation data were collected at 1 reading per second, while the temperatures were recorded every 10 seconds, through a PC data acquisition system. 3.4

CONSOLIDATION IN THE EXPERIMENTAL RIG

The second smaller experimental test rig, Figure 4, was designed and manufactured with the aim of more accurately studying and characterizing the consolidation process over a wider range of heating rates than was possible in the larger oven-based process. This rig employed conductive heating to bring the sample up to the processing temperature and provided improved repeatability in the measurement of the through-thickness temperature profiles and the consolidation during heating and cooling.


The main body of the rig was manufactured from aluminium alloy. A rubber diaphragm was employed to seal the vacuum, and the design of the chamber enabled this to be situated in a cool part of the rig, some distance from where the laminate pack was heated and consolidated. This eliminated problems of vacuum failure associated with the temperature limitations of sealing materials and facilitated the easy processing of PET homopolymer composites without the need for special bagging materials to resist the higher processing temperature. The sample laminate pack was heated by locating the test rig on a 1.5 kW electrical flat plate heater, temperaturecontrolled via a programmable three term controller to achieve the necessary range of heating rates. Heat was transferred to the sample by conduction through the base of the aluminium container and through the lower part of the tooling. The upper part of the tooling was insulated from the upper part of the rig by a block of calcium silicate board (‘Cape’ board) and the entire rig was placed inside a glass-wool insulated chamber. Consolidation pressure was achieved through the pressure of the ambient air acting on the internal piston so, apart from the somewhat different experimental configuration, the local consolidation conditions were quite similar to those pertaining in the larger oven vacuum bag rig. The vacuum maintained in the small rig was 900 mbar below ambient pressure- just a little more favourable than that used in the hot air oven. The laminates consolidated in this rig (R-samples) were of the same thickness as in the case of the O-samples (i.e. 10 layers of the fabric) and similar consolidation and thermal data were collected during processing. Figure 5 illustrates the main components, auxiliary equipment and the sample lay-up to be consolidated. At the end of the cycle the base of the rig was water-cooled to room temperature, the vacuum being maintained throughout. 4.

RESULTS AND DISCUSSION

4.1

PROCESSING CYCLES OF COMFIL® LAMINATES

Figure 5 illustrates a typical output obtained during heating cycle of a Comfil® GF/CoPET composite consolidated in the oven. The air temperature inside the oven is plotted in this figure, along with that of all the key components. Temperatures measured during processing of Comfil® composites in the small scale test rig are shown in Figures. 6 and 7. Due to the upper temperature limit of the convective oven, PET homopolymer composites were processed only in the test rig. The composites in the rig were consolidated at 1.0 kW and 1.5 kW heater power to achieve different heating rates.


250 Oven air Breather Tem perature oC

190 Lam . Top Lam . Mid. 130

70 Lam . Bot.

10 0

700

1400

2100

2800

Tim e sec.

Figure 5: A typical heating cycle of a commingled Comfil® GF/CoPET composite consolidated in the oven (O-sample) by vacuum bag moulding technique.

In the hot air oven, heat is transferred to the laminate mainly by convection to the upper surface of the pack. Consequently, higher temperatures are observed at the top surface, as seen in Figure 5. By contrast, the construction of the small scale consolidation rig results in conductive heating through the lower surface of the laminate, as can be seen in Figures 6 and 7.

composite, Figure 8(b), do not seem to show any clear influence of the heat of crystallization. This is probably due to crystallization taking place over a broad temperature range and, again, the effect being diluted by the presence of the glass fibre. To a good approximation, therefore, the cooling cycle could also probably be modelled without taking account of this exotherm.

It is interesting to note that the through-thickness temperature gradients developed in the R-samples increase with heater power, as might be expected, reducing the total cycle time. Temperature gradients and thermal lags during processing are typical of thermoplastic composite parts, even of moderate thickness, due principally to the low thermal conductivity of the polymer. A small perturbation can be seen in the temperature profile for the middle and upper surface of the laminate, Figure 7. This corresponds to the melting point of the PET homopolymer. The amorphous CoPET, by contrast, does not show this behaviour. The endothermic effect of the melting transition, however, does not seem to greatly influence the temperature field within the laminate. This is presumably due to the size of the transition being relatively small in PET, and the effect being further diluted thermally by the presence of the glass fibres. It would appear that the heating behaviour of the homopolymer performs could probably be modelled with sufficient accuracy without taking any account of the PET melting endotherm.

4.2

Thermal profiles during cooling of Comfil® copolymer and homopolymer composites are shown in Figure 8. The oven samples cooled with relatively little throughthickness temperature gradient due to the relatively low cooling rate. By contrast, the water-cooled samples in the small-scale rig show a more pronounced gradient. This will be discussed and modelled in a later paper. The cooling curves for the semi-crystalline PET matrix

CONSOLIDATION BEHAVIOUR

As mentioned, the consolidation of the mouldings was measured during processing by an LVDT transducer. Figure 9 shows the through-thickness compaction of Comfil® GF/PET and GF/CoPET composites respectively, consolidated in the test rig at the heater power of 1.0 kW. The temperatures measured at the midplane of these laminates are also shown. It can be observed that a primary consolidation that occurs at roughly the Tg of the matrix in each case. For the semicrystalline PET matrix, Figure 9(a), a further sharp consolidation can be seen to take place around the melting point of the matrix. In the case of the copolymer, Fig. 7(b), there is also a second less pronounced consolidation step, which in this case takes place over a rather broader temperature range. As described above, the curves in Figure 9 show that the consolidation in Comfil® composites takes place in two stages. The transition between these two stages is sharper in the case of semi-crystalline PET, Figure 9(a), than for amorphous CoPET, Figure 9(b). This two-stage consolidation process was first reported by Gibson et al [13]. The first stage, which will be referred to as ‘process A’, occurs at a temperature near to the Tg of the polymer. Since liquid behaviour cannot be expected at this temperature the consolidation process probably involves some solid-state stress relaxation and softening of the polymer fibres, accompanied by local visco-elastic or



visco-plastic deformation, but no bulk flow of the Darcy type. This preliminary de-bulking accounts for a substantial proportion of the total consolidation in the case of amorphous CoPET, whereas for the semicrystalline matrix, PET, it represents about 40% of the total. The second stage, ‘process B’, occurs over a significantly higher temperature range. This second process is probably a true ‘melt’ process involving impregnation flow and wetting. Although consolidation is actually driven by the external air pressure on the evacuated laminate pack, both compaction stages can be regarded as being ‘thermally’ activated in the present process, since it is the achievement of a particular temperature in the laminate pack that really enables the process to occur. It is probably to be expected, therefore, that the behaviour will not be very sensitive to the type of model chosen. Two-stage consolidation behaviour in the vacuum induced process is in contrast to the behaviour in the more commonly used press moulding or stamping process. This is because, in the vacuum case, changes in consolidation result primarily from the laminate pack reaching particular temperatures during the heating up process. In press moulding, by contrast, the entire charge is heated to an approximately uniform temperature prior to moulding, so these effects are lost, and two-stage behaviour is not seen. The two-stage consolidation effect may have a practical application: by heating to an intermediate temperature prior to moulding it may, in certain cases, be possible to ‘pre-consolidate’ the commingled precursor, reducing the final volume change. This could be useful because the contraction that occurs on consolidation is regarded as a nuisance, as it reduces moulding accuracy and can cause rupture of the bagging material. 4.3

COMPARISON BETWEEN HOMOPOLYMER AND COPOLYMER

PET

Comparing the compaction behaviour of the homopolymer and copolymer-based composites enables further conclusions to be drawn about the mechanisms of Process A and Process B. In the homopolymer case, Figure 9(a), Process A extends over a broader range of temperature compared to that of the copolymer sample, Figure 9(b), but accounts for a smaller proportion of the overall consolidation. This would be in keeping with a consolidation mechanism involving compression or compaction of a fibre network. Van Wyck [16] and later, Toll [1], have suggested a relationship of the following form for networks of similar fibres:

(

P = kE φ m − φ 0m

)


(1)

where k is a constant and E is the Young’s modulus in bending of the fibres. φ0 and φ are the fibre volume fractions when there is no compaction pressure and when under pressure, P, respectively and m is an exponent evaluated from the experimental data. This model arises from the assumption that the fibres behave as a network of bending beams of circular cross section, when under compressive load. Often the φ0 term is small enough to allow the relationship to be reduced to:

P = kEφ m

(2)

This type of model has yet to be adapted to describe mixtures of fibres such as glass and PET, but one could well expect a fairly similar type of relationship to apply in this case. The solid-state consolidation that occurs around and above Tg may be capable of being modelled mainly in terms of the change in polymer fibre Young’s modulus, using a relationship of this type. The fact that the drop in Young’s modulus at Tg is much less in the case of the homopolymer, which is semi-crystalline, would then account for the lower degree of consolidation in Process A. It would also account for the continuing consolidation of the homopolymer over a wider range of temperature- corresponding to the continuing fall in modulus between Tg and Tm. 4.4

MELT IMPREGNATION PROCESS AND TEMPERATURE GRADIENTS

Of particular interest, here is process ‘B’, that represents the consolidation taking place in the melting region of the semi-crystalline PET. As can be observed from Figure 9(a) completion of the first stage consolidation around Tg, is followed by a significant period in which little consolidation occurs until the bottom surface of the laminate approaches the melting point (~260oC) of the matrix. The consolidation curve then shows a sharp rise which continues until the top surface reaches the melting point as is evident from Figure 10(a). Thus, the consolidation time ∆t, during this stage appears to be primarily determined by the thermal field existing within the laminate, or more specifically by the size of the temperature gradient. The time taken by other concurrent consolidation phenomena e.g. intimate contact between adjacent plies, interply adhesion, fibre deformation and movement etc., seems to be negligibly small in this stage. This consolidation characteristic can be better visualised when the consolidation is plotted against the bottom and top surface temperatures of the laminate as in Figure 10(b) for the Comfil® PET matrix. Since these points roughly mark the beginning and end of the second stage consolidation process, the points ‘1’ and ‘2’ on the consolidation curves both lie approximately at the melt temperature (~260oC) of the PET matrix.


5.

CONSOLIDATION MODELING

It is relevant to begin with a discussion of Darcy’s Law which, under some simplifying assumptions, states that

dx Sp = dt η x

(3)

where x is the depth of melt penetration, t is the time, S is the permeability, p is the driving pressure and η is the viscosity of the resin. Melt impregnation processes have a characteristic flow length L along which the resin must percolate to achieve full impregnation. It is possible, therefore, to define a degree of impregnation X such that

X=

x L

(4)

In the present case, it is better to use the more general phrase, ‘degree of consolidation’, to take account of the fact that process A involves no liquid flow. In addition, the consolidation process is non-isothermal and the viscosity would probably depend on temperature according to the usual type of Arrhenius relationship, so that

η = ηo exp( H kT )

(5)

where H is the activation energy, k is Boltzmann’s constant and T is the absolute temperature. Incorporating Equations 4 and 5 into the expression for Darcy’s Law gives a relationship that might be used to describe the temperature dependent evolution of degree of impregnation

dX Sp exp(− H kT ) = ηo XL2 dt

(6)

One feature of this model, which is again characteristic of Darcy’s Law, is that, for a constant driving pressure, the rate of impregnation declines as the degree of impregnation progresses. With the present experiments it was tempting to equate the degree of impregnation with the degree of consolidation, as measured by the LVDT. However, the processes, referred to as A and B in the previous section, are clearly different and separate. For process A, which does not seem to involve any bulk liquid-like flow, Darcy’s Law would seem inappropriate. For process B, Darcy’s Law was attempted, but did not give a very good fit to the results. This consideration led to the investigation of more widely applicable laws for consolidation.

In the most general case, it can be expected that the impregnation rate in a thermally activated process would be given by the product of a rate constant K and a function of degree of consolidation, X , as expressed below:

dX = K (T ) F ( X ) dt

(7)

It will be assumed that the temperature dependence of the rate constant, K, follows the Arrhenius law and consequently;

dX = F ( X ) exp(− B / T ) dt

(8)

Equation 6 could be regarded as a special case of this relationship. The consolidation rate, dX/dt is influenced by two factors: the temperature (through the Arrhenius term) and the level of impregnation. It was possible to produce master curves giving the form of F(X) vs X by choosing an appropriate value of the parameter, B, (related to the activation energy), since

F(X ) =

dX exp( B / T ) dt

(9)

In Figure 11 master curves for both processes of PET laminates, consolidated in the rig at different heating rates are shown. These curves were drawn by selecting, using trial and error, a value of parameter B that would closely match the consolidation rates for the different thermal histories to which the laminates were subjected during processing. From these master curves it was found that for both processes (with different values of B, of course) F(X) decreased as X increased. It was necessary, therefore, to consider laws that reflected this behaviour. Taking into account models based on Darcy’s Law for instance, it might be expected that the compaction rate would be inversely proportional to the degree of impregnation because of the increasing resistance as impregnation proceeds. However, this model did not fit the results- it was found that the compaction rate actually declined much more rapidly than the Darcy Law prediction as each of the two phases reached completion. It was therefore decided to look at an empirical model that would simulate this rapid decline. One such empirical law, often used in the modelling of thermoset cure reactions, is the Kamal equation [15]. The model of Kamal and Sourour [15] is a very well known empirical model that makes it possible to express the curing rate as a function of degree of cure and the temperature in thermosetting resins. Its mathematical form is: .

dX = A [ exp(− B / T ) ] X n (1 − X ) m dt

(10)



It can be seen that, when n = -1, m = 0, Equation 8 would correspond to Darcy’s Law. The power of this empirical relationship lies in its ability to fit a wide variety of possible forms for the F(X) relationship. One disadvantage is that when used in modelling, dX dt is always predicted to be zero for X=0. This often necessitates the addition of a second term, so that

dX =  A′ exp(− B′ / T ) + A exp(− B / T ) X n  (1 − X ) m dt  (11) The purpose of the new term is solely to ensure that dX dt has a finite value for X=0. The term in (1 - X) describes empirically, the final decline in impregnation rate. In the present problem, it is not necessary to consider the X- dependence of the early part of the impregnation process, so the first term can be taken as unity (n = 0), giving significant simplification. The two separate impregnation phases can be described by:

dX A = AA [ exp(− BA / T )] (1 − X A ) mA dt

(12)

dX B = AB [ exp(− BBT ) ] (1 − X B ) mB dt

(13)

and

In addition to the six constants needed to describe the two rate equations, a seventh is needed to represent the respective size of the contributions from the two processes, so

X = fX A + (1 − f ) X B

(14)

Because of the separation of the two processes in temperature it was possible to determine f simply by inspecting the consolidation curves. Each individual Kamal expression contains three parameters. These were determined, firstly by establishing the value of the activation parameters, BA and BB as in Figure 11, as the values, which most closely brought together the curves at Matrix PET CoPET

f 0.40 0.72

AA 4.0 x 1027 2.0 x 1027

BA 22000 22500

different heating rates to form one single master curve. The values of mA and mB were then determined as the values which most optimally fitted the decline of the consolidation rate with X. Finally, AA and AB were determined by the values that best fitted the initial consolidation rates for the two processes. The results in Figure 11 show that there was reasonable success in brining together the results to form master curves, although the exact form of the decay with X was not reproduced with complete accuracy. It may be that a function other than the power law in equations 12 and 13 may have accomplished this more successfully. The modified Euler method (predictor-corrector method) was employed to numerically integrate equations 12 and 13, using the constants shown in Table 1, and the measured process data. The results are plotted against time and compared with the experimental data in Figures 12 and 13. The empirical model yielded an accurate comparison with the experimental data for the R-samples of GF/PET composites consolidated at two different heating rates in the rig, as illustrated in Figure 12. Similarly, good agreement was seen between the predicted and measured values in Figure 13 for the two rig samples of GF/CoPET composites. For the O-sample of CoPET composites, however, with the same set of constants the modelling results in both stages showing some deviations from the measured results, Figure 13(c), although the overall trend was fairly consistent with the experimental curve. One possible reason for the discrepancy might be the slightly better vacuum conditions that were achieved in the in the rig, compared to the oven, which may have resulted in a slightly faster consolidation rate. Another reason could be that the LVDT compaction measurements were more reproducible in the case of the rig, while those in the oven were rather less so. By removing or minimizing these dissimilarities between the two processing environments, an improvement in the predicted results for the oven-consolidated samples can be anticipated. Thus, by bringing the processing conditions close to each other, the applicability of the proposed kinetic model could be established for both types of the heating modes used in this work. The existing model is probably accurate enough for the purposes of process design and modelling.

mA 9.5 6.0

AB 3.2 x 1036 3.2 x 1031

BB 45000 33000

Table 1: Constants used to compute Equations. 10-12 for Comfil® preforms.


mB 3.0 7.0


6.

CONCLUSIONS

An improved understanding of the vacuum-induced consolidation of commingled composites was achieved. Measurements showed, for the first time, that consolidation in this process occurred in two stages: a low temperature solid state de-bulking near to Tg, followed by full melt impregnation at a higher temperature (at and above Tm in the case of semicrystalline matrices). The melt-impregnation and consolidation (compaction) were observed to occur rapidly at the melt temperature. The experimental rig described here has proved to be very useful in determining consolidation characteristics on a small scale. Both impregnation stages of the consolidation process were modelled separately using a simplified version of the Kamal equation, and reasonable agreement was found between the measured and calculated results for the laminates consolidated in the test rig. In the case of the PET homopolymer the effect of the melting endotherm on the temperature profile was quite small. On cooling, the effect of the crystallization exotherm was negligible. This was due mainly to the presence of the glass reinforcement diluting these effects. The result implies that, for crystalline polymers it may be possible to neglect these phenomena during modelling. It is hoped that this investigation of the vacuum consolidation production process of commingled thermoplastic matrix composites developed provides the first stage in the understanding necessary to further explore the application of these materials in the production of larger marine structures. 7.

REFERENCES

[1] Hamada H, Maekawa Z, Fujita A, Matsuda M, and Hokudoh T, “Impregnation and mechanical properties of carbon fiber reinforced thermoplastic composites with commingled spun yarn”, 39th International SAMPE Symposium, Anaheim CA, USA, April 11-14, 1994, 1507-1520. [2] Shonaike G O, Matsuda M, Hamada H, Maekawa Z, and Matsuo T, “The effect on impregnation technique on bending properties of glass fiber reinforced polyethylene terephthalate composites”, Composite Interfaces, 1994, 2, 157-170. [3] Lystrup A, and Andersen T L, “Autoclave consolidation of fiber composites with a high temperature thermoplastic matrix”, Journal of Materials Processing Technology, 1998, 77(1-3), 8085.

[4] McDonnell P, Mcgarvey K P, Rochford L, and Bradaigh C M O, “Processing and mechanical properties evaluation of a commingled carbonfiber/PA-12 composites” Composite Part A, 2001, 32, 925-932. [5] Bernhardsson J, and Shishoo R, “Effect of processing parameters on consolidation quality of GF/PP commingled yarn based composites”, Journal of Thermoplastic Composite Materials, 2000, 13, 292313 [6] Wakeman M D, Cain T A, Rudd C D, Brooks R, and Long A C, “Compression molding of glass and polypropylene composites for optimized macro and micro mechanical properties- I Commingled glass and polypropylene”, Composites Science and Technology, 1998, 58, 1879-1898. [7] van West B P, Pipes R B, and Advani S G, “The consolidation of commingled thermoplastic fabrics”, Polymer Composites, 1991, 12(6), 417-427 [8] Ye L, Friedrich K, and Kastel J, “Consolidation of GF/PP commingled yarn composites”, Applied Composite Materials, 1995, 1, 415-429 [9] Ye L, Friedrich K, Kastel J, and Mai Y-W, “Consolidation of unidirectional CF/PEEK composites from commingled yarn prepreg”, Composites Science and Technology, 1995, 54, 349358 [10] Klinkmuller V, Um M-K, Steffens, Friedrich K, and Kim B S, “A new model for impregnation mechanism in different GF/PP commingled yarns”, Applied Composite Materials, 1995, 1, 351-371. [11] Cain T A, Wakeman M D, Brooks R, Long A C, and Rudd C D, “Towards an integrated processing model for a commingled thermoplastic composite” ICCM-11, 1997, V, V-366 V-375. [12] Bernet N, Michaud V, Bourban P E, and J.-A.E. Manson J-A E, “An impregnation model for the consolidation of thermoplastic composites made from commingled yarns” Journal of Composite Materials, 1999, 33(8), 751-772. [13] Gibson A G, Ijaz M, Dodds, N. Sharpe A, and Knudsen H, “Vacuum bag molding of large thermoplastic parts in commingled glass/PET”, Plastics, Rubbers and Composites, 2003, 32(4), 160166 [14]. M. Ijaz, “Vacuum Consolidation of Commingle Thermoplastic Matrix Composites” PhD Thesis, 2005, University of Newcastle upon Tyne, UK



[15] Kamal M R, and Sourour S, “Kinetics and thermal characterization of thermoset cure”, Polymer Engineering & Science, 1973, 13(1), 1973, 59-64

[17] Toll S, ‘Packing mechanics in fiber reinforcements’, Polymer Engineering and Science, 1998, 38, 13371350.

[16] van Wyck C M, “Note on the compressibility of wool”, Journal of Textile Institute, 1946, 37, T285T292. 8.

FIGURES

Temperature oC

230

180

Lam . Bot.

130 Lam . Mid. Lam . Top

80

30

(a)

0

700

1400

2100

2800

Tim e sec.

230

Temperature oC

Lam . Bot. 180

Lam . Mid. 130 Lam . Top 80

30

(b)

0

600

1200

1800

2400

Tim e sec.

Figure 6: Heating cycles of amorphous Comfil® (GF/CoPET) composites consolidated in the test rig at heater powers of (a) 1.0 kW, and (b) 1.5 kW



300 Lam . Bot. Tem perature oC

230 Lam . Mid. 160 Lam . Top 90

20 0

1000

2000

(a)

3000

4000

Tim e sec.

300 Lam . Bot.

Tem perature oC

230 Lam . Mid. 160 Lam . Top 90

20 0

700

(b)

1400

2100

2800

Tim e sec.

Figure 7: Heating cycles of semi-crystalline Comfil® (GF/PET) composites consolidated in the test rig at heater powers of (a) 1.0 kW, and (b) 1.5 kW 240 Lam . Bot. Lam . Mid. Tem perature oC

190 Lam . Top 140

90

40

(a)

0

700

1400

2100

2800

Tim e sec.



300 Lam . Bot. Lam . Mid.

Lam . Top

230 Tem perature oC

Lam . Top Lam . Mid.

160

90 Lam . Bot. 20

(b)

0

250

500

750

1000

Tim e sec.

Figure 8: Cooling cycles of Comfil® composites (a) GF/CoPET copolymer O-sample cooled inside the oven, and (b) GF/PET homopolymer R-sample cooled by water. 100

300

230 Mid. Tem perature

50

160

25

Tem perature oC

%Consolidation

75

90 % Consolidation

0 0

(a)

700

1400

2100

20 2800

Tim e sec. 100

220 % Consolidation 170

Mid. Tem perature

50

25

120

Temperature oC

%Consolidation

75

70

0 0

700

(b)

1400

2100

20 2800

Tim e sec.

Figure 9: Consolidation and mid-plane temperatures measured for Comfil® composites consolidated in the rig (a) GF/PET homopolymer, and (b) GF/CoPET copolymer.



125

300

%Consolidation

75

2

240

180

Bottom tem p.

Top tem p.

50

120 ∆t

25

Tem perature oC

1

100

60

%Consolidation 0 0

(a)

600

1200

1800

0 3000

2400

Tim e sec.

100 2

%Consolidation

75

Top tem p. 50 1 25 Bottom tem p.

260 oC

0 0

(b)

60

120

180

240

300

o

Tem perature C

Figure 10: Illustration of the melt-consolidation behavior during processing of Comfil® GF/PET composites.

1

F(xa)

0.0001

fast heating calculated

1E-08

slow heating

1E-12

1E-16

(a)

0

0.25

0.5

0.75

1

xa



100 fast heating

F(xb)

1

slow heating

0.01 calculated

0.0001

0.000001 0

0.25

(b)

0.5

0.75

1

xb

Figure 11 Measured and calculated master curves for (a) process A, and (b) process B of PET laminates consolidated in the rig at different heating rates.

100

% Consolidation

75 calculated 50 m easured

25

0 0

900

(a)


1800 Tim e sec.

2700

3600


100

% Consolidation

75

50

m easured

calculated

25

0 0

(b)

700

1400

2100

2800

Tim e sec.

Figure 12: Comparison of modeled and measured (symbols) consolidation of Comfil® GF/PET composites (Rsamples) at the heating power of (a) 1.0 kW, and (b) 1.5 kW 100 calculated

% Consolidation

75

m easured

50

25

0

(a)

0

700

1400

2100

2800

Tim e sec. 100 calculated

% Consolidation

75 m easured

50

25

0

(b)

0

500

1000

1500

2000

Tim e sec.



100

240

% Consolidation

75

180 tem perature

50

120 m easured %cons.

25

Tem perature oC

calculated %cons.

60

0 0

1000

(c)

2000

3000

0 4000

Tim e sec.

Figure 13: Comparison of modeled and measured (symbols) consolidation of Comfil® GF/CoPET composites (Rsamples) at the heating power of (a) 1.0 kW, and (b) 1.5 kW, and (c) oven sample.



ENVIROPEEL SYSTEMS - SETTING NEW STANDARDS T Davison, Enviropeel Systems Ltd., UK 1.

INTRODUCTION

Increasingly, products and services have to meet recognised standards in order to take their place in the modern world. And, since we now recognise that selling sweets laced with opium, or train sets with 240-volt live rails was probably misguided, it is generally accepted that such controls are a ‘good thing’ and a necessary part of commerce. Standards work on many levels; they govern the toxicity of products released into the environment and the quality of the food that we eat. Internationally, they allow companies and governments to work with common goals and materials and, in their most basic form, they govern such things as surface preparation and the size of the blast media we use. They are also used to differentiate between products. Standards for a given product type, for example, allow an Engineer to judge what is best for his project, what level of protection he can expect, how long the work will take, how long it will last – all based on known criteria, published standards and his own experience. But, what if the same Engineer knows he has a problem to solve and is offered a solution which he has not come across before and for which no relative standards exist, how can he compare chalk with cheese – what criteria should be used in such a situation? Enviropeel is a new system, offering solutions in some quite intractable areas - but merely offering an attractive solution is not enough. This paper examines the process of introducing a new class of product, how existing criteria can be adapted to measure its qualities against those that have preceded it and how new standards can be set that meet the needs of the engineers and managers that want to use it. 1.1

MEASURING UP - COMPARING LIKE WITH LIKE

Enviropeel is a corrosion-inhibiting sprayable thermoplastic polymer (CIST) and, because CIST application is a new way of providing protection, it has to be compared with other solutions to see how it measures up. If we look at the application’s original target substrate, a bolted system like a flange, there are a number of products offering a cure for its ailments. Bolt caps, flange protectors, tape wraps, shrink wraps - it’s a long list, because flanges have a lot of problems, but the majority only offer partial solutions, so, while one component may be protected, other parts are left to rot away.


Figure 1: Rusting Multi-Flange System with Bolt Caps Traditional coating solutions may offer temporary respite but require constant maintenance and can never work where corrosion already exists within a working system. Tape wraps do offer a whole system approach, but problems of accessibility and disposal make them impractical for multi-flange systems. In order for comparisons to be made between such disparate solutions, some form of protection standard which could be applied to each system would have to be devised. Engineers want measurable solutions, a graph that defines, for example, film thickness against longevity, with predictable results. And they don’t want to be caught up in a ‘hardware/software’ argument - or in the case of protective coatings, preparation and application v. product performance. Equally, a manufacturer making assertions about product performance has to be able to measure that performance against standards that customers will understand. So how, if a product is not obviously the same as other products in the market, can a rational assessment be made? Testing under standard conditions, ASTM testing for example, is one way of pointing to performance specifics. During the development of Enviropeel, the material was tested in many ways, its reaction to UV (ASTM G53-96), film integrity


2.

STANDARDS CURSE?

-

A

BLESSING

OR

A

For a manufacturer, testing standards are both a blessing and a curse. As a Group, Alocit has had considerable experience in America with lists of required tests for approval of its coatings by the US Navy. For bilge coatings, which come under Performance Specification Mil PRF 23236C, there are 19 classes, 8 types and 4 grades to choose from and you can see how many ASTM tests this involves.

Figure 2: Test piece from 3000 hour ASTM B117 showing protected and unprotected areas (ASTM G62-87), hot salt fog testing (ASTM B117-97) cryogenic testing, immersion testing - it’s another long list - but no matter how long the list, tests like these are only indicators, there is no ASTM test for the performance of sprayable thermoplastics against corrosion - at least, not yet!

The curse is the number of tests that need to be undertaken, the blessing is that, once you have ticked all the boxes, you’re approved. The number of tests may explain why an Alocit coating is one of only two approved for use in bilge areas – getting all the boxes ticked is a long and expensive process. Yet, even after all this, the real test comes when the product is used. In the case of the coating above, reports from applicators and inspectors speak of ease of use, low odour and excellent performance characteristics - but there are few objective tests for these - this is where experience has a key role to play in product selection and it is measuring such performance characteristics in an objective way that I will come to later with regard to Enviropeel.

Figure 3: 23236 ASTM test list



Figure 4: Applications large and small on the USCGC Healey and Kukui There are some advantages if you have a product that doesn’t fit into a standard category. For Enviropeel, the fact that there is no Performance Specification in which it has to fit has meant that, when the US Coastguard, who closely follow Navy specifications, saw that it was a product which had a lot to offer them, they were free to try it anywhere they saw a potential benefit. Once the system had satisfied their strict environmental criteria, applications could go ahead. Nevertheless, we were fortunate that the Coastguard immediately understood the potential of a sprayable thermoplastic and, because they had a serious problem for which they had no other solution, they were willing to experiment. In the mining industry in Australia, a similar willingness to think outside the box has seen, over the past three years, progress from application on small stored items to the specification of Enviropeel for conveyor bearing protection by BHP Billiton and widespread adoption of the system by other companies. But a manufacturer cannot rely on the vision of others to develop its business, customers want facts, history and, above all, guarantees that what they choose will give them a return on their investment. It used to be said that nobody would lose their job for specifying IBM, implying that there are risks in trying something new, and we operate in an environment that is increasingly risk-averse. So, it is in a company’s interest to show that,


far from taking a risk, there are great benefits to be gained from using its products and part of this process, at least as far as Enviropeel is concerned, is to create a ‘box’ into which it can be put, a definable, measurable category that meets the needs of conservative consultants and specifiers. 3.

NEW STANDARDS

The key difference between the solution offered by a corrosion-inhibiting sprayable thermoplastic and other solutions for problems in bolted systems is its multifunctionality.


It is the whole system approach, together with the dual, active/passive performance, that sets the new standard for bolted system protection. The fact that it is a new approach was recognised by DNV in 2001 when they contacted Enviropeel about becoming part of their longterm bolting materials testing programme. This programme, testing for which started in 2003, is to be over 10 years and is with the participation of DNV, ConocoPhilips, BP, Enviropeel and Nexans.

Special test pieces were designed and two exposure sites were designated, one splash zone site in North Carolina and the other at Bergen in Norway. Inspection is programmed at regular intervals, with removal and testing of any failing samples. After 18 months, samples in the North Carolina splash zone were performing so badly that all but two were removed.

Surprisingly enough, at least to me, although there are all sorts of standards for bolt materials and coatings, the prime motivation for the programme was that the performance of bolted systems has been subject to very little systematic research, although some work has been done on bolting materials and coatings. In many cases, design engineers have had to rely on a manufacturers’ claims for bolt performance statistics without a full understanding of how a system using such fastening materials would react in all environments. For example, an oil company with excellent results using PTFE on a phosphate primer in the Middle East, found that this coating system was ineffective in the North Sea. It was felt by DNV that, although the main focus of the programme was to test bolting materials for optimum performance in marine conditions, so many failures were occurring that it was also important to seek remedies in situations where corrosion had already taken place, and be able to provide on-site protection for existing systems where problems were likely to occur; hence the inclusion of Enviropeel in the programme.

Figure 7: Bolt and nut failure after only 18 months The fact that one of the two surviving systems was our thermoplastic coating was very satisfying for us but it may be of more overall significance that, despite a two year search and input from manufacturers and professional engineers, nearly all the systems failed after less than two years in splash zone conditions. 4.

Figure 5: Test piece rack

BACK TO BASICS

These findings confirmed the Company’s belief, that a new approach to corrosion protection and standard setting is required, one based on system performance rather than that of individual components. Of course, even the simplest flange is subject to a complex range of forces and it would be a very complex design calculation that allowed for all the material, component and environmental factors - as well as the accountants’ bottom line. How much better it would be if such a system could be designed on the basis of system requirements without taking into account environmental factors?

Figure 6: Splash zone exposure site

There are, of course, more reasons than just corrosion for flange failure. The most common cause for failure in bolts is fatigue cracking – and this can be as a result of a number of factors, wrong preloading, poor connection design or improper assembly. Corrosion fatigue, hydrogen embrittlement, stress corrosion cracking – it’s another long list but, no matter how much we would like to have a panacea for every possible situation, it is certainly beyond the scope of a coating, applied after manufacture and assembly, to be able to address all these problems!



Yet, if we examine a number of the factors involved, we can see that the effects of corrosion have to be taken into consideration at all levels. An assembly of separate components, which would almost certainly be of different materials, would be prone to galvanic corrosion. Thus the bolt material would need to have a free corrosion potential that is more positive than the base material of the flange, otherwise the corrosion effect on the much smaller surface area of the bolt would be severe. 4.1

TYPICAL BOLTS

CORROSION

ON

FLANGE

In fact, in the offshore industry, despite awareness of the importance of compatible materials, there has been increased use of corrosion resistant alloys in piping systems but, because compatible bolt materials are often unavailable, low alloy steel [ASTM A193 B7] bolts are in common use. Severe, rapid corrosion of a B7 bolt will result unless isolating sleeves can be installed and constant vigilance is required in a system where such incompatibilities exist.

If galvanic corrosion is the most likely mechanism for the acceleration of damaging corrosion in bolted systems, other factors also exist. Pitting, crevice corrosion and stress corrosion cracking all need to be taken into account, even for corrosion resistant alloys, and, where coating systems are used to protect low alloy materials, damage to the coating from stress in use or during assembly is common. Corrosion affects the system in two ways: Primary safety risk, by reducing the load-bearing ability of the bolt and hence the integrity of the system - with high potential for human, environmental and economic damage, and secondary economic risk from the economic and production consequences of seized bolts and the need to cut them out. Thus, eliminating corrosion as a factor in the consideration for bolted system survival would be a significant breakthrough in both safety and economics and, if such a breakthrough was concomitant with a reduction in environmental impact, it would seem an extremely worthwhile objective. Yet the introduction of sprayable thermoplastics appears to offer just such a combination and it is quantifying the value of this approach that is of such importance for its general acceptance. 5.

SETTING THE STANDARD

The Company has adopted a multi-channel approach to the issue of parameters for the performance of its thermoplastic coating system:

Figure 8 The table shows a typical group of candidate bolting materials and the likely corrosion effect that would result from their use.

5.1

ASTM AND OTHER TESTING OF SPECIFIC QUALITIES AND FUNCTIONS

A wide variety of testing has been undertaken to establish the suitability of materials and equipment as mentioned earlier. Testing continues as specific requirements come to light and, as the Company adopts a policy of continuous development, new material developments are being tested all the time.



5.2

LONG-TERM TESTING WITH WORLD LEADING AUTHORITIES, SUCH AS DNV

We are currently in the third year of the ten-year testing process with very satisfactory results so far. As the programme develops, bolted test pieces that have failed with their original coating are to be coated in CISTP and returned to the test zones to test the system’s ability to arrest corrosion once it has occurred – a key element in providing a solution for world-wide problems with flanges, valves and other bolted systems as well as providing insight into its performance in other areas, such as for stored, stand-by and transit protection. 5.3

In Britannia’s case, manufacturing faults on bolts supplied for the construction of the pipeline infrastructure had caused them to fail in as little as six months. With a particular problem on the smaller flanges and valves because of the disproportional effect of corrosion on low-diameter connections, a rolling programme of applications was started on all bolted connectors on 4 inch nominal pipes and below.

LONG-TERM APPLICATION PROGRAMMES

The Company has adopted a strategy of seeking longterm application programmes in specific target areas in order to audit progress and provide background data on all aspects of CIST applications. A key motivator in the early stages of product development was to provide a long-term ability to preserve steel structures of all kinds with a system that could be applied as a short-term reactive remedy but which would also form part of a long-term pro-active asset maintenance programme. Working on the joint Chevron/Conoco platform, the Britannia, in the North Sea, both in terms of application for the last three years and Applicator development for more than a year prior to the first applications, has allowed an unparalleled continuity of development. As with the US Coastguard, Britannia had a particular problem that they needed to solve and were prepared to look at innovative solutions because everything they had tried had failed.

Figure 10: Applications on the Britannia platform At first the Britannia operators kept a watching brief to see if Enviropeel lived up to its claims. After the first year, everybody was very happy with what had been accomplished and when, during the second year of the project, some applications from the previous year were stripped off and the substrates examined, it was clear that coating performance was excellent.

Figure 9: Rusty flanges and valves on the Britannia

Although the Britannia project had not arisen out of the connection with the DNV bolting programme, because ConocoPhilips were involved operationally on the platform and as committee members in the bolting tests, it was decided that a review of the practical experience on the platform could be usefully included in the longterm tests.



Data collection from Enviropeel applications, in the form of daily reports and photography, where permissible, had always formed part of the Company’s asset-maintenance strategy. With advice from DNV, the system was refined and now individual applications are tagged, timed and dated with operational and weather data for every one, with photography of every stage of application on representative samples for each day.

Figure 11: Pic report sheet and tag - Data recording of this type now forms part of every Enviropeel application

6.

PROGRAMME DEVELOPMENT RECOGNISED STANDARD SPRAYABLE THERMOPLASTICS

FOR FOR

7.

APPLICATION DEVELOPMENT WITH MAJOR COMPANIES AND ORGANISATIONS

Much of the focus of this presentation has been on bolted system protection because of the particular problems that such systems present. But the Company philosophy of providing flexible ‘whole system protection’, whether it be for flanges offshore, stored engineering products or standby equipment, has lead to unexpected developments. Working with BHP Billiton and Dampier Salt in Western Australia, Robil Engineering, who pioneered the use of Enviropeel in Australia, have been able to establish the remarkable ability of CIST to not only prevent corrosion but also the ingress of contaminants and abrasives into rotating systems such as bearings. Following a three-year testing programme Enviropeel has been specified for the protection of conveyor bearings at BHP Billiton and Dampier, who are part of the Rio Tinto Group. During the test period stored equipment failures were reduced from 40+ % to zero and conveyor bearings, which had been averaging 9 months between failures, are now expected to last more than three years, with no bearing failures reported since CIST applications began. In fact the Robil results were so outstanding that they won the Engineers Australia 2005 Small Company Project Engineering Excellence award!

As indicated earlier, it is the Company’s intention to provide a functional standard for CIST applications that can be accepted internationally. A joint DNV/Enviropeel programme is currently being developed to provide performance benchmarks for consultants and engineers based on standardised tests with bolted systems, these should be available in 2006.

Table 1: BHP/Dampier test results



7.

CONCLUSION

Enviropeel and its corrosion-inhibiting sprayable thermoplastic system is well on its way to establishing that taking a broader view and working to protect a system as a whole rather than a sum of its parts is a worthwhile objective. That specific criteria need to be met is understood, but the adoption of a more integrated, holistic approach offers significant advantages where one solution can be shown to be effective for a variety of problems. Prolonging system life cycles already provides sound economic justification for the use of systems like CIST but significant reductions in hazard to personnel, as well as increased environmental safety should make it a priority.



A NON CHROMATE CONVERSION COATING PROCESS FOR CORROSION PROTECTION OF AL2024 ALUMINIUM ALLOYS IN A MARINE ENVIRONMENT W C Tucker and M G Medeiros, Naval Undersea Warfare Centre, USA R Brown and D Maddala, The University of Rhode Island, USA SUMMARY To provide corrosion protection of Aluminium 2024 and other alloys in a marine environment, usually a chromate based conversion coating process is applied. However, chromates are environmentally unfriendly and efforts are underway to replace them. To develop a coating that works in similar manner, an ion with many of the same properties as chromate was required. The titanate ion was chosen for its many similarities. Electrochemical testing indicated that the titanate ion would retard corrosion in the same manner as the chromate ion. Based on successful electrochemical testing a complete conversion coating process was developed. Variables investigated included conversion solution pH and composition, process time, pre-treatment and temperature. Initially 2.5 by 10 cm samples were coated then tested in a salt spray marine environment. Several reached the 1000 hour goal with no corrosion evident. Recently, the process was scaled up to coat 7.5 by 25 cm samples. Again surface pre-treatment, process time and pH were investigated. After 336 hours of exposure, corrosion was absent on the exposed surfaces. The optimum treatment results in a "drop in" replacement for the chromate conversion coating, with only minor, advantageous variation. For example, no cyanide based chemicals are employed and the pH is more neutral at 5 rather then lower used by the chromate, both of these are environmentally friendly in addition to the removal of chromate. ¶ Several studies investigated the role of chromates in 1. INTRODUCTION chemical conversion coatings in protecting aluminum Chromate conversion coatings have been employed for and its alloys. Incomplete reduction of the hexavalent many years to protect aluminium alloys from corrosion. chromate in the dichromate ion to its trivalent form in a However the US Department of Defence [DoD] has chromium hydroxide film retained a reservoir of mandated the elimination of chromates. At present there hexavalent chromate in the conversion coating film. This is no other chromate free protective system available that reservoir could then heal defects in the film by dynamic is effective on aluminium alloys such as Al2024T3. A reduction during service exposure, so continually considerable amount of effort has gone into chromate repairing the film and providing a self-healing capability replacement technology. An excellent review of the [2]. Other later studies also confirmed the presence of mechanisms for corrosion inhibition of aluminium and its hexavalent chromium ion in the conversion coating films alloys by soluble chromates, chromate coatings and [3,4] chromate free coatings was recently presented [1]. The following hypothesis for mechanisms of inhibition were A 20% ratio of hexavalent chromium ion was reported in presented the conversion coating, which decreased on exposure to a) The hexavalent chromate ion transports to local air, but still was thought capable of being reduced at corrosion sites where they are reduced to the defects to repair the film during exposure [5]. Migration trivalent state and inhibit oxygen reduction. of the chromate to artificial defect regions was shown by b) The hexavalent chromate ions inhibit pit Raman spectroscopy and provided protection from initiation and dissolution of active intermetallic corrosion. Pits were specifically targeted by the chromate phases. ion. The mechanism proposed for chromates protecting c) Hexavalent chromates modify the chemical aluminium and its alloys was that the chromate ion composition of the surface passive oxides and “seeks” out pits and stop their growth while forming an passivated intermetallic phases by adsorption insoluble hexavalent chromium product [6]. Chromate and buffering decreased the cathodic reaction kinetics for aerated 0.1M d) Adsorb on aluminium oxides discouraging the NaCl for Al 2024 T3. Alternates to chromate also formation of aluminium chlorides which decrease the cathodic reaction kinetics, for example promote dissolution aluminium cerium based coatings, permanganate coatings and lithium based talcs [7,8,9].Chromates were shown to be a While chemical conversion coatings provide a good major inhibitor of the oxygen reduction in water to form supply of hexavalent chromate ion necessary for some of hydroxyl ion reaction shown below, especially when these mechanisms and inhibit anodic and cathodic chloride to chromate ratios are very high 105:1 [10]. O2 + 2H2O + 4e- -> 4[OH] reactions. This inhibition also takes place over a wide range of cathodic potentials.



1.1

OTHER CHROMATE FREE COATINGS.

A competing technology based on the trivalent chromium ion state was developed[11], but as an oxidizing treatment is applied it is expected that the trivalent ion is oxidized to the hexavalent form and thus uses the properties from the hexavalent ion to produce corrosion inhibition. The likely presence of the hexavalent chrome in the new coating may lead to the same health hazards that produced the desire to decrease traditional hexavalent chromate use. Alternatives to chromates that are non-chromium based exist [12] but are not very successful against aggressive alloys such as Al2024T3. Multivalent metallic ions exist and form the basis of some attempts, but not all offer large pH stability ranges for their passive films which can be determined form their respective Pourbaix diagram. Cerium based coating treatments received much attention [7], but is a complicated treatment process. Technology based on refractive metals oxide precursors such as Ti exists, but as noted are difficult to stabilize in aqueous solutions [1] even though a coil coating on aluminium alloys based on titanium ions was reported [13]. Recently, URI and NUWC (Naval Underwater Warfare Centre) developed a viable non-chromate alternative titanate conversion coating [14]. Recent data on a vanadates [15] based coating system showed breakdown at –0.5V (SCE) compared to nearly +1.8 volts (SCE for the titanate based conversion coating. A cerium based coating also on Al 2024T3 exhibited a pitting potential of around –0.6V(SCE) and was the same as untreated alloy[16]. The untreated alloy in figure 1 has similar behavior. The chromate pitting potential was 0.25V (SCE). For salt spray testing the vanadate coating pitted after 72 hours until the end of the test at 168 hours. 2.

EXPERIMENTAL PROCEDURES

2.1

COATING PROCESS

2.2

CORROSION TESTING

Two types of corrosion testing were conducted. In the first test, potentiodynamic scans were conducted to determine the cathodic and anodic behaviour of the alternate coating in contrast to untreated and chromate treated. The test solution was 0.5 N NaCl. The samples were purged with oxygen for 5 minutes prior to the start of the test. In the second test, salt spray testing at 95% humidity of 5% NaCl at 95oF was conducted on untreated, chromated and the alternate to chromate treatment. Digital photographs were taken of salt spray samples at the start of testing and when any interesting features appeared on the exposed surfaces. 3.

RESULTS

3.1

COATING PROCESS

The process results in several surface changes to the alloy. For example after the alkaline clean a black surface is apparent figure 1.:

Figure 1: Appearance after sodium hydroxide treatment.

The sequence for conversion coating Al 2024 T3 on either 2.5 by 10 cm panels or 7.5 by 25 cm panels is shown below:Solvent cleaning with acetone Mechanical cleaning with abrasive such as Scotchbrite Solvent cleaning with acetone Rinse in de-ionized water Chemical cleaning with sodium hydroxide at ph 12.5 for 10minutes at 40C Rinse in de-ionized water Deoxidize in proprietary solution of Smut-Go Rinse in de-ionized water Conversion coating in titanate solution at 62C for 3 minutes Rinse in de-ionized water Air dry.

Figure 2: Appearance after de-oxidizing treatment



The uniformity of this black layer is very dependent on the processes applied prior to alkaline cleaning. A solvent clean only with no abrasive cleaning produced a non-uniform layer. The black layer was removed by the de-oxidizing cleaning process after several minutes of treatment at room temperature as shown in figure 2. The third stage of conversion coating produces a dull surface with insignificant colour change as shown in figure 3.

The data indicates that in the test solution of 0.5N NaCl, the untreated Al2024 T3 exhibits no passive region, while the chromate and titanate conversion coating both show extensive passive behaviour. It should also be noted that the cathodic behaviour below the open circuit potential is significantly lower for the titanate coating compared to the chromate coating with the untreated showing the highest cathodic current density. 3.3

SALT SPRAY DATA

Initially, the small panels were tested under salt spray conditions. Prior to testing photographs were taken to show the representative finishes of untreated, titanate coated and chromate coated. These are shown in figure 5. A silver looking dull appearance characterised the titanate and the usual green appearance was present for the chromate.

Figure 3: Appearance after titanate conversion coating process It is in this state that further testing was conducted either by potentiodynamic scans or salt spray testing 3.2

POTENTIODYNAMIC RESULTS Figure 5: Appearance prior to testing of small panels.

Comparison data for uncoated, chromate coated and titanate coated Al 2024 T3 is shown in figure 4.

URI Coating

Base Sample

Chromate Coating

2.5 2

Esce(V

1.5 1

0.5 0

Figure 6: Coated samples after 30 days of salt spray

-0.5 -1 -1.5 -10

-5

0

2

5

10

Log i (uA/cm )

Figure 4: Potentiodynamic data for Al 2024 T3 alloy with different surface finishes in 0.5 N NaCl purged with oxygen.


The appearance after thirty days of salt spray exposure is shown in figure 6 for the small samples, with the untreated, titanate and chromate coated showing differing behaviour. The untreated was extensively corroded, while only a very small amount of discoloration was noted on the titanate coating. The chromate was not significantly changed with only some slight surface discoloration.


Large panels with the titanate coating were tested in salt spray to evaluate the efficiency of the process for larger sized components. Several different process conditions were tried and the best data is shown in figure 7. Three of these large 25 by 7.5cm panels were slightly discoloured after 16 days of salt spray exposure. The fourth panel in the figure degraded within the first five days of exposure. The three panels on the left of the figure degraded after the sixteen days of exposure. What was apparent from this testing is that the process is inconsistent when scaled up.

Previous work on the role of chromates on inhibiting cathodic reactions was reported [10,16]. In one study chromate ions in solution were shown to inhibit the cathodic reaction of oxygen forming hydroxyl ions. Whether a chromate ion in solution behaves in the same manner as one incorporated into a very thin film on the surface by a multistep process is of interest. However, for conversion coatings on Al2024T3, the cathodic reaction was also clearly inhibited, suggesting that the same processes are available to ions whether in solution or incorporated into films and that it is an inhibiting reaction for aluminium 2024 T3 alloy. A titanate film also cathodically inhibits the oxygen reduction reaction as indicated in figure 4. Its inhibition is an order of magnitude better compared to the chromate under the same test conditions. Titanium is a well known absorber of oxygen under many different conditions and here it seems to be an important role in inhibiting corrosion.

Figure 7: Large coated samples after16 days of salt spray exposure. 4.

DISCUSSION

The data from both potentiodynamic testing and from salt spray suggested that the titanate conversion coating will be a possible substitute for the chromate conversion coating. Potentiodynamic behaviour indicated a passive region for both the chromate and titanate, with the titanate range of passivation extending to 1.8volts prior to breakdown. For the chromate the breakdown potential was around 0.25 volts, significantly lower. Vanadate coatings showed a breakdown potential of -0.5 volts while cerium based coatings had a breakdown potential near the open circuit potential of -0.6volts. However the cerium coating showed good impedance data for corrosion resistance, so the value of anodic data for evaluating corrosion resistance of alternates to chromates must be in doubt. What may be more important is the cathodic data, especially when the solution is purged with oxygen. This is often a condition in service when a plentiful supply of oxygen is available for reaction. In the present study, the chromate coating showed a decrease in cathodic reaction rate in comparison to the uncoated material. The titanate coating showed a further decrease in cathodic reaction rate over both the untreated and chromate treated surfaces.

The second question of this project is whether the corrosion protection indicated by the potentiodynamic data has temporal longevity. Chromates are well known to provide good corrosion protection and this was further reinforced by the thirty day salt spray data on small panels. The chromate conversion coating was only slightly stained with no evidence of any white or copper coloured corrosion products on the surface, indicating the strength of corrosion protection. The titanate coating developed some surface discoloration after 30 days, but no white corrosion products or copper coloured corrosion products or pits which was encouraging that the titante coating was resistant to corrosion under salt spray exposure. For larger panels, the data was not as encouraging. The process was scaled up, but relatively few of the panels lasted more than 16 days. The ability to last fourteen days is the requirement of MIL-DTL-81760B and so the large panels met this requirement. However, after sixteen days the panels showed evidence of degradation in the salt spray, so the coating was marginal at best. Other coatings, such as vanadate based, lasted only 72 hours prior to pitting [15]. The salt spray testing of the large panels pointed to the coating process being inconsistent in results, with panels meeting the requirements but on an unpredictable manner. Further work is required to optimise the coating process for a consistent coating that will meet the necessary requirements. 5.

CONCLUSIONS

Titanate based conversion coatings hold promise as a replacement for chromates on Al2024 T3e alloys Further investigation will be required to turn the titanate conversion process into an industrially accepted process.



6.

ACKNOWLEDGEMENTS

The financial support of NUWC from their ILIR program is gratefully acknowledged. 7.

REFERENCES

[1]

M.W. Kendig and R.G. Buchheit, Corrosion, 59,5, 379, 2003 H. A. Katzman, G. M. Malouf, R. Bauer and G. W. Stupian, Applied Surface Science, 2,416 1979. K. Asami, M. Oki, G.E Thompson, G.C. Wood and V. Ashworth, Electrochemical. Acta, 32, 337, 1987. T. Drozda and E. Maleczki, J. Radioanal. Nucl. Chem. Lett. V 95, 339 1985 A.J. Davenport, H.S. Isaacs and M.W.Kendig, Corrosion Science 32, 653, 1991. Jun Zhao, G. Frankel and R. L. McCreery, J. Electrochemical Society, 145, 2258, 1998. F. Mansfeld, “Chemically Induced Passivity of Aluminum Alloys and Al-based Metal Matrix Composites”, Report to ONR for Project N00014-91-J-1041,1995. B. Srinivasan, S. Sathiyanarayanan, C. Marikkannu ,”, Corrosion Prevention and Control, 42, 147, 1995. B.R.W. Hinton,, Metal Finishing, 89, 15, 1991 A. Sehgal, G.S. Frankel, B.Zoofan, S Rohkhlin, J. Electrochem Soc, 147,140, 2000 F. Pearlstein and V.S. Agarwala, Plat. Surface Finishing 81,50, 1994 I. Danilidis, J.M. Sykes, J.A. Hunter and G. Scamens, Surface Engineering 4,401, 1999 M. A Smith., J.M Sykes., J.A Hunter., J.D.B Sharman., and G.M Scamans., Titanium based conversion coatings on aluminum alloy 3003. Surface Engineering, Volume.15, Number.5, 1999. US Patent 6,638,369, "Non-Chromate Conversion Coatings”. WC. Tucker, M G. Medeiros, and R Brown, awarded October 2003. H. Guan and R. G. Buchheit, Corrosion, 60#3,284-296, 2004 C. Wang, F. Jaing and F. Wang, Corrosion 60,#3, 237-24, 2004.

[2]

[3]

[4] [5] [6] [7]

[8]

[9] [10] [11] [12] [13]

[14]

[15] [16]

8.

AUTHORS’ BIOGRAPHIES

Richard Brown is a professor at the University of Rhode Island, where his research group investigates degradation of materials in the marine environment and environmentally friendly coatings. Dharma Maddala is a graduate student at the University of Rhode Island, investigating titanate coatings for corrosion protection of Aluminium alloys.


Wayne C. Tucker is a science and engineering fellow at the Naval Undersea Warfare Centre in Newport, Rhode Island. Maria G. Medeiros is a research scientist at the Naval Undersea Warfare Centre, Newport , Rhode Island.


FATIGUE CRACK GROWTH IN ANODISED ALUMINIUM ALLOYS A M Cree, Britannia Royal Naval College, UK G W Weidmann, The Open University, UK SUMMARY It is well known that the presence of a corrosion resistant anodized surface film on aluminium alloys will influence the fatigue performance of the underlying substrate material. This is generally attributed to initiation effects associated with the presence of process cracks in the oxide layer. What is less clear however, is the effect that such a film has on the growth behaviour of an already initiated, and growing, fatigue crack. This paper will report the findings of fatigue life assessment studies carried out on 2000 series aluminium alloys that had been anodized using the recently developed boric-sulphuric acid anodizing process. The results obtained demonstrated the decrease in fatigue life experienced by these alloys depends not only on the easier initiation of fatigue cracks but also on film-assisted fatigue crack growth brought about by the presence of the protective oxide film. NOMENCLATURE a A

crack length Paris intercept parameter film crack slant angle film crack length fatigue crack growth rate stress intensity with crack closure

β

c da/dN Kcl Kcl* Kfc KEqfc Kmax Kmin ∆K ∆Keff ∆Keff*

stress intensity with reduced crack closure film-crack interaction stress intensity equivalent mode I stress intensity maximum stress intensity minimum stress intensity stress intensity range effective stress intensity range reduced effective stress intensity range threshold stress intensity range crack tip stress intensity range control crack tip stress intensity range anodized crack tip stress intensity range

Kth ∆Ktip (∆Ktip)c (∆Ktip)a (∆Ktip)a+ Pcl Pcl* R

σ

∆σ tc 1.

partitioned crack tip stress intensity range fatigue crack closure load reduced fatigue crack closure load load ratio (Kmin/Kmax or σmin/σmax) applied stress applied stress range film thickness (coating)

INTRODUCTION

Many engineering components fail in service when their surfaces cannot withstand the external forces or environmental conditions imposed upon them. To overcome these problems a variety of coating systems have been developed which improve the physical and mechanical integrity of the underlying material, particularly in relation to corrosion resistance, friction and wear. For aluminium and its alloys, protection is usually achieved by anodizing. This electrolytic process produces a much thicker (typically 3-25 µm) oxide film


than would form naturally. For structurally significant items, chromic-acid anodizing (CAA) [1] and, to a lesser extent, sulphuric-acid anodizing (SAA) [2] are the usual preferred finishing operations. However the results of numerous studies [3,4,5,6] have shown that any form of anodizing is potentially detrimental to the fatigue life of the underlying substrate material. Thus the benefits gained in terms of corrosion protection must always be weighed against the likely disadvantage of reducing the effective fatigue strength of the component. The main contribution to fatigue failure of anodizing is that of encouraging crack initiation, leading to premature failure [7]. The SAA process is much worse than the CAA process in this respect. The latter, as used mainly for aerospace applications, has very little effect provided the film is kept thin (typically ~8 µm) and the normal specified industrial practices strictly adhered to. However in this era of environmental conscientiousness the use of the CAA process is unlikely to continue for reasons of its inherent toxicity. Of particular concern is the emission of hexavalent chromium vapour during the anodizing process. Existing legislation has stimulated the development of several less toxic alternatives. Currently, the most useful replacement candidate for CAA appears to be the boric-sulphuric acid anodizing (BSAA) process. This modified sulphuric-acid process allows thin and compact (2-3 µm) oxide films to be formed consistently and has the added advantage of being more economical to make up and operate. It is claimed [8] that BSAA is equivalent to CAA in all respects including its effects on corrosion resistance, paint adhesion and fatigue performance. In recent years the BSAA anodizing process has been fully accepted as a replacement for CAA and is currently in use on many US Naval aircraft. 1.1

ANODISING AND FATIGUE

Despite its many benefits, anodizing has been shown to adversely affect fatigue performance. Of the two stages of fatigue, namely crack initiation and crack growth, most attention has been concentrated on initiation. Anodized films are hard and brittle and readily crack when deformed. Since the oxide film grows out from the substrate, and is very adherent to it, any cracks that


develop in the film act as stress raisers and contribute many excellent sites for the initiation of fatigue failure. This contribution has been shown to depend on the type of anodizing process employed and the thickness of the oxide film [9], the base substrate material [10], and the presence of residual stresses after sealing [11]. A reduction to fatigue life of up to 30 %, as determined by standard S-N tests, is possible for all types of anodizing.

Maximum stress, MNm

-2

S-N data for 2024:T4 aluminium alloy anodized using the boric acid-sulphuric acid process is shown in Figure 1. 200 150 control

100 50 0 10000

anodised

100000

1000000 10000000 100000000

Number of cycles, N

Figure 1: S-N data for boric acid–sulphuric acid anodized 2024:T4 aluminium alloy. Although the influence of anodizing on fatigue crack initiation is understood, evidence presented in the literature for other coating systems [12,13,14] suggests that fatigue crack growth rates may also be affected by the presence of a thin surface film or coating. This paper reports the findings of a study carried out to assess the influence of BSAA on fatigue, and particularly crack growth rate, in a typical high strength aluminium alloy. 2.

EXPERIMENTAL PROCEDURES

2.1

FATIGUE CRACK GROWTH.

Fatigue crack growth rates can be represented in terms of LEFM using the Paris-Erdogan equation, Paris & Erdogan [15], modified by Elber [16], to account for the influence of fatigue crack closure. Crack closure gives rise to the premature contact between crack faces, due to a residual plastic stretch left in the wake of an advancing crack, thus resulting in a reduction to the effective stress intensity driving fatigue crack growth. The equation is a power law relation between the crack growth rate, da/dN, and stress intensity factor range ∆K, expressed in the form: da = A(∆K eff )m (1) dN where ∆Keff = Kmax - Kcl is the effective stress intensity factor range corrected for the influence of closure. Kcl is the stress intensity at closure where Kcl > Kmin. The intercept and slope parameters A and m depend on material variables such as the stress ratio (R), frequency, environment and temperature. Equation (1) applies only to the growth of long cracks in the Paris regime and assumes that the crack tip plastic zone and the grain size are both small relative to the crack length. 0.04 Crack length, m

In this study fatigue crack growth data were obtained from single edge notched (SEN) specimens machined from the centre of a single sheet of 18 gauge (1.3 mm thick) 2014:T6 aluminium alloy. Half of the specimens were anodized using the BSAA process which produced a compact oxide film approximately 2 µm thick. This film is shown in Figure 2. A control set of unanodized, but otherwise identical, specimens was also tested.

Constant amplitude fatigue crack growth tests were carried out in accordance with ASTM Standard E647 (1991). Specimens were cycled in tension about a mean load of Pmean =1.8 kN to a maximum load Pmax =3.25 kN at a sinusoidal frequency of 25 Hz and using a load ratio of R=0.1. These conditions imposed an initial stress intensity range of ∆K=5 MNm-3/2, a value slightly above that required to initiate and grow a fatigue crack, viz. ∆Kth. The constant load amplitude dictated that ∆K increased from its initial value to about ∆K=30 MNm-3/2 during each test. During testing the crack lengths were measured manually using a travelling microscope. To ensure that the measured growth rates were not affected by the notch tip stress field, no measurements were taken until this region had been surpassed. The resulting crack growth curves for the two alloys, in the anodized and control conditions, are shown in Figure 3.

anodised

0.03

1

2

control 1

3

23

0.02 0.01 0.00 0

20000

40000

60000

Number of cycles, N

Figure 2: SEM micrograph of film-substrate interface.


Figure 3: Fatigue crack growth curves for control and anodized 2014:T6 aluminium alloy.


2.1 (a) Fatigue Crack Closure Measurement

3.1

In this study a side face strain gauge technique was used to determine the magnitude of the crack closure load Pcl. This method utilizes compliance measurements taken from two strain gauges, G1 and G2, positioned on the side of the SEN specimens used as shown in Figure 4.

The modified Paris-Erdogan relationship indicates that one of the major factors controlling crack growth rates must be the extent to which crack closure effects reduce ∆K. For thin sheet materials, closure is primarily due to plasticity effects which occur in the wake of an advancing crack. These effects tend to be largest in the surface layers of the material. Thus the importance of a brittle, and constraining, surface film to the closure behaviour of the underlying substrate can readily be appreciated, particularly if the film is in a state of residual stress. Figure 5 indicates the values of the closure loads, Pcl and Pcl*, obtained for both the control and the anodized conditions.

∆σ strain gauges

5 mm notch

G1

G2

PLASTICITY-INDUCED CLOSURE

crack 9 mm

8 mm

notch tip strain field

∆σ

Figure 4: Strain gauge locations used for crack closure determination. One anodized and one unanodized specimen (3a and 3c) were used to determine closure. Closure measurements were taken at crack growth intervals of approximately 1 mm upon interruption of the cyclic loading with the specimen then taken manually through one complete fatigue cycle. The crack closure load, Pcl, was defined from each compliance trace using the normal deviation from linearity method. The values of Pcl thus obtained are shown as a function of increasing crack length in Figure 5.

3.


From Figure 3 it is evident that, given equal and constant loading conditions, fatigue cracks grow at an increased rate in the anodized condition. In terms of the number of cycles to failure, for an already initiated fatigue crack, these data represent a significant reduction of between 25% and 35%. When compared to the control alloy, the greater spread for the anodized condition may be attributed to film variability in terms of thickness, the presence of residual stress and structural defects in the anodized film. These results indicate that the decrease in fatigue life seen in anodized materials depends not only on the easier initiation of fatigue cracks but also upon enhanced crack growth rates in the presence of the anodized film. To understand these effects the nature of fatigue crack growth under plane stress conditions, and the influence of plasticity-induced closure, must be considered.


Figure 5: Crack tip closure loads for anodized and control specimens. As can be seen, the presence of the anodized film reduces the closure load from a steady state value of 530 N in the control condition to a value of 450 N when anodized. Thus a 15% reduction in the developed closure load occurs. Why this happens is not entirely clear at present, however it is possible that it may occur for the following reasons. It has been shown that this type of anodized film is in a state of considerable residual tension, Cree & Weidmann (in preparation). Under these conditions stress transfer from the film across the film-substrate interface will induce an opposite compressive stress in the underlying substrate material. The effect of this compressive stress, coupled with any difference in Young's modulus between film and substrate, will serve to reduce the extent of the residual tensile deformation left in the wake of the advancing crack tip. Since it is the extent of this remnant plastic stretch that is responsible for plasticity-induced crack closure, then any reduction to it will manifest itself as a decrease in Pcl. Thus the extent of crack closure experienced by the anodized material would be reduced, from Kcl (control) to Kcl* resulting in an attendant increase in ∆Keff to ∆Keff* so increasing the observed crack growth rate. Figure 4 shows the material response in terms of the Paris plots, the linearised form of equation (1) viz. log(∆K) v log(da/dN), as derived from Figure 3 and corrected for plasticity-induced closure. In determining the values of ∆Keff and ∆Keff* the steady state closure loads Pcl and Pcl* for the control and anodized conditions


were used. Invoking the concept of reduced plasticityinduced closure does reduce the disparity between the crack growth rates for the two conditions, although not completely. It is evident therefore that the difference in crack growth rates cannot be accounted for entirely by reduced plasticity-induced closure. Some other additional mechanism must also be contributing to the higher growth rates observed in the anodized material.

Figure 6: Paris-Erdogan plots for anodized and control specimens

Figure 7: Optical micrographs showing crack trajectory in anodized specimen.

3.2

3.3

INFLUENCE OF FILM CRACKS

Micrographic evidence presented in Figures 7 indicates another possible contributory factor due to the brittle nature of the anodized film. As can be seen, the anodized film in the vicinity of the crack tip is significantly cracked. These film cracks develop due to the strain field around the growing fatigue crack which is large enough around the crack tip to initiate micro-cracks in the brittle anodized layer. In the presence of a tensile residual stress, and assisted by the cyclic loading of the specimen, growth of these film cracks is inevitable. Once they reach a substantial size they are able to influence the growth characteristics of the crack in the underlying substrate material. Thus the growing fatigue crack in the substrate is able to intermittently follow, under the influence of a diminishing mode II component, the energetically more favourable direction of the cracks in the film. It is interesting to note that although the crack path in the anodized specimen shows a much greater degree of deviation from the normal mode I direction (tortuosity), the measured closure load is consistently and markedly lower, and the crack growth rate faster, than that observed in the control alloy. This is contrary to the commonly held belief [17] that a rougher crack flank profile will increase crack closure, due to roughnessinduced closure mechanisms so decreasing the observed crack growth rate. For these roughness mechanisms to operate, an additional residual displacement of the fatigue surfaces, perhaps due to unconstrained surface layer plasticity, is required. Since this condition cannot be met, due to the extra constraint imposed by the anodized film, closure will not increase in the manner expected.


COMBINED CLOSURE AND FILM CRACK EFFECTS

The accelerated crack growth rate observed in the anodized material can be explained in terms of two concurrent mechanisms; firstly reduced plasticityinduced closure due to an increase in the substrate constraint imposed by the presence of the anodic oxide film and secondly, the introduction of an intermittent mode II component caused by the interaction of the main fatigue crack with cracks present in the anodic oxide film. The combined effect of these two mechanisms is to increase the local stress intensity range at the crack tip. This situation can be modelled (qualitatively) in terms of a local applied stress intensity range, ∆Ktip, where this can be defined as; ∆K tip c = K max − K cl (2)

( ) (∆K tip )a+ = (K max − K cl* )+ K fc

(3)

The term Kfc describes the temporary local increase in stress intensity caused by the interaction of the main crack with a film crack. It is also assumed that for both conditions the developed closure load will always exceed Kmin. It is important to note that the anodized stress intensity range, (∆Ktip)+a , defined by equation (3) is only operative while the main substrate crack is interacting with a film crack. Once the main crack has passed through this region, the influence of Kfc falls to zero and the anodized stress intensity range is then simply defined by

(∆ K tip )a

= K

max

− K

cl *

(4)


Evaluating equation (3) is not trivial, since Kfc combines mode I (ki) and mode II (kii) components. Combining ki and kii cannot be achieved simply by superposition since the slant crack growth driven by Kfc invalidates the standard expression for the strain energy release rate. However, the maximum principal stress criterion of Erdogan and Sih [18], for such a mixed mode loading condition, provides a simple and elegant solution to the problem. Erdogan and Sih assumed that the rate of fatigue crack growth in mixed mode I and mode II conditions would be equivalent to the mode I case provided the principal stresses were equal. If the stress intensity in the equivalent mode I case for Kfc is defined as KEqfc, the criterion becomes

K Eqfc = k i cos 3

β 2

− 3 k ii cos

2 β 2

sin

β 2

(5)

where β =(90−θ) and θ is the slant angle of the film crack as determined from Figure 6. By resolving the uniaxial stress acting on the film crack, ki and kii can be determined from

k i = Y 0 σ πc sin

2

θ

k ii = Y 0 σ πc sin θ cos θ

(6) (7)

where c is the film crack half length and Y0 is the dimensionless geometrical correction factor for this surface crack configuration. Combining equation (5) and equation (3) defines the local anodized stress intensity range at the crack tip (in terms of only mode I components) as;

(∆K tip )a+ = (K max − K cl* )+ (k i cos 3 β2 − 3k ii cos 2 β2 sin 2β )

(8) This represents the temporary stress intensity range at the tip of the main crack as it merges with that in the film. Once the main crack passes through this region, the stress intensity range returns to its normal lower level given by equation (4). Since the trajectory of the main crack is not made to curve significantly by a continuous mode I-mode II combination, the stress intensity factors need not be modified to account for this. Thus the localised stress intensity range applied to the anodized material varies between (∆Ktip)a and (∆Ktip)a+ as crack growth proceeds. A reasonable estimate for KEqfc was obtained by considering a film crack to be a semielliptical surface flaw of film thickness depth (tc) and length 2c, inclined at an angle to the loading direction as defined earlier. For this type of crack Shiratori et al [19] showed that Y 0 could be calculated from ψ Y0 = (9) λ where


2

t  t ψ = B 0 + B1  c  + B 2  c t  s   ts t  for 0 ≤  c  ≤ 1 .  c 

   

4

 tc    c 

1 65

λ = 1 + 1 .464 

and B0, B1 and B2 are curve fitting parameters. By combining equations (5), (6), (7) and (9), the solution for the film crack stress intensity factor, in its equivalent mode I form, can be calculated from

K Eqfc = ψσ

πc ⋅ f (β, θ ) λ

(10)

where  β β β   f (β , θ ) =  sin 2 θ cos 3  − 3 sin θ cos θ  cos 2 sin  2 2 2    To evaluate equation (10) a number of simplifying assumptions, and measurements, were necessary. Firstly, under an applied cyclic loading the semi-elliptical surface flaw (film crack) was assumed to initiate and grow to a maximum size, c. At this size only then was it able to influence the growth rate of the main substrate crack. The size of c and the crack slant angles were obtained from Figure 7. To correlate the observed crack growth rate with the applied stress intensity range, given that this varies intermittently between (∆Ktip)a and (∆Ktip)a+, it was decided to simply partition the effects of (∆Ktip)a and (∆Ktip)a+ in relation to measurements taken from Figure 7. These measurements suggested that a conservative estimate for this partitioning would be a 70:30 ratio. Thus

(∆ K tip )a

(

= 0 .7 ∆ K

tip

)a

(

+ 0 .3 ∆ K

tip

)a +

(11)

where |(∆Ktip)a| is the weighted stress intensity range for the combined influence of the film cracks and reduced plasticity-induced closure. The physical meaning behind this approach is shown schematically in Figure 8 together with the other relevant stress intensity ranges. Stress Intensity K max + Kfc K max ∆ K eff ∆K

(∆Ktip )

a

∆Ktip ) (∆K )+ a tip a

K cl K cl* K min

Frequency

Figure 8: Influence of reduced plasticity-induced crack closure and film crack effects on crack tip stress intensity range.


The contribution made by (∆Ktip)a+ is only "switched on" when the main crack intersects intermittently with a film crack. Evaluating equation (11) allowed the combined effect of reduced plasticity-induced closure and the presence of surface cracks in the oxide film to be included in the calculated stress intensity range. The results obtained using this procedure are shown below in Figure 9.

5.

ACKNOWLEDGEMENT

Gratefully acknowledged is the support of the Defence Research Agency, Farnborough.

6.

REFERENCES

1. Ministry of Defence, Defence Standard 03 - 24, Issue 2, Directorate of Standardisation, Procurement Executive, MOD, Glasgow, UK, 1988. 2. Ministry of Defence, Defence Standard 03-25, Issue 2, Directorate of Standardisation, Procurement Executive, MOD, Glasgow, UK, 1988. 3. G.W. Stickley and F.M. Howell. ‘Effects of anodic coatings on the fatigue strength of alloys’, Proceedings of American Society for Testing and Materials, 50, 735, 1950.

Figure 9: Fully corrected Paris-Erdogan plots for anodized and control specimens. These results demonstrate very clearly that the observed fatigue crack growth rate for anodized material can be correlated to a high degree of accuracy (R2 = 0.99) with the applied stress intensity range. Within the confines of the proposed model, this confirms that reduced plasticityinduced closure coupled with the presence of surface film cracks can explain the accelerated fatigue crack growth rates observed in boric-sulphuric acid anodized aluminium alloy.

4.

CONCLUSIONS

The results of this study have confirmed that fatigue crack growth in aluminium alloys can be markedly affected by the presence of a thin anodic oxide film, with enhanced growth rates of up to 35% being possible in the anodized condition. The increase in the crack growth rate can be explained in terms of two concurrent mechanisms. Firstly, constraint of the substrate material related to tensile residual stress in the oxide film, and its inherent higher modulus, results in a significant reduction to the plasticity-induced closure experienced by the growing fatigue crack. Secondly, the highly cracked surface morphology of the oxide film in the vicinity of the crack tip, alters the growth characteristics of the propagating fatigue crack by increasing the stress intensity local to the crack tip. The combined influence of these two mechanisms results in a faster growing crack, despite an observed increase in crack path tortuosity. These effects are likely to be most prevalent at low values of ∆K where the influence of crack closure is greatest and may occur, given the necessary prerequisite conditions, in other coating systems.

4. S.E. Larssen. ‘The influence of anodizing processes on the fatigue strength of aluminium alloys in a noncorrosive environment’, Proceedings of 8th I.C.A.F Symposium. Lausanne, Switzerland, June 1975. 5. R.G. Rateick, T.C. Binkowski and B.C. Boray. ‘Effect of hard anodize thickness on the fatigue strength of AA6061 and C355 alumnium’, Journal of Materials Science Letters, 15, 1321, 1996. 6. R. L. H. Wanhill, ‘The effects of cladding and anodizing on flight simulation fatigue, NLR Report No. TR85006U, National Aerospace Laboratory, The Netherlands, 1985. 7. E. Abramovici, P. Leblanc and B. Weaver, ‘The influence of etch pits on the fatigue life of anodized aluminium alloys, Proceedings of the International Conference and Exhibits on Failure Analysis, Montreal, Canada, July 1991. 8. R. Koop and Y. Moji, ‘Boric/sulphuric acid anodize – an alternative to chromic acid anodizing, S.A.E Technical Report No. 920944, Warrendale, U.S.A, 1992. 9. C.E Alvey, G.C. Wood and G.E. Thompson (1980), ‘The mechanical properties of porous anodic films formed on aluminium’, Proceedings of the 10thWorld Congress on Metal Finishing, Kyoto, Japan, S. Haruyama ed., Metal Finishing Society of Japan, 275-280, October 1980. 10. R. L. H. Wanhill ‘Effects of cladding and anodizing on flight simulation fatigue of 2024-T3 and 7475T761 aluminium alloys’, NLR Report No. TR85006U, National Aerospace Laboratory, Netherlands, 1985. 11. A.V. Karlashov, R.G. Gainutdinov and A.T. Pankov, 'Comparative study of the effect of anodizing and of



cladding followed by anodizing on the cyclic strength of D16T sheet material', 12 (1), 76-79, 1976. 12. T. Torri and K. Honda, Fatigue crack growth testing of films using pre-cracked base plates’, Advances in electronic packaging, American Society of Mechanical Engineers, USA, 1992. 13. Y.G. Li and J.S. Qiao, ‘Effects of brush electroplating and shot peening on fatigue strength of a medium carbon steel’, Fatigue and Fracture of Engineering Materials and Structures, 15, 431-436, 1993. 14. A.M. Cree, G.W.Weidmann and R. Hermann, ‘Filmassisted fatigue crack propagation in anodized aluminium alloys’, Journal of Materials Science Letters, 14, 1505-1507, 1995. 15. P.C. Paris and F. Erdogan, ‘A critical analysis of crack propagation laws’, Journal of Basic Engineering, 85, 528-534, 1963. 16. W. Elber, ‘Fatigue crack closure under cyclic tension’, Engineering Fracture Mechanic, 2, 37-45, 1970. 17. S. Suresh and R.O. Ritchie, ‘ A geometric model for fatigue crack closure induced by fracture surface roughness’, Metallurgical Transactions, 24, 6, 803819, 1986. 18. F. Erdogan and G.C. Sih, ‘On the crack extension in plates under plane loading and transverse shear’, Journal of Basic Engineering, 85, 519-527, 1963.

7.


Eur Ing Dr Alistair Cree, CEng CPhys MInstP FIES is a Senior Lecturer, and Team Leader, in Engineering Science and Ship Technology at the Britannia Royal Naval College, Dartmouth. His research interests include fatigue in engineering materials, the development of microstructure during heat treatment of ferrous alloys and the use of marine environmental data for the fatigue assessment of ships structures. Dr George Weidmann, CEng FIMMM is a Senior Lecturer and Staff Tutor in Technology with the Open University. His research interests include the fracture of inorganic, polymeric and metallic glasses, mechanical properties of composites and fatigue in coated materials.



COMPOSITE OVERLAY FOR FATIGUE IMPROVEMENT OF A SHIP STRUCTURE G Guzsvány and I Grabovac, Defence Science and Technology Organisation, Australia SUMMARY In this paper the impact of a composite overlay on stress levels and consequently fatigue performance of a superstructure on a Royal Australian Navy (RAN) vessel is investigated. The Finite Element Methodology (FEM) using a ‘top-down’ approach is applied to model the structural response. For the overlay modelling, a thin composite plate and zero bending assumptions are applied. Sea loads are modelled by the effects of a characteristic hogging-sagging wave and the corresponding probability of encountering such a wave over the lifetime of the ship. The stress distribution in the structure due to the composite overlay is approximated using the Weibull probability distribution and the fatigue damage coefficient estimated by adopting the Palgrem-Miner Model. The results show that a composite overlay in the highstress concentration area can effectively modify the stress distribution and fatigue damage accumulation in the supporting structure thus reducing the likelihood of a fracture initiation. The impact of overlay thickness on the stress pattern is also investigated. By increasing the number of plies in the overlay, the fatigue damage in the critical region has been found to be reduced. However, increasing the number of plies tends to intensify an accumulation of fatigue damage at the overlay ends. The technique, developed and presented in this paper can be used to determine the location, the number of plies and the overlay orientation to optimise overlay effectiveness.

NOMENCLATURE

Pp,l( )

Significant wave height Hs Zero crossing period Tz F(Hs, Tz) Percentage of time the ship spends in sea state characterised by Hs, Tz T Normalized wave period R Normalized wave amplitude Characteristic period τCh Normalized characteristic wave period TCh (τCh/(1.0897 Tz) Characteristic wave amplitude ACh Normalized characteristic wave amplitude RCh (ACh/0.354 Hs) Characteristic lifetime probability of excedence FCh,l λ Harmonic wave length R Normalized wave amplitude T Normalized wave period ν Spectral width parameter Correction function for broad banded process L(ν) Ship length between perpendiculars Lpp B Ship breadth at water line T Ship Draft ∆ Displacement of the ship C Fatigue constant of the structural detail material property m Fatigue exponent of the structural detail material property Constant amplitude stress cycle number to crack N initiation s Stress range Stress range at N=107 where s-n curve s* properties change ds Stress range differential Characteristic stress range resulting from the SCh characteristic hogging-sagging wave Life time total number of stress cycles from all NL stress range bins

DFAT k A E( ) γ( ) Γ( )

Lifetime probability density function of stress range Fatigue damage ratio Weibull distribution shape parameter Weibull dist. scale parameter A=-SCh ln(FCh,l)-1/k Statistic expectation function Lower incomplete Gamma function Upper incomplete Gamma function

1.

INTRODUCTION


This paper presents a recent study of the effect of a composite overlay on the fatigue performance of a ship superstructure. The focus of this work is at the crackprone knuckle region of an FFG class vessel [1] for which the fatigue damage accumulation is modelled and compared in two configurations: (i) superstructure without overlay and (ii) with overlay present. In this study the following engineering areas are addressed using mostly probabilistic concepts: • • •

Sea loads hydrodynamics Ship structural response Fatigue damage accumulation (Fatigue Analysis techniques)

In order to calculate the benchmarking load used in comparison of the two configurations of the ship structure, the model of the Characteristic HoggingSagging Wave is introduced. The model neglects the effect of the overlay on the ship motion response and consequently on the sea loads over the lifetime of the ship.


The fatigue damage accumulation is modelled using FEM techniques. The model is validated using the strain gauge measurement during the sea trials. The fatigue accumulation patterns in the vicinity of the knuckle region are obtained using the validated FEM model. Based on the calculated fatigue accumulation patterns the importance of level ageing in relation to the overall ageing of the structure is discussed. 2.

BRIEF HISTORY

Two carbon fibre (CF) composite overlays (5m x 1m) were installed onboard an FFG Class frigate in 1993 to address persistent fatigue cracking of the aluminium superstructure at the knuckle, 02-deck, port and starboard, Figure 1. The work was performed by the Defence Science and Technology Organisation (DSTO) for the Royal Australian Navy (RAN) [2]. The initial design of the composite overlay was primarily based on an empirical approach. Traditional solid mechanics principles were used in combination with the crack repair history to engineer the composite reinforcement that was to prevent the reoccurrence of the cracks. Composite Overlays Weld Forward

For the accurate prediction of fatigue crack occurrence, it is required to conduct seakeeping analysis for all possible sea conditions and assess their lifetime impact on cracks in a ship structure [3]. The aim of this study however, was to compare the performance of two structural configurations over an invariable life-time sea-load, rather than to accurately predict the crack occurrence. Hence, the complex seakeeping analysis was replaced by a more appropriate benchmarking approach. The fatigue performance calculation of the two structural configurations (without and with the overlay) was based on a characteristic harmonic wave i.e. hogging-sagging wave. The hoggingsagging wave loading condition is traditionally used in preliminary design, initial classification compliance check, for screening structural weak points, etc. It is relatively simple to model and it includes the worst case scenario i.e. quasistatic hull–girder bending condition analogous to a beam supported at the ends or in the middle. Dahle et al. [4] justified the use of simplified sea load models in the early stages of the ship design process. In the hull girder analysis, the maximum wave bending moment on a ship generally occurs at a wavelength to ship-length ratio of λ/Lpp=1 which corresponds to quasistatic hogging-sagging condition. For this reason, the characteristic benchmarking wave load in this study was based on the above maximum bending moment condition. The statistical approach derived below follows the concepts described in [4]. The rational behind replacing the random sea wave with a single characteristic regular wave can be understood by considering the following: •

Figure 1: Location of composite overlays on the superstructure This new overlay technology for the RAN was demonstrated in service over a seven-year trial period after which the outcome was considered as a success that met Navy’s objective. In that time no aluminium cracking at the knuckle (frame 196) was reported and the reinforcements have remained in place to provide continued service. 3.

CHARACTERISTIC SEA LOAD MODEL

To demonstrate the fatigue improvement of the ship superstructure after installing the composite overlay, the lifetime operational sea loads need to be characterised and modelled since they are considered as the primary drivers of the fatigue process.

•

The short term stationary random waves as well as structural responses can be considered to be composed of a number of simple harmonic wave functions. Within the approach of Dahle et al. [4] we only consider the amplitudes of components within a narrow frequency band. The central frequency of that band corresponds to the hogging-sagging wave frequency.

Using the fundamental equation for harmonic waves and meeting the λ/Lpp=1 condition, the period for the characteristic hogging-sagging wave can be found as: τ Ch = 2πλ / g = 2π Lpp / g . The characteristic wave amplitude ACh, is usually chosen by an analyst to be Lpp/20 or a similar design criterion. Despite the ACh and τCh being fixed for the hoggingsagging condition, the normalised characteristic values i.e. RCh=ACh/(0.354Hs) and TCh=τCh/(1.0897Tz) change with sea definition (Hs and Tz).



The life-time probability of exceeding the characteristic bending moment, resulting from the hogging-sagging condition was obtained from the following equation (1):  ∞ 1 05⋅TCh  FCH , L = ∑∑ F ( H S , Tz ) ⋅  ∫ ∫ P( R, T ) ⋅ dTdR  ...... ( 1) Hsi Tz j  RCh 0 95⋅TCh 

In evaluating equation (1), F(HS,TZ) was based on the standard operational profile of RAN vessels which was defined in reference [5]. The function P(R,T) was obtained from equation (2) which was derived by Longuet-Higgins [6]:

P ( R, T ) =

R 2 − R2 [1+ (1−1/ T )2 /ν 2 ] ⋅e ⋅ L (ν ) (2) π 1/ 2ν T 2 2

⋅

A more accurate calculation of the life-time stress-range statistics may include all short term stress response statistics including that of the transient load response. This can be achieved by conducting sea keeping and Finite Element (FE) analysis for all sea conditions defined by the RAN operational profile and by performing statistical analysis of the results. 4.

STRUCTURAL MODEL

In order to quantify the fatigue improvement after applying the composite overlay to the crack-prone area of the superstructure, the structural response to the benchmarking characteristic hogging-sagging wave was obtained using an FE analysis and a ‘top-down’ approach. This method uses a coarse global model combined with a fine grid local model. The global FE model was developed according to the Maestro1 global ship modelling conventions and the local FE structural model for the detailing of the superstructure according to Trident2 conventions. The advantage of the ‘top-down’ approach of the superstructure (i.e. separating the course global model of the entire ship from the local model), lies in the efficiency of modelling, solving and post processing of the FE problem. The global and the local models are connected through boundary forces acting at the common boundary nodes. This way the boundary forces are obtained in the global FE analysis and transferred to the local model as displacements. 4.1.

the global model for the FFG frigate in this work are shown in Table 1. FFG frigate - Basic details Ship length between 125.8 m perpendiculars, Lpp Ship breadth at water line, B 13.7 m 5.05 m Ship Draft, T 4290 t Displacement of the ship, ∆ Table 1: Principal data of the FFG frigate (Adelaide Class). The breakdown of the element types used in the model is given in Table 2. The influence of stiffeners on the plate elements was idealised (smeared stiffeners) using stiffened panel elements. The frames, beams and girders were formed using 4 node beam elements. The end fixity of the above elements was specified by the end-bracket parameters. Element Type

Number of Elements

Girder Strake panel Transverse Beam Pillar Additional Beam Triangular plate Table 2: Global model details.

213 3268 1116 206 287 798

Figure 2 shows the Maestro global model of the ship structure subjected to bending modes due to hydrodynamic loading.

Area of the local model analysis

GLOBAL MODEL

The purpose of the global model is to screen the principal compliance of the ship primary structure with the design requirements as well as to locate the areas of high stress concentration. These areas can subsequently be analyzed using a refined local model. The principal dimensions of 1 Maestro is the trade-name of FE software created by Proteus Engineering, US 2 Trident is the trade-name of FE software created by Martec Limited, Canada


Figure 2: Hogging-sagging conditions of the FFG frigate hull (global model). The mass distribution corresponding to the nominal displacement weight of the ship in this paper was defined as the longitudinally distributed sectional weight and spread over the sectional nodes in accordance with the Maestro conventions. Based on the above nominal weight distribution of the ship, the hydrodynamic load


conditions were defined. For each of the above conditions the quasi-static equilibrium between ship weight and the buoyancy force was applied. 4.2.

LOCAL MODEL

The refined area of the superstructure included a part of the 02-deck structure (port and starboard) spanning between Frames 188 and 212. A breakdown of the elements used in the local model is given in Table 3 while Figure 3 shows the configuration and the details of the refined local model. Element Type

Number of Elements Triangular plate 204 Quad shell 9954 18 Bar Table 3. Local model details.

Forward

4.3.

FATIGUE MODEL

DAMAGE

ACCUMULATION

The cyclic loads acting on the ship hull over the lifetime of the ship are the cause of the fatigue damage accumulation in the ship structure. Therefore, the benchmark statistics for the cyclic stress (stress ranges) was integrated into the traditional Palgrem-Miner fatigue initiation model. Generally, the total stress response of a structural detail has to account for primary hull girder bending loads/responses, secondary substructure-level loads/responses as well as tertiary local loads/responses. In terms of ‘top-down’ analysis this was achieved by the characteristic stress response obtained from a reasonably detailed model and the assumption that the local loads would be absorbed by the local reinforcement. The local model in this work was considered appropriately detailed, whereas the tertiary response for this study was not considered important. A more comprehensive study of the dynamic response may require solving the FE model using the spectral sea load model including transient loads, which ideally solve the structural and hydrodynamic problem simultaneously (hydro-elastic solution) instead of using the benchmark characteristic sea load. In order to determine the fatigue damage ratio, in this analysis the fatigue limit of the material was formulated as follows (3):

sm

(a)

C

= 1/ N .............................................................( 3)

Where a nominal fatigue property of the structural weld type-23 was selected in this study defined by m=3 and log C=9.91 [7]. The lifetime stress range probability distribution was modeled using the two parameter Weibull probability density function as follows (4):

Pp ,l ( s ) =

k s Ak

k −1 − ( s / A ) k

e

........................................( 4)

The scale parameter A was calculated using the hoggingsagging stress range SCh of the structural detail in question obtained from FE analysis and the corresponding lifetime probability FCH , L of RCh being exceeded.

(b) Figure 3: Local FE model : (a) Area shown in Fig 2., (b) Mesh definition.

Calculation of the Palgrem-Miner fatigue damage ratio was based on the characteristic stress range SCh and a set number of expected stress cycles, NL= 108, over the lifetime of the ship. For a two segmented S-N curve the following formulation, (5) was derived: ∞ m ∞ m  s sm  .... ( 5)  sm  s s *

DFAT = NL ⋅ E  = NL ∫ Pp,l (s )ds = NL  ∫ Pp,l (s)ds + ∫ Pp,l (s)ds  0 C  C C C 0 s*  



By combining (4) and (5) above, the fatigue damage ratio of a particular structural detail was obtained (6):

DFAT

4.4.

* ∞ m m m * k * k  s s m k k −1 − ( s / A)k s k k −1 − ( s / A)k   A 1  m1  s   A 2  m2  s    ..... ( 6) ,    s e ds + ∫ s e ds  = N L  = NL ∫ ⋅ γ 1 + ,    + ⋅ Γ 1 + k   C Ak C k  A   C2 k  A    0 C A  s*     1 

COMPOSITE OVERLAY PROPERTIES

The composite overlay was modelled using thin plate elements. To model the adhesive connection between the aluminium surface and the composite overlay, both the composite overlay mesh grid-points and the mesh gridpoints of the supporting aluminium structure were selected identical and connected. The properties of the carbon laminate used in the FE analysis are given in Table 4. Longitudinal modulus of elasticity, (E1 ) Transverse modulus of elasticity, (E2) Shear Modulus of elasticity, (G) Poisson’s ratio, (µ) Ply thickness

140 GPa 5 GPa

For example, for those design details where high stress concentrations and consequently the fatigue damage is predicted, the composite overlays could offer a reduction of stress concentration of the design-detail and yet, not exceed weight requirements. The principle of ‘even ageing’ in the structure (levelling ageing) can be achieved by smart distribution of overlays in the ship structure which may also provide a solution to a more cost effective through-life ship structural maintenance.

10 GPa 0.25 0.3 mm

Table 4: Properties of the Carbon-vinyl-ester laminate.

5.

Furthermore, this analysis provides an insight into the structural fatigue process which is of particular interest to ship operators. Advanced computational techniques can nowadays provide a useful prediction of the accumulated fatigue damage allowing inclusion of the fatigue process into the design, operation and maintenance optimisation.


The benchmark method introduced here is rudimentary. It is, however considered, that the calculated structural response trends to characteristic global loading are sufficiently accurate.

Lastly, the information obtained from the analytical tools could be used to optimise the design and the strength of an overlay to achieve the desired strength effects while minimising the cost of the application.

6.3E01

7.3E1

3.5E01 0.0E0

0.0E00

-5.2E1

(a)

-1.1E2

Knuckle

(b)

-2.8E01 -6.3E01

2.9E01

5.4E01 02

1.5E01 2.2E01 01

0.0E00 0.0E00 -1.5E01 -2.2E01

(c)

-4.3E01

(d)

-3.6E01

Figure 4: Stress distribution output for hogging (a & b) and sagging (c & d) condition –port side; (a & c) without and (b & d) with composite overlay; dotted line represents the position of the overlay (Stress units: MPa) © 2006: The Royal Institution of Naval Architects


5.1 STRESS ANALYSIS RESULTS AND COMPARISON WITH EXPERIMENTAL DATA The stress distribution of the superstructure area given for the local model (Figure 3) is shown for the port side in Figure 4. Figure 4 is composed of four separate analytical results addressing both hogging and sagging conditions for the original and modified structure which includes the presence of the composite overlay. By comparing each load condition, the change of the general stress distribution pattern due to the overlay effect clearly indicates the disappearance of the high stress concentration at the superstructure knuckle area, Figure 4(b & d). For example, in the knuckle region the calculated stress range between sagging (compression) and hogging (tension) is reduced from 71.1MPa to 57.9 MPa. This amounts to about an 18.6% reduction in stress. In other areas adjoined to the overlay position such as the forward and aft end, the calculated increase in hogging-sagging stress range is from 58.9 MPa to 70.9 MPa at the forward end, and 50.3 MPa to 60.4 MPa at the aft end. These

changes in the stress condition amount to an increase of 20.4 % and 19.8% respectively. A relief in hogging-sagging stress range is also predicted at superstructure sides both at 01- and 02-deck levels. At 01-deck level the stress range is reduced from 19.4 MPa to 16.2 MPa amounting to a calculated reduction of 16.5 %. At the 02-deck level a reduction of the hoggingsagging stress range from 45.1 MPa to 38.6 MPa is calculated, which amounts to a reduction of 14.4 %. The positions at which the above results were obtained in the local model correspond to the locations of the straingauges used in sea trials, Figure 5. The strain gauge measurements were recorded at sea on the original (unmodified) structure as well as overlay fitted structure and the measurement results are reported in [8]. This experimental data was used for the calculation of the stress variation and consequently the structural fatigue loading in the structure under random wave loads. In Table 5, the characterised stress range trends are given, of both the experimental and analytical data. The results of the experimental and analytical data presented in Table 5 are in reasonably good agreement.

G11 G12

G10

G20

G15

Figure 5: Redistribution of the stress intensity level on port and starboard after installation of composite overlays [7] (Strain gauges G15 & G20 were located on the corresponding positions, port side 01- and 02deck respectively).



EXPERIMENTAL DATA (Sea trials) Strain Gauge Measured Stress Change, Location, (%) (see Fig. 5) G10 G12 G11 G20 G15

ANALYTICAL DATA (Local model) Stress at corresponding Strain Gauge Locations, PORT side Without overlay, With overlay, (MPa) (MPa)

-18 71.1 57.9 +18 58.9 70.9 +18 50.4 60.4 -13 45.1 38.6 -15 19.4 16.2 Table 5. Comparison of experimental and analytical data.

(a)

Calculated Stress Change, (%) -18.6 +20.4 +19.8 -14.4 -16.5

1.5E 00

1.9E 00

1.2E 00

1.5E 00

8.4E-01

1.1E 00

5.1E-01

5.4E-01

8.4E-02

(b)

1.1E-01

1.8E 00

1.6E 00

1.4e 00

1.2E 00

9.1E-01

8.4E-01

5.0E-01

4.4E-01

(c)

(d) 1.0E-01

8.9E-02

Figure 6: Fatigue damage change due to altered stress distribution – port side (a & c) without and (b & d) with composite overlay, dotted line represents the position of composite overlay (Scale: Dimensionless units)

5.1.

FATIGUE ANALYSIS RESULTS

The distribution of fatigue accumulation on the local model was determined following the procedure outlined in Section 4.3. The fatigue analysis included calculations for both configurations of the superstructure (without and with the composite overlay). The results are shown in Figure 6 and the values presented in Table 6, for corresponding strain gauge positions as identified in Figure 5 From Figure 6 it is evident that the accumulated fatigue (Fig. 6a) is reduced in areas adjacent to both sides of the overlay (Fig. 6b). Similarly, in the same location the structural framework (Fig. 6c) shows a reduction in fatigue accumulation underneath the overlay (Fig. 6d). However, as observed in the stress analysis (Section 5.1), increased accumulation of fatigue is present at the


overlay front and rear end. The side of the superstructure down to 01-deck is also relieved from fatigue accumulation resulting in a delay of crack initiation in that area. The above trends are supported by data presented in Table 6. While the knuckle area and superstructure sides receive some reduction in fatigue accumulation, it is noticed that a significant increase in fatigue occurs at overlay ends. This procedure therefore could assist in the design and strategic location of the composite overlay(s) by predicting the areas of high fatigue accumulation within the ship structure.

Stress Range, MPa

Advanced Marine Materials & Coatings, London, UK 90

90

85

85

80

80

75

75

70

70

65

65

60

60

55

55

50

50

45

45

40

40

Stress at Knuckle Stress at Ends

35

35 30

30 0

5

10

15

20

25

Overlay Plies

Figure 7: Effect of overlay strength on stress in the structure

Fatigue Level

0.8

0.8

Fatigue at Knuckle Fatigue at Ends

0.7

0.7

0.6

0.6

0.5

0.5

0.4

0.4

0.3

0.3

0.2

0.2

0.1

0.1 0.0

0.0 0

5

10

15

20

25

Plies of C arbon Fibre Reinforcem ent

Figure 8: Effect of overlay strength on fatigue of structure In order to carry out this optimisation the relationship between overlay design and the stress and fatigue accumulation rate in the structure need to be determined. One of the possible optimisation objectives that is apparent from this case study is to achieve balanced ageing of structure at both the knuckle area and the area adjacent to the overlay ends. In this context the optimum overlay strength (number of plies) that would yield the required effect could be obtained using the data as shown in Figure 7.

Fatigue Damage Accumulation Change, Without With (%) Overlay Overlay G10 0.29 0.26 -10 G12 0.38 0.92 +142 0.25 0.65 +160 G11 0.63 0.38 -40 G20 0.084 0.08 -5 G15 Table 6: Changes in fatigue damage accumulation attributed to composite overlay. Reference Point

5.3.

EFFECT OF OVERLAY PARAMETRIC STUDY

STRENGTH

-

The general trends presented in the preceding Sections on stress distribution and fatigue accumulation in the structure could potentially be used to optimise the design of a composite overlay. As shown in Figures 4 - 6, the overlays provide a means of redistributing stresses and changing a local map of fatigue accumulation. While reduction in the critical regions prone to fracture is beneficial, the increase in stress and fatigue accumulation at both ends is undesirable, yet the adverse effect can be minimised through an optimisation process.

Presently, the superstructure is fitted with composite overlays consisting of 25 plies of carbon fibres. No cracking in areas adjacent to the overlays has been reported; therefore it appears that the cracking has been contained. However, referring to data in Figure 7, some refinement in balancing the stress level between the knuckle and the ends is achievable by using the overlay consisting of only 10 – 18 plies of carbon fibres (assuming material of same properties is used).



Similar observation can be made on interdependency between fatigue and overlay thickness. As shown in Figure 8, an overlay comprising about 10 – 15 plies would impart more balanced level of fatigue accumulation at the knuckle and the overlay ends. In summary, the retrospective modelling results of overlay effect on the ship structure are shown to be in good agreement with the experimental data. The original overlay design was performed according to traditional engineering principles which include somewhat conservative approach to safety factors.

Future work may involve a study of the overlay structural response. In this paper the overlay was modelled by thin plate elements neglecting the bending and shear stresses. By refining the model it is aimed to predict the fatigue damage accumulation within the laminate that may lead to different failure mechanisms (i.e. disbonding at the overlay/metal interface, delamination, fibre fracture, etc.). Potentially, the standardisation of the repair overlays for the permanent and emergency repair of structural damages is also of interest.

However, the numerical methods can offer a greater freedom in overlay optimisation. For example, once the stress histories at both the knuckle area as well as the structure at overlay ends are known, the stress and fatigue accumulation ‘hot-spots’ can be shifted to an area where still sufficient fatigue life remains. Thus by overlay optimisation and designing fatigue accumulation take-over, a structural design is achievable where the stress and fatigue accumulation in the structure is evenly distributed.

7.

In future applications, this approach could extend the service life of all structural hot-spots and therefore offer the operator a reduced through-life cost of ownership.

[2]

6.

Authors wish to acknowledge Drs. Craig Gardiner, Nigel StJohn, Robert Hughes and Craig Flockhart for reading the manuscript and providing valuable comments.

8.

REFERENCES

[1]

Directorate of Naval Ship Design, ‘FFG Aluminium Superstructure Cracking’, Technical Memo: A012695, Canberra, Australia, 1986. GRABOVAC I., PEARCE P.J., CAMILLERI A., CHALLIS K. and LINGARD J., ‘Are Composites Suitable for Reinforcement of Ship Structures?’, Paper presented at the 12th International Conference on Composite Materials (ICCM-12), Paris, France, July 5-9, 1999. GU X.,K., AND MOAN, T., ‘Long-Term Fatigue Damage of Ship Structures Under Nonlinear Wave Loads’, Marine Technology, Vol. 39, No. 2, April, pp. 95-1004, 2002. DAHLE E.A., MYRHAUG D., AND WIST H.,T., ‘Maximum Wave Bending moment –back to basics?’, The Naval Architect, , London, UK, September 2002. Royal Australian Navy, ‘Standard Materiel Requirements for RAN SHIPS and Submarines’, Volume 3: Hull System Requirements, Part 6: Seakeeping, Document No. A016464., Canberra, Australia. LONGUET-HIGGINS M.S., ‘On the joint distribution of wave periods and amplitudes in a random wave field’, Proceedings of Royal Society London, A389, pp. 241-253, 1983. FRICKE W., PETERSHAGEN H., AND PAETZOLD H., ‘Fatigue Strength of Ship Structures’, Part 2: Basic Principles, Germanishe Lloyd, Hamburg, Germany, 1997. PHELPS B.P., ‘Bonded Repairs to RAN FFG Superstructure-Strain Gauge Data Analysis’, DSTO Report DSTO-RR-0046, Melbourne, Australia, 1995.

CONCLUSION

The work presented above is seen as a step forward in general understanding of the relationship between the effect of composite overlay and the response of ship structure. A number of engineering areas were addressed in an attempt to illustrate the multidisciplinary nature of the fatigue analysis. It was shown that by combining the FE top-down analysis with the quasistatic loading and including the statistical characterisation of the lifetime loads, realistic fatigue accumulation figures can be obtained. The modelling results were in good agreement with the experimental data, therefore the methodology used in this analysis could be applied in modelling the effects of composite overlay on a metal structure. This information could assist an engineer in designing and maintaining a level-ageing structure and therefore contribute to the overall process of ship management in reducing the cost of maintenance.

[3]

[4]

[5]

[6]

[7]

[8]


ACKNOWLEDGEMENTS


9.


Gáspár Guzsvány is a naval architect specialising in structural analysis and hydrodynamics of surface ships. His work in fatigue analysis of welded structures and in extreme lifetime load analysis provides DSTO with structural adequacy assessment capability. He is also contributing to DSTO capability development in the area of composite fatigue modelling. Ivan Grabovac is a Senior Research Scientist at DSTO with over thirty years of experience and expertise in adhesives, resins and composite materials development for aerospace and maritime applications. He was the principal investigator and is responsible for the development, installation and monitoring of carbon fibre composite overlays now in service on a navy ship.



AUTHORS’ CONTACT DETAILS Raouf Kattan Safinah Ltd, UK [email protected] Pramod Kumar Mazagon Dock Ltd., India [email protected] Morten Sorensen MC Technology, Belgium [email protected] Joao Azevedo Euronavy, Portugal [email protected] Rob Mutton University of Newcastle upon Tyne, UK [email protected] Guy Seabrook Magellan Companies inc., USA [email protected] Brian Glover Alocit Systems Ltd., UK [email protected] Arran Flowerday University of Newcastle upon Tyne, UK [email protected] Peter Wright University of Newcastle upon Tyne, UK [email protected] Tim Davison Enviropeel Systems Ltd., UK [email protected] Richard Brown University of Rhode Island, USA [email protected] Alistair Cree Britannia Royal Naval College, UK [email protected]

© 2006: Royal Institution of Naval Architects

Ivan Grabovac Defence Science and Technology Organisation, Australia [email protected]

Composite Overlay for Fatigue Improvement of a Ship Structure Gáspár Guzsvány, Ivan Grabovac Defence Science and Technology Organisation RINA, Advanced Marine Materials & Coatings, London, UK 22- 23 February 2006

Overview

DSTO and the Defence of Australia FFG class frigate superstructure investigation Fatigue modelling Operational profile and random sea states Characteristic Sea Load model Finite Element Model of the structure Fatigue Damage Accumulation model Results Structural maintenance recommendations Concluding remarks

DSTO and the Defence of Australia Mission: Expert, impartial and innovative application of science and technology to the defence of Australia and its national interests.

Vision: Leading science in Australia's defence.

Values:

Excellence - We pursue scientific excellence. People - We value our people and help them to excel. Innovation - We seek innovative solutions for Defence. Integrity - We provide expert, impartial advice and we stand by it. Teamwork - We work together.

DSTO Resources

The People Platforms Sciences Laboratory Systems Sciences Laboratory Information Sciences Laboratory Headquarters* Total staff Staff with PhD *includes HQ staff based in laboratories.

Further Information: http://www.dsto.defence.gov.au

554 838 529 302 2223 ~650

The Royal Australian Navy’s challenges projected to DSTO – MPD tasks Increasing parent-navy responsibility of the RAN defines the Navy’s role in risk management of the fleet. Due to its strong OH&S culture the RAN is committed to safe work place policy. The RAN is increasingly conscious about the through life cost of ownership of its ships. RAN is increasingly taking part in diverse missions.

FFG class frigate structural problem

FFG frigate - Basic details Ship length between perpendiculars, Lpp

125.8 m

Ship breadth at water line, B

13.7 m

Ship Draft, T

5.05 m

Displacement of the ship, ∆

4290 t

Composite repair of structural cracks

Strain gauge 2

Strain gauge 1 (fore)

Other structural risk management systems

Regular Survey Classification based risk management Regular repair of fatigued structural parts Stress monitoring Fatigue monitoring Fatigue prediction and crack prevention

Modeling Synopsis Given: Operational profile Hull structure Material fatigue property

Objective: Fatigue damage over the lifetime of the ship Stress patterns with and without the overlay Overlay optimisation suggestions

Calculation models: Characteristic wave model FE model of the structure Miner rule of fatigue damage accumulation

Fatigue modeling procedure

Operational Profile Average Lifetime Exposure

Sea Load Model: Hogging-Sagging Wave

Structural Model:

Quasi-static FE model

Fatigue Damage Miner rule

Operational Profile definition

Operational Profile

Sea Load Model: Hogging-Sagging Wave

Structural Model:

Quasi-static FE mode


RAN has developed an average lifetime wave scattering table combining all ocean areas it operates in and the history of fleet activities. This average wave scattering data is used for lifetime service simulation.

Sea Load model simplification Pros: Calculation reduction DLA compatibility

Cons: Non-linearity Transient loads Viscous effect

Operational Profile Average Lifetime Exposure

Sea Load Model:

Structural Model:

Quasi-static FE mode


Rationalization: As validation data is available for linear “quasi-static” conditions only, the cons are considered irrelevant. Outcome: VBM DLP is selected to define characteristic sea loads, applying known quasi-static conditions for maximum loading.

Characteristic Sea Load definition Maximum VBM generally occurs at hogging sagging condition. Probability of excedens of the hogging-sagging condition drives the probability of exceeding stresses generated by VBM. The probability of exceeding the generated stress is calculated from lifetime random sea surface statistics. Most severe condition: head sea, following sea.

Characteristic Sea Load definition τ Ch = 2πλ / g = 2π Lpp / g

ACh RCh = 0.354 H S TCh =

FCH , L

τ Ch 1.089TZ

 ∞ 1.05⋅TCh  = ∑∑ F ( H S , Tz ) ⋅  ∫ ∫ P( R, T ) ⋅ dTdR  Hsi Tz j  RCh 0.95⋅TCh 

R 2 − R 2 [1+ (1−1/ T )2 /ν 2 ] P ( R, T ) = 1/ 2 ⋅ 2 ⋅ e ⋅ L(ν ) π ν T 2

Structural model simplification Pros: Calculation reduction DLA compatibility Simplicity in validating the model

Cons: Losing stress concentrations Dynamic effects Nonlinear effects

Operational Profile Average Lifetime Expo

Rationalization: Dynamic and non-linear effects negligible Sea Load Mode at this stage, however stress concentration is important for fatigue Structural Model damage accumulation. Outcome: Top-down FE model selected to be analysed quasi-statically. Hogging-Sagging

Fatigue Dam Miner rule

Finite Element Model of the structure Global model

Element Type

Number of Elements

Girder

213

Strake panel

3268

Transverse Beam

1116

Pillar

206

Additional Beam

287

Triangular plate

798

Finite Element Model of the structure Local model

Element Type

Number of Elements

Triangular plate

204

Quad shell

9954

Bar

18

Miner’s sums as the Fatigue Damage Accumulation model Pros: Simple Availability of material data (S-N curves) Operational Profile

Cons: Sequencing error Not good for low cycle fatigue Inappropriate for plastic responses

Average Lifetime Exposur

Sea Load Model: Hogging-Sagging Wa

Structural Model:

Quasi-static FE mo

Fatigue Damag

Rationalization: In the comparative study it is reasonable to neglect the low cycle fatigue. In normal operation the structure is not stressed beyond elastic limit. Outcome: A small sequencing error is expected.

Fatigue Damage Accumulation model Adequacy criteria Stress Level of a structural detail is lower than the fatigue endurance limit Stress level of the structural detail such that the lifetime fatigue accumulation is below failure limit.

Log S Structural Property

m=3

K-x

m=5

Log N

107 Life time loading

B

DFAT = ∑ i =1

B

ni =∑ N i i=1

N L ⋅ Pp ,l ( si ) ⋅ ∆s C / Sim

Endurance Limit

Fatigue Damage Accumulation model Analog formulation ∞ m  s sm  s  s s DFAT = NL ⋅ E  = NL ∫ Pp,l (s )ds = NL  ∫ Pp,l (s)ds + ∫ Pp,l (s)ds  0 C  C * C C 0 s   m

∞ m

*

∞ m  s s m k k −1 − ( s / A)k s k k −1 − ( s / A)k  = NL ∫ s e ds + ∫ s e ds  = ... k k C A  0 C A  s* *

DFAT

 Am1  m  s*  k  Am2  m  s*  k   = NL  ⋅ γ 1 + 1 ,    + ⋅ Γ 1 + 2 ,      k  A   C2 k  A    C1    

A = − SCh ln( FCh,L )−1/ k

Results

Calculated results - Stress

6.3E0

7.3E

Hog

3.5E0

0.0E0 0.0E0

Knuckle

-5.2E

-2.8E0

(b)

(a)

-6.3E0

-1 1E2

Sag 2.9E0

5.4E01 02

1.5E0 2.2E01 0.0E00

01

0.0E00 -1.5E0

(c)

-2.2E01 -4.3E01

(d) -3.6E0

Calculated results – Fatigue Damage

Hog

1.5E 00

1.9E 00

1.2E 00

1.5E 00

8.4E-01

1.1E 00

5.1E-01

5.4E-01

(a)

(b) 1 1E 1

Sag (c)

1.8E 00

1.6E 00

1.4e 00

1.2E 00

9.1E-01

8.4E-01

5.0E-01

4.4E-01

(d)

Strain gauge measurement details G11 G10

G12

G20

G15

Calculated effects of the overlay

ANALYTICAL DATA (Local model)

EXPERIMENTAL DATA (Sea trials) Strain Gauge Location, (see Fig. 5)

Measured Stress Change, (%)

Stress at corresponding Strain Gauge Locations, PORT side Without overlay, (MPa)

With overlay, (MPa)

Calculated Stress Change, (%)

G10

-18

71.1

57.9

-18.6

G12

+18

58.9

70.9

+20.4

G11

+18

50.4

60.4

+19.8

G20

-13

45.1

38.6

-14.4

G15

-15

19.4

16.2

-16.5

Calculated effects of the overlay

Reference Point

Fatigue Damage Accumulation

Change, (%)

Without Overlay

With Overlay

G10

0.29

0.26

-10

G12

0.38

0.92

+142

G11

0.25

0.65

+160

G20

0.63

0.38

-40

G15

0.084

0.08

-5

Structural Maintenance recommendations

Level ageing philosophy Within the Fleet Within a Structure

Fatigue damage monitoring system (sea state monitoring) Implementation of the overlays at design level reduces maintenance cost.

90

90

85

85

80

80

75

75

70

70

65 60

0.8

55

45 40

Stres Stres

35 30 0

0.8

Fatigue at Knuckle Fatigue at Ends

50

Fatigue Level

Stress Range, MPa

Level ageing philosophy - optimisation

0.7

0.7

0.6

0.6

0.5

0.5

0.4

0.4

0.3

0.3

0.2

0.2

0.1

0.1

0.0

0.0 0

5

10

15

20

Plies of Carbon Fibre R einforcem ent

25

Fatigue Damage Monitoring System DSTL’s Fatigue logger installed on HMAS Arunta Continuous strain measurements for 18 months New software DERAFAT is used to analyse the data Port wave and Slam 10000

1397 1396

1000 311

105 76

100 48

38

25 1718

13 10

Strain Gauge on longitudinal stiffener

13

1010

10

6

7

4 3

3

2

2

1

1

1

2

1

2

1 1

1

1

1

1

Strain Gauge on longitudinal stiffener Strain

650

610

570

530

490

450

410

370

330

290

250

210

170

90

130

50

10

-30

-70

-110

-150

-190

-230

-270

-310

-350

-390

-430

-470

-510

-550

-590

-630

0.1 -670

Number of occurrences

339

Installation at design stage -Maintenance Cost Cut TOTAL COST OF OWNERSHIP VS RISK MANAGEMENT STANDARDS Total Cost Accumulated

Initial Cost

Low class structure

High class structure End of Life

Future work

Extend measurement data to higher sea states. Sea load modelling in non-linear waves (steep, short crested). Transient load modelling (slamming). Multi-axial fatigue modelling. Fatigue damage modelling within the laminate that may lead to different failure mechanisms (i.e. disbonding at the overlay/metal interface, delamination, fibre fracture, etc.). Standardisation of the repair overlays for the permanent and emergency repair of structural damages.

Concluding Remarks

This methodology could be applied in modelling the effects of composite overlay on a metal structure.

Assist an engineer in designing and maintaining a levelageing structure and therefore contribute to the overall process of ship management in reducing the cost of maintenance.

Questions ?

ENVIRONMENTALLY FRIENDLY MARINE ANTIFOULING ADDITIVES

MAGELLAN COMPANIES, INC. USA

PHYTOCHEMICALS • COMPOUNDS ISOLATED FROM BOTANICAL SOURCES • PREVALENT IN THE HUMAN DIET • ACTIVE AGAINST MARINE FOULING ORGANISMS

TARGET ORGANISMS • • • •

ALGAE CRUSTACEAN MOLLUSK BACTERIA

CHARACTERISTICS OF PHYTOCHEMICALS IN MARINE COATINGS • LEACH OUT OF PAINT & DISPERSE WELL INTO WATER • DO NOT ACCUMULATE TO UNACCEPTABLE LEVELS IN THE ENVIRONMENT

• THE ANTIFOULING EFFECT IS ON THE SURFACE AREA IN IMMEDIATE CONTACT WITH SURROUNDING WATER ONLY • PHYTOCHEMICALS BECOME INERT WITHIN MILLIMETERS OF LEAVING THE COATED SURFACE

MARINE APPLICATIONS

• ANTI-FOULING PAINTS FOR PLEASURE BOATS, SHIP HULLS & MARINE STRUCTURES • SHIP BALLAST BEFORE DISCHARGE • AQUACULTURE SYSTEMS • WATER INTAKE & DISCHARGE COATINGS FOR SUBMERGED PIPE

EXAMPLE OF MARINE PAINT TUTICORIN, INDIA SACRED HEART MARINE LABORATORY

STATIC & DYNAMIC CONDITIONS

12 MONTHS IMMERSION

EXAMPLE OF MARINE PAINT

BATTELLE LABORATORIES FLORIDA, USA

STATIC CONDITIONS

12 MONTHS IMMERSION

FOULING OF WATER INTAKE & DISCHARGE SUBMERGED PIPE ZEBRA MUSSEL INFESTATION $7 BILLION ECONOMIC IMPACT

ZEBRA MUSSEL TEST PANELS

AQUACULTURE SYSTEMS • PRODUCE $53.7 BILLION PER YEAR GLOBALLY IN REVENUES • OVER $1 BILLION IN ANNUAL MAINTENANCE COSTS • 10% LOSS IN PRODUCTION DUE TO FOULING • NO KNOWN SAFE ANTI-FOULING METHODS CURRENTLY IN PRACTICE

AQUACULTURE ROPE TEST 4 MONTHS TOTAL IMMERSION

SUMMARY • DEVELOPMENT OF ENVIRONMENTALLY FRIENDLY ANTI-FOULING ADDITIVES FOR THE MARINE INDUSTRY • PHASE IN PHYTOCHEMICALS WITH CUPROUS OXIDE UNTIL A 100% GREEN ALTERNATIVE IS ACCOMPLISHED

SUMMARY

• TO CONTROL ZEBRA MUSSEL INFESTATION • TO CONTROL FOULING OF AQUACULTURE SYSTEMS WHERE NO SAFE ALTERNATIVE EXISTS • TO CONTROL INVASIVE SPECIES FROM BALLAST DISCHARGE

PARTNERS SOUGHT • MARINE INDUSTRY

• PHILANTHROPIC ORGANIZATIONS • GRANTS • PERSONAL INVESTMENTS

Fatigue Crack Growth in Anodized Aluminium Alloys. Dr. Alistair Cree 1 Dr. George Weidmann 2 1

Britannia Royal Naval College 2 The Open University U.K.

Coatings for protection • Components and structures often fail when their surfaces cannot withstand the external forces or environmental conditions imposed upon them. • Problem can be minimised by using a variety of coating systems which improve physical or mechanical integrity in relation to corrosion, friction and wear. • For aluminium and its alloys protection is usually achieved by anodising prior to service.

The anodising process • Electrolytic process used to thicken the natural oxide film on aluminium and its alloys (range 8-25 µm). • Modern industrial processes based on sulphuric acid, chromic acid or phosphoric acid as electrolyte. • Improves corrosion resistance and possible surface treatment prior to painting or adhesive bonding. • Environmental concerns with chromic acid anodising replacement processes currently being sought. • Main candidate is the boric acid-sulphuric acid process.

Film thickness ~ 2 µm

SEM micro-graph showing a compact boric-sulphuric acid anodised film.

Anodising and fatigue • Anodised films are hard and brittle (alumina) and readily crack under load. • This facilitates the initiation of embryonic fatigue cracks in the surface oxide film. • Cyclic loading allows these sub-critical cracks to grow into the substrate material. • Catastrophic failure eventually occurs so significantly reducing fatigue strength (life).

Reduction in fatigue strength Max Stress (MPa )

200

control anodised

150 100 50 0

10

4

10

5

10

6

10

7

10

8

Number of cycles

S-N data for boric acid-sulphuric acid anodised 2024:T4 aluminium alloy

Micrograph showing oxide film cracks initiating a fatigue crack in substrate.

Current understanding • Anodising facilitates the initiation of fatigue cracks. • 20→ 30 % reduction in fatigue life likely due to this initiation mechanism. • All anodised components subject to fatigue loads will suffer from this problem - extent depends on process. So • Benefits gained in terms of corrosion protection must be balanced against the disadvantage of reduced fatigue strength.

Crack growth in anodised aluminium • Although initiation effects are quite well understood, the influence of anodising on fatigue crack growth rates (FCGR’s) has not been previously studied. • This presentation reports the findings of a study which considered the the influence of boric acid sulphuric acid anodised films on FCGR’s in 2000 series aluminium alloys.

Experimental details • FCG data obtained from SEN specimens made from 2000 series (T4 & T6) aluminium alloys. • Data obtained (anodised and control) under identical constant load amplitude conditions (ASTM std E647). • Specimen thickness (1.3 mm) corresponding to plane stress conditions. • ∆K range during test 5 MPa√m → 30 MPa√m at load ratio R = 0.1 • Crack measurement (travelling microscope) only after notch stress field surpassed.

70 mm 140 mm

Specimen geometry and testing arrangement.

Results: crack growth curves (2014:T6)

Crack length (m)

0.04

anodised

0.03

1

2

3

control 12

3

0.02

0.01

0.00 0

200000

400000

Number of cycles

600000

Fatigue crack growth rates FCGR’s can be represented in terms of LEFM using the Paris - Erdogan relationship: da/dN = C(∆K)m where ∆K→ stress intensity factor range ∆K = Kmax - Kmin= ∆σ √πa α C , m → intercept and slope parameters (material). α → correction factor (crack length and geometry). a → crack length

Crack growth rates (2014:T6)

Log10 da/dN (m/cycle)

-6 control anodised

-7 ~ 30% increase

-8

10 MPa√m -9 6.6

6.8

7.0

7.2

Log10 ∆K (MPa√m)

7.4

7.6

Effect on crack growth rates • For the mid range of the Paris regime results indicate a 25-35% increase in FCGR. • Similar results obtained for 2024 (T4) alloy also tested at R = 0.1 . • Tests carried at R = 0.2 and with thicker material showed no increase.

Crack length (m)

Results: crack growth curves (2024:T4) 0.020

0.015

0.010

control anodised

0.005

0.000 0

100000

200000

Number of cycles

300000

Crack growth rates (2024:T4) Log10 da/dN (m/cycle)

-6 control anodised

-7 ~ 80 % increase

-8 6.8

10 MPa√m 6.9

7.0

7.1

7.2

7.3

Log10 ∆K (MPa√m)

7.4

7.5

Comment • Results demonstrate that FCGR’s in these alloys can be markedly affected by the presence of a thin anodised film. • Increases of up to 80% were possible for alloys tested at R = 0.1. • 2024:T4 appeared more susceptible than 2014:T6. • Seems to be a plane stress effect. Two possible causal mechanisms are suggested

Mechanisms for enhanced FCGR 1.

Constraint of substrate related to residual stress in oxide film reduces the plasticity-induced closure experienced by the growing fatigue crack.

2.

Highly cracked surface morphology of the oxide film alters the crack growth characteristics due to the degradation (embrittlement) of the material ahead of the crack tip.

Plasticity-induced closure increasing ∆K developing plastic wake

notch

closure load (Pcl) plastic zone at crack tip growing fatigue crack

Modified Paris-Erdogan relationship In the presence of plasticity-induced closure the crack growth rate equation becomes: da/dN = C(∆Keff)m where ∆Keff = Kmax - Kcl Kcl → stress intensity with plasticity- induced closure.

∆Keff - plasticity-induced closure

Stress Intensity

Kmax ∆Keff

∆K

Kcl Kmin

Time

Measurement of closure load (Pcl) Pcl was determined for both control and anodised conditions using a side-face strain gauge technique. strain gauges

5 mm notch

9 mm

8 mm

notch tip strain field

Plasticity-induced closure loads

Closure load (N)

800

600 Pcl = 530 N Pcl∗ = 450 N 400

200 control anodised 0 0

10

20

Crack length (mm)

30

Closure results • Results indicated a reduction in the closure load experienced by the anodised material. • Reduction due to extra constraint imposed by oxide film on developing plastic zone. • Constraint related to residual tensile stress in film (measured value ∼ 400 MPa).

Crack front curvature (additional evidence for constraint)

• Reduced curvature of crack front for anodised material reflecting enhanced growth rate in surface region. • Implies a localised increase in ∆K in that area consistent with a reduced closure mechanism.

Anodised & control crack growth rates Accounting for closure in both material conditions the Paris-Erdogan crack growth rate equations become: (da/dN)control = C(∆Keff)m (da/dN)anodised = C*(∆Keff*)m* where ∆Keff = Kmax - Kcl

(control)

∆Keff* = Kmax - Kcl*

(anodised)

C*, m* → slope and intercept parameters (anodised)

∆Keff* - reduced plasticity-induced closure Stress Intensity

Kmax ∆Keff

∆K

∆Keff*

Kcl Kcl* Kmin

Time

Growth rates corrected for closure (Paris regime data) Log10 da/dN (m/cycle)

-6.50

-6.95

-7.35 control anodised -7.80 6.80

6.95

7.10

7.25

Log10 ∆Keff (MPa√m)

7.40

Comment on ∆Keff* • The mechanism of reduced plasticity-induced closure does narrow the disparity between the growth rates for the two conditions but not entirely. • An additional mechanism must also be contributing to the higher growth rates observed in the anodised material. • Optical microscopy showed presence of many small cracks in the oxide film adjacent to the crack tip these may also be a contributory factor.

Cracking in anodised oxide film

Focus on film cracks

0.1 mm

Focus on substrate crack

The role of oxide film cracks • Main substrate crack interacts with film cracks temporarily altering its direction and growth rate. • (∆Keff*) now a combination of normal mode I loading and an acquired a mode II component from film. • Stress intensity intermittently increases locally above normal ∆Keff* value. • Therefore (da/dN)anodised increased further.

Mixed mode I and mode II loading

film cracks

main substrate crack notch

KII

θ

KI crack slant angle

Britannia Royal Naval College

Reduced closure and film crack effects Combining these two mechanisms now fixes the effective stress intensity, (∆Keff*)+, as (∆Keff*)+ = (Kmax - Kcl*) + Kfc

(anodised)

where Kfc → temporary increase in stress intensity caused by the interaction of the main substrate crack with a film crack.

Calculating Kfc Problem How to determine Kfc since intermittent combination of mode I and mode II components ? And how long is the additional mode II component switched on for ? Solution Maximum principal stress criterion allows calculation of an equivalent mode I solution (KEqfc) for this type of mixed mode loading situation. Partition this according to the coincidence time of film and substrate cracks.

Maximum Principal Stress Criterion So in this instance KEqfc = kicos3(β/2) - 3kiicos2(β/2) where ki = Yo σ √πc sin2θ kii = Yo σ √πc sinθ cosθ β = θ/2 c and θ obtained by direct measurement and Yo by treating a film crack as a semi-elliptical surface flaw.

Growth rates corrected for closure and film crack effects Log10 da/dN (m/cycle)

-6.50

R2 = 0.99 -6.95

-7.35 control anodised -7.80 6.80

6.95

7.10

7.25

Log10 (∆Keff*)+ (MPa√m)

7.40

Conclusions • Fatigue crack growth in anodised aluminium governed by both initiation and enhanced growth rates effects. • Two concurrent growth rate mechanisms operate; i) constraint of substrate reduces plasticity - induced closure (due to tensile residual stress in film), ii) presence of film cracks supplying an additional, but intermittent, mode II loading component. • Combined influence of these mechanisms can be modelled using normal LEFM methods. • Film-assisted fatigue crack growth may also occur in other surface coated materials.

Questions ?

Improved Corrosion Resistance and Durability with Single Component Moisture Cure Urea MC TECHNOLOGY BELGIUM BV

SINGLE COMPONENT MOISTURE CURE URETHANE

This is why Moisture Cure was developed

30 30 Years Years Ago Ago •

•

Offshore Maintenance calls for coating with application capabilities in cold / hot highhumidity conditions. The most viable solution was the single component moisture cure urethane or more appropriately polyurea.


What is a Moisture-Cure Urethane? Single-Component Liquid Coating +

Polyurea HUMIDITY

Dense, pore-free, chemical resistant coating

=

Initial Initial MCU MCU Problems Problems • Stability • Intercoat adhesion • High film build resulted in failures • This drove away most major coating mfg.’s listing MCU as another good idea that just does not work. • However a small group of smaller firms produced this technology, still in use today.

Wasser Wasser High-Tech High-Tech Coatings Coatings • In 1980 Bill Brinton, established Wasser High-Tech Coatings, where he was able to develop unique formulations based on a development of proprietary resins and additives. And offer superior; • Stability in the can over long periods of time. • Long term substrate and intercoat adhesion • Extended corrosion protection • Flexibility, will not become brittle over time • UV protection & chemical resistance

Use Use of of Micaceous Micaceous Iron Iron Oxide Oxide • MIO is a laminar pigment additive • Allows for micro-permiability – reducing gas & moisture entrapment • Yields better UV and atmospheric chemical resistance • Better abrasive & erosive qualities than aluminium & glass flake • Used in most all Wasser coatings and systems.


The The innovative innovative technology technology -- polyureas polyureas Ê Ê Ê Ê Ê Ê Ê Ê Ê Ê Ê Ê

No dew point limitation No humidity limitation (6% to 99%) No temperature limitation (down to -12oC) No pot life limitation, single component No induction time Very surface tolerance (surface preparation) No peeling on old coatings when used as an overcoat Long term flexibility Cure fast even under water, application in fog conditions No gloss loss with rapid exposure to condensate No mixing error in ratio or types VOC compliant

Eliminate hidden costs

Work Shut downs / Down time Delay penalties Idle manpower Premature Failure Consumables waste Weather limits Two-pack repair Hidden agenda’s

TIME TIME IS IS MONEY MONEY Savings: • No weather related delays • Ability to paint sweating gas & water pipes without line shut down • Faster turn around • Less down time • Fastest cure offered - 3 coats in as fast as 1 hour. • Immersion service in less than 15-30 minutes. • Longest lasting corrosion protection

Wasser benefits ship owners to save time and money by being more successful completing projects at sea! Savings of riding crew over dry dock • Loss of daily vessels earnings • Cost of dry dock • Yard % of contract labour • Yard % of materials acquired


Ê Ê Ê Ê

ISO 12944 C5M and NORSOK QUV ASTM G53: Accelerated Weathering - Pass 500hr. ASTM D4548 : Water Resistance - 10,000 hr no change ASTM B117 : Corrosion Resistance Salt Fog - Pass 20,000 hr.

Ê Ê Ê

ASTM D522 : Flexibility Mandrel Bend - Pass 0.5 inch ASTM D2794: Impact Resistance - Pass 150 inch LBS ASTM D4060 CS17: Abrasion Resistance, wheel, 1000g

Load, 1000 cycles: less than 35mg loss. Ê ASTM D4541: Adhesion, more than 500 psi

TYPICAL EPOXY MASTIC SYSTEM

Combined Cycle Test 30 cycles / Corrosion Salt Spray 1,000 hours EPOXY MASTIC 350 DFT

MC Miozinc / MC Ferrox B / MC Luster 75 + 75 + 75 Total DFT = 225 microns

Combined Cycle Test 30 cycles / Corrosion Salt Spray 1,000 hours

MC Miozinc, MC Ferroguard X 2 75 + 150 + 150 Total DFT 375 microns

Combined Cycle Test 30 cycles / Corrosion Salt Spray 1,000 hours

Examples Examples of of Failure Failure Analysis Analysis • This demonstrates an improper coating chosen for a tank internal that was not able to flex at the same coefficient as the weld seam.

Example Example of of Failure Failure Analysis Analysis • This is also an example of the wrong coating chosen for an internal tank, which became brittle and cracked during hull flexing. • The corrosion on the stringer show improper surface preparation and the use of a coating not surface tolerant.

Example Example of of long-term long-term edge edge retention retention and and excellent excellent performance performance even even coated coated on on poor poor steel. steel. This ballast-tank holed when UHP prepared at sea, had the Wasser MC Miozinc and MC BallastCoat system applied. The steel was not replaced, however after over 3 years even on the stress of coating a knife edge no corrosion was visible.

OUR STRENGTH Gain 10% min. of coverage per person. Ê Gain 15% min. of time saving to clean spray equipment. Ê No environment condition restrictions, shorten project time by 30%. Ê Easy maintenance especially on splash zone or wet condition. Ê Out last epoxy base systems, solves epoxy mastic problems. Ê

European European Marine Marine Projects; Projects; • • • • • •

John Pederson & Sons – Bulk Vessels, Cargo holds Stena –Fast Catamaran – Aluminium corrosion solutions World Wide // Bergesen DY – Ballast tanks Allseas – Pipe laying vessel externals and sea bed crib Hoegh – Ballast tanks RoRo decks; Wagenborg, Swedish Orient, Transatlantic, Goliat (Torline). • Stolt Neilsen – Chemical Vessels • Ensco Offshore – Platforms & Rigs, externals, internals, tanks • Others such as; Herema, Boskalis,


SURFACE SURFACE PREPARATION PREPARATION Previously blasted Steel. Hydro-blast or Hydro-Jet Ê Poorly Prepared Surfaces. ST 2 & ST 3 Ê Overcoating Existing Coatings on Steel, by removing, grease & oils, loose paint and loose rust Ê High gloss over-coating, must be naturally weathered or slightly abraded; mechanically or chemically. Ê


BLASTED BLASTEDSTEEL: STEEL:POORLY POORLYPREPARED PREPAREDSURFACES SURFACES //UHP UHP&& OVERCOATING OVERCOATINGEXISTING EXISTINGCOATINGS COATINGSTYPICAL TYPICALSYSTEMS SYSTEMS

Ê Ê Ê Ê Ê Ê Ê Ê

Primer: MC-Miozinc / or / MC Prepbond Followed by; 1/ MC Prepbond 2/ MC-Ferroguard / MC- Ferroguard 3/ MC -Miomastic / MC –Luster or Ferrox A 4/ MC-CR / MC –Luster or Ferrox A 5/ MC- Luster or Ferrox A 6/ MC BallastCoat / MC BallastCoat

Typical Application: ONE GRADE

• Brush. • FOR ALL APPLICATIONS • Roller. • ONE THINNER FOR ALL • Pad. COATINGS • Air Spray (Conventional). • ONE COMPONENT NO MIXING • Airless Spray.

Applicable Applicable Substrates; Substrates; Ê Ê Ê Ê Ê Ê Ê Ê

Steel, mild & galvanized & metalized Iron Stainless Steel Aluminium Copper Brass Plastics Fibre Glass


AREAS OF APPLICATION

Ê

Tanks: Ballast, Drinking Water, Fuel, Grey & Black water, mud Sweating Exterior Piping Hull, Deck Superstructure Offshore platforms, exterior – interior – all tanks – incl. mud tanks Port & Onshore Facilities: Cranes, Dry-docks, Sheet piles, Pilings, Piers All field repainting in damp conditions: And Virtually any areas need corrosion control & protection.

Ê

Not Suitable for Chemical Tanker Linings!

Ê Ê Ê Ê Ê Ê


Full Range Wasser MCU-Coatings Primers: Intermediates:

Topcoats:

Special Coating:

MC-Miozinc MC-Prepbond

World Class Quality

MC-Ferrox B MC-Miomastic MC CRPW / MC BallastCoat

MC-Ferrox A MC-Luster MC-Aroshield MC-Clear MC-Ferroguard MC-Aluminium MC-High Heat Aluminium - on order MC-High Heat Zinc - on order

MC MC MIOZINC MIOZINC • • • • • • • •

Zinc, surface tolerant, ST 2 – SA 2.5, WJ 2, primer No maximum recoat time, infinitely re coat-able Will adhere to existing coatings Recommended DFT 75, 200 acceptable Recommended for immersion Compatible with zinc anodes Compatible with other coatings Recommended for steel, iron, failing galvanizing & metalizing

MC MC PREPBOND PREPBOND • • • • •

Surface tolerant aluminium, penetrating primer/sealer Highly abrasion resistance Designed initially poor surface preparation All metal and concrete surfaces Will penetrate loose rust, recommend to remove scale & apply mechanically • Compatible with other coatings • Overcoat within 3 – 5 days

MC MC CRPW/BallastCoat CRPW/BallastCoat • Light coloured (off-white & beige) coating • For Ballast, drinking water, black & grey water and fuel tanks, cargo holds, offshore mud pits, • ANSI/NSF approved potable water • Intermediate coat for superstructure • Infinitely re coat-able • Compatible with other coatings • Over 15 years on vessels, & offshore platforms

MC MC Luster Luster & & MC MC Ferrox Ferrox A A • Aliphatic, uv and atmospheric chemical resistance. • Excellent resistance to acid and oil and diesel spills • Does not amine blush with exposure to moisture or condensate prior to curing • Infinitely re coat-able • Can receive foot traffic from 4 hrs. to as low as 30 minutes after application in 10 C + • Compatible with other coatings

Extended Extended Performance Performance Job Job Profiles Profiles • Coating projects with Wasser has shown longer life durability and possible life extension with significantly reduced maintenance costs. • Projects completed and documented in ballast tanks, offshore platforms, RoRo decks

Shell References - Gulf Of Mexico Production Platform (Eugene Island 189) Comparison between 3 projects at same time, 2 with epoxy and 1 with Wasser MCU. The MCU job summary; Application completed two weeks ahead of schedule •Saving in man-hour •No hazardous waste •No down time due to weather conditions •No loss of product due to mixing of two component ( Approximate total saving of nearly US$3.50 per square foot )

* Test Evaluation : Shell Development, Houston Texas

• Why remove good coating, unless necessary? Wasser adheres to most without abrading.


Ê Excellent

adhesion over all other

Coatings. Ê Performs well with anodic protection. Ê A true corrosion coating system, that does not need additional external protection system such as CP. Ê Pass corrosion resistance test, ASTM B117 Salt Fog: 20,000 hours.

ENSCO ENSCO INTERNATIONAL INTERNATIONAL DRILL DRILL PLATFORMS PLATFORMS

• MAINTENANCE OF ALL PLATFORMS IN N. SEA • HYDRO BLASTING • COATINGS SUPPLIED BY WASSER

Hual Hual Trubador, Trubador, Car Car Carrier Carrier

Hual Hual Trubador, Trubador, Ballast Ballast Tank, Tank, Hoegh Hoegh Fleet Fleet Services Services AS AS • Completed by riding crew in 2000, water jetting surface prep. • After 4 years, the coating is in 100%, with no coating breakdown • No signs of corrosion

Stena Stena MV MV Discovery Discovery

This fast going aluminium catamaran ferry developed pittings some severe, in the untreated areas, specifically in the turbine intake rooms. • Epoxy coatings were not able to penetrate into the pittings and offered unsuitable adhesion. • See the story highlighted in the March Issue of RINA's Ship Repair and Conversion Technology magazine.

MV MV Discovery Discovery –– Test Test Application Application

•

• •

A test was carried out on a hatch plate form the air intake room. The plate was prepared 50% with a chloride remover, then entirely blasted with 3000 bar. MC Prepbond was applied 50% as a 1-coat and 50% as a 2coat. After 2 months in service evaluation tests were completed, by Stena

MV MV Discovery Discovery –– upon upon completion completion and and after after 12 12 months months

Superior Superior Abrasive Abrasive resistance resistance • Commercial RoRo vessels have typically had coating adhesion problems on decks in abrasive situations, often having the coatings peeling away within 3 months. • Wasser has been used as the repair and recoat on Canadian Navy vessels as well as on on several Goliat Torline vessels and is now being used to repair and recoat decks on Wagenborg vessels, Transatlantic and Swedish Orient Lines. • With a three coat system at 30% of the price of glass flake epoxy, 25% of the ceramic epoxy and 50% of HS epoxy the Wasser system is performing over 3 times longer – TO DATE.

Wagonborg Wagonborg Shipping Shipping Test Test 2004 2004 • Due to premature failures within 3 – 6 months of full blast and recoat and within 12 – 18 months of new build, a comparative test of 4 systems was conducted. • Systems tested on abrasive blasted SA 2.5, included: • Ceramic pigmented solvent free epoxy • Glass flake pigmented solvent free epoxy • Anti-abrasive formulated solvent free epoxy • Solvent was added to the epoxies in order to get better adhesion • Wasser single pack MCU, MC Miozinc & MC Prepbond • All systems had Aluminium Oxide for non-skid due to heavy loads required.

Wagonborg Wagonborg Test Test –– system system 11 • Ceramic filled epoxy • The coated surface has been hit several times with a sharp and heavy steel flat bar. • The adhesion is not very good, the coating breaks easily after a few hits. Most likely because of the added ceramic parts. • Undercutting and corrosion seen.

Wagonborg Wagonborg Test Test –– system system 22 • Glass flake filled epoxy • The coated surface has been hit several times with a sharp and heavy steel flat bar. • Undercutting and initial corrosion seen. • The adhesion is not very good, the coating breaks after a few hits.

Wagonborg Wagonborg Test Test –– system system 22

• The coated surface has been hit several times with a sharp and heavy steel flat bar. • The adhesion of the primer coat very good but the system breaks after hitting with the steel bar in the primer coat.

Wagonborg Wagonborg Test Test –– system system 44 • The coated surface received several severe impacts with a sharp and heavy steel flat bar. • The adhesion is very good, the topcoat was removed. • The MC Miozinc stayed intact.

Wagonborg Wagonborg Material Material Test Test Conclusion Conclusion

• Considering all facts; lower cost per square meter, strength against mechanical impact, incredible corrosion protection, required pre-treatment, hardly any restrictions and large window for application, fast curing and easy onboard maintenance (single component) I have to conclude that the best choice would be the Wasser system. • Jan Wink (NACE 3), Wink inspections BV

Wagenborg Wagenborg –– Spaarneborg Spaarneborg rehabilitation rehabilitation

• In 2005 12 RoRo vessel decks were rehabilitated with Wasser’s MCU system. One vessel had a 60+ ton container fall gauging the deck system topcoat – but without going down to the bare steel. Owners are now specifying Wasser on new builds, 3 scheduled for delivery in 2006.

Norwegian Norwegian Coast Coast Guard Guard • These structures are often accessible by helicopter and time is of the essence to prep. and coat. • Cold and damp conditions have lead to poor maintenance. • Wasser has been met with outstanding results.

Norwegian Norwegian Coast Coast Guard Guard • 4 years contract since 1999, Wasser coatings have been used for the more than 10,000 navigational light houses

2006 NEW DEVELOPMENT MC TECHNOLOGY BELGIUM WW EXCLUSIVE HIGH SOLIDS MOISTURE CURE • UP TO 85% VOLUME SOLIDS • VOC AS LOW AS 165 • FILM BUILD CAPABILITIES OF 75 – 600 DFT, WITHOUT GAS-ENTRAPMENT OR BUBBLING • REDUCE THREE COAT TANK APPLICATIONS TO TWO COATS • SAME VISCOSITY – GOOD WETTING OUT PROPERTIES EVEN IN AREAS WITH PITTINGS.

HIGH-TECH COATINGS

HOW MUCH TIME AND MONEY COULD YOU SAVE IF YOU COULD PAINT IN DAMP CONDITIONS

•ABRASIVE BLASTING NOT REQUIRED - HP or UHP, hand and power tool acceptable in many situations. •UNLIMITED RECOAT WINDOW ON MOST PRODUCTS •GREAT ALLOWANCES For Difficult Field Conditions •PASSES OVER 20,000 HOURS ASTM B 117 SALT FOG TESTING •EVERY TEST AGAINST EPOXY HAS PROVEN MCU IS FAR BETTER

WASSER IS MADE FOR DAMP CONDITIONS www.wassereurope.com

TEAM WASSER - TRUSTWORTHY – RELIABLE Thank you for your attention

Our team will be pleased to discuss and present our coatings with you and review any potential projects for the Wasser system and will quote on projects on an job specific basis.

SINGLE COMPONANT MOISTURE CURE UREA

Royal Institution of Naval Architects Advanced Marine Materials & Coatings London Feb 22nd – 23rd 2006

Managing Coatings Through Life

M.R Kattan Ph.D, C.Eng, FRINA & R.H.Towers C.Eng, MRINA

Coating Management - Two Major Phases Newbuilding Design, selection of & coating installation 2-3 yrs for most commercial ships 3-5 yrs + for warships Operational life Maintenance of coating systems 20-25 yrs commercial ships 25-30 yrs warships Doc No. XXXCYYYpres

2

The Long Phase

M&R

Annual operational cost

seastock paints s/structure, dks, E/R, internals

Drydocking paint cost 24, 36, 60 month intervals

A/F+ (BT, TS)

Major system upgrades infrequent /10-20 yrs

hull UW cargo tanks cargo holds WBT, other

Doc No. XXXCYYYpres

3

Planned Maintenance Today a key function of ship management Some form mandatory for all ships under class Owners use - own systems - proprietary software Provide records and info feedback for - ship management - Class - Flag / Port State control auth. Doc No. XXXCYYYpres

4

Planned Maintenance Main elements cover - main propulsion system - navigational machinery and equipment - electrical power generation - hull structure

Doc No. XXXCYYYpres

5

Maintenance Management Remote MMS Database

Central MMS Database

Routine Completions

Current Workload & Defects Automatic Workload Reschedule

PMS-1 Completions

Maintenance Planning Convert Defect to Work Order

Schedule Work Order

PMS-2 Forward Workload Check Resource Availability Labour & Spares

Spares a s Demands

Record Defects

Update History

Meter Readings

Replicate Bi-Directional Replication epl c n

Commit PMS-2 Work Spares Sp r Demands

Complete Work Order

Complete PMS-2 Work Enter Completions

Base Maintainers & Workshops

Basic Process Cycle for Planned Maintenance

Planned Maintenance What about coating systems ? Some owners approach coating maintenance through ‘Partnering Agreements’ with paint suppliers Planned Maintenance systems can however include coatings and some owners do this but the general case seems to be that coating systems are not subjected to proper planned maintenance Doc No. XXXCYYYpres

7

Reasons for this A combination between Owners perception about - the cost of paint in the initial NB investment package - the role which paint plays in the operational life of the ship Shipbuilders control over the initial specification Avoidance of another additional cost ? Doc No. XXXCYYYpres

8

Paint in the NB Investment Package The Shorter Phase VLCC 300K NB in Korea current price 2/06 $ 118-125 million steel weight ± 35,000 tons steel cost / Korea $ 700 per ton paint cost / full ship currently equates to

± $ 2 million 1.6 – 1.7% of ship price

Doc No. XXXCYYYpres

9

Paint in the NB Investment Package 300K VLCC ± 35,000 tons steel Full ship Paint application & surface preparation

Man hours Korean yard

Man hours per ton steel

± 500,000

≈ 14

35,000-40,000

≈ 1

Paint application costs in relation to man hours per ton steel in relation to total man hours Doc No. XXXCYYYpres

small 7- 8% of total 10

Paint in the NB Investment Package VLCC 300K NB in Korea current price steel weight paint cost / full ship paint application man hrs

$ 118-125 million ± 35,000 tons 1.5 - 2% of ship price 7- 8% of ship man hrs

paint package cost will be

± 9-10% of ship cost

paint package cost larger than generally appreciated Doc No. XXXCYYYpres

11

Paint in the NB Investment Package The Shipbuilder will only seek the Owners approval of the paint supply, 1.5 - 2% of cost The Purchaser will consequently have limited scope to influence the Shipbuilder in this specification item The decision has become one - of lower priority driven by contract cost - without long term preservation considerations - unlikely to have early effect on operational costs The longer term consequences arising from the package can however be very costly as maintenance items Doc No. XXXCYYYpres

12

Paint in the NB Investment Package In ships with fully coated cargo tanks eg; chemical tankers, & VLPCs the total paint package will represent 15% + of total cost The decision here is - driven by Owners operational requirement - the need for coating reliability and performance - about impact on the overall investment - fundamentally more important

Doc No. XXXCYYYpres suggests paint package cost will be

± 9-10% of ship cost

13

Paint in the NB Investment Package This differential in package cost is reflected in the derivation of paint specifications Functional specifications for cargo tanks and holds Performance specifications for antifoulings (+WBT in future) Generic specifications for most other locations There is no common approach Doc No. XXXCYYYpres

14

Role of Coatings Long term preservation against corrosion Prevention of fouling of hull Cosmetic appearance of hull Protection from corrosive effects of cargoes liquid chemicals, petroleum products and dry bulk

Protection from corrosive effects of having to carry non earning liquids SW ballast, FW, sewage, grey water, etc Doc No. XXXCYYYpres

15

Coating Performance Through Life Correct performance is dependent upon Initial coating selection Properly specifying performance requirement Correct installation of coating system Subsequent proper ‘on board’ maintenance Reinstatement or upgrade of full system later in life

Doc No. XXXCYYYpres

16

Factors Influencing Coating Selection at NB Shipbuilder may have own preferred supplier Owner may have own fleet suppliers – different Owners paint suppliers may offer commercial incentives Contract terms may limit yard supplier changes Key factors often overlooked Will selected coatings perform for duration of NB schedule ? Will selected coatings be ready to accept owners first cargo ?

Doc No. XXXCYYYpres

17

Suggestions to Improve the Coating Selection Process NB contract discussions could usefully include Location, duration and predominant season of build Environmental exposure during intended trading What planned maintenance system for coatings ? Methods for assessing ‘on board’ condition Resources available ‘on board’ for painting maintce Doc No. XXXCYYYpres

18

Generic Specifications In widespread use by Owners and Shipbuilders NB & Maintenance Specifications Location

deck

Generic Description

Paint Type

No of Coats

DFT mic

Colour

epoxy

a/c

1

80 mic

grey

epoxy

finish

1

50 mic

green

‘The traditional approach’ Doc No. XXXCYYYpres

19

Functional Specifications Owners and Shipbuilders have differing functional requirements of coating systems Location

Function

Requirement

for

anti-corrosion

3 yr life / 1% spot rust

owner

s/structures

gloss

Initial gloss retention & after 24 months

owner

cargo hold

Taber resistance

Taber value

owner

main deck

Max overcoating time

6 months

builder

Drying time to walk at 10ºC

12 hours

builder

general

‘The advanced materials approach’ Doc No. XXXCYYYpres

20

Correct Installation of Coating Systems It is normal for newly applied coating systems to look fresh and bright The underlying assumption is that the work must have been done well but coats of fresh paint can easily render an underlying problem invisible Doc No. XXXCYYYpres

21

Analysis of Coating Failures over 8 Years percentage 45 40 35 30 25 20 15 10 5 0 Design

Spec

Applic

Chem

Doc No. XXXCYYYpres

Operation

Other

22

Causes of Failure Application

- most frequent contributor

Operational pressure - cargo commitment - off hire time - local weather, tides etc Selection & Specn

- more frequent than may be realised

Doc No. XXXCYYYpres

23

Supervision of Installation Considerable time & cost spent on approving the installation both during NB and DD In NB, Builder accepts inspection of paint supplier but can exert pressure on inspectors re schedules, and as paymaster. Owners get limited records. For DD, Owners rely heavily on s/vision by paint supplier as their ‘in service‘ partner. Owners get full DD painting report For ‘on board’ maintenance painting, Owners usually list actions completed rather than compile useful records Finding Investigations have shown that Owners paint records are usually somewhat inconsistent and poor in their detail Doc No. XXXCYYYpres

24

Painting Maintenance ‘on board’ Trends of last 30 years - ship size has increased - crew size has reduced - coating materials have changed alkyds chlor rubber epoxy

Consequences

- less ‘onboard’ painting done. - product knowledge is weak

Crew have little knowledge of paint or application Work is poorly planned and often ‘ad hoc’ The low layer thickness of a brush or roller application is generally not understood. Doc No. XXXCYYYpres

25

Painting Maintenance ‘on board’ Maintenance painting by crew deck

epoxy

a/c

1

80 mic

grey

epoxy

finish

1

50 mic green

Typical brush / roller application will achieve 30 mic per coat. In practice a/c will require 3 coats to fully reinstate finish will require 2 coats to fully reinstate Prognosis Result

crew will not apply 5 coats due other priorities corrosion likely to occur again at same location Doc No. XXXCYYYpres

26

Painting Maintenance ‘on board’ What might improve maintenance painting by crew ? Preparation of appropriate work packages with timings and specification details Schedules of work by locations with agreed standards Standardise the system of reporting coating work carried out Planned Maintenance systems are able to provide this type of information

even for paint

Doc No. XXXCYYYpres

27

Shipbuilders Warranty incompatible with Owners long term asset preservation need In the Shipbuilding industry - 12 months has long been the shipbuilding global commercial standard for all supply items but - up to 5 years warranty on AF (by paint makers) - up to 4 years warranty on tank linings (chm tkrs) In the Car industry - up to 7 years warranty on the body paint system is given by the Car manufacturer Doc No. XXXCYYYpres

28

Long term preservation v. low current cost The longer term view poses a financial conflict and choice will depend upon owners longer term intention on ownership The short term traditional view is driven by low cost - when freights are good, there is money but no time to do the work - when freights are low there is never enough money Doc No. XXXCYYYpres

29

Reinstatement of Coating Systems Examples might be the failure of coating system early in the ships life ( NB problem ) or major re-coating of hull, cargo holds,etc The costs of this work can be dramatically different from yard NB costs and can reach up to $50 m2 This only emphasises the importance - of ensuring the best installation at NB - of managing the maintenance of coating systems through life Doc No. XXXCYYYpres

30

A Non Chromate Conversion Coating Process for Corrosion Protection of Al 2024 T3 Aluminum Alloys in a Marine Environment Wayne C. Tucker and M.G Medeiros Naval Undersea Warfare Center, Newport, Rhode Island Richard Brown and Dharma Maddala Department of Chemical Engineering The University of Rhode Island Kingston, Rhode Island

Example of Chromating

Aluminum Heat Sink

Objectives • • •

Replace chromates employed as corrosion resistant conversion coating. Minimum to no disturbance of present process. Corrosion resistance to meet specification for chromate.

Overall Process. • • • • • • • • • •

Solvent cleaning with acetone Mechanical cleaning with abrasive such as Scotchbrite Solvent cleaning with acetone Rinse in de-ionized water Chemical cleaning with sodium hydroxide at ph 12.5 for 10minutes at 40C Rinse in de-ionized water Deoxidize in proprietary solution of Smut-Go Rinse in de-ionized water Conversion coating in titanate solution at 62C for 3 minutes Rinse in de-ionized water

Why Titanium Based Solution Multiple Valence Passive Film Low Passivation Potential Ionic Form

Chromium 6+or 3+ Yes Yes

Titanium 4+ or 2+ Yes Yes

Yes

Yes

Chromate Conversion Coating • The current chromate coatings (Alodine type) are composed of: • 0.025M Na2Cr2O7 • 0.024 M NaF • 0.015M K3Fe(CN)6 • pH range 1.2 to 2.2 adjusted with HNO3

Titanate Conversion Coating • The titanate coatings are composed of the following: • (0-1M) Na2Ti3O7 • (0-1M) NaF • (0-1M) K3Fe(CN)6 – usually zero • pH range 1.0 to 5.5 adjusted with nitric acid

Formulation Used The titanium bath formulation for all samples consisted of:6g/L of K2TiO3, 4g/L of NaF and pH adjustment with nitric acid to either 2 or 5.5 at 62 C. No cyanide additions were used

Process Conditions • Non uniform coating without an abrasive treatment. • 10 minutes in NaOH for uniform surface. • 10 minutes in SmutGo for uniform surface • 3 minutes in conversion bath – longer leads to a powdery deposit on surface.

Abrasive and Solvent Cleaning

As received Al 2024 T3 Panel

Post abrasion with Scotchbrite and acetone wipe Remove any visual surface marks such as small scratches and debris along with oils and other surface contaminants

Chemical Cleaning Rinse tank DI water

1 minute in NaOH

NaOH SmutGo tank pH meter tank 10 minutes in NaOH

Aim is to produce a uniform film on the surface, using NaOH at 40 C and pH 12.5. A 10 minute immersion was found to produce the necessary uniform surface. A rinse in de-ionized water for 1minute followed.

Chemical Cleaning

SmutGo 3 minute

SmutGo 10 minutes

After NaOH and rinse, 10 minutes in SmutGo a proprietary de-oxidizer solution at room temperature was needed to clean the surface.

Conversion Coating

Conversion Coating

Finished panel after 3 minutes in conversion coating solution at 62 C, wash in de-ionized water, then air dried.

Electrochemical Data ELECTROCHEMICAL TESTING APPARATUS.

Computer controller

URI Coating

Potentiostat

Reference electrode (SCE)

Counter electrode Platinum grid

2

∆A ∆V

Chromate Coating

2.5

1.5 Electrolyte level

Esce(V

Working electrode

Base Sample

1

0.5

Glass vessel Test Sample

0 -0.5 -1 -1.5 -10

O ring seal

-5

0

2

5

10

Log i (uA/cm ) Capilliary tube to connect SCE to solution near working electrode

Test apparatus

Data comparing uncoated to chromate to titanate coatings on Al2024 T3.

Impedance Data Chromate coating

Titanate coating

After 27 days of exposure to 0.5N NaCl, the titanate coating was at A higher impedance than the chromate coating on Al 2024 T3.

Salt Spray Testing

Small panels 2.5 by10 cm

Samples after 30 days of salt spray

Salt Spray

Three days of exposure some panels have already failed

16 days of exposure – panels with no corrosion meet MIL-DTL-81760B

Issues • Same treatment – some panels meet 14 day specification of MIL-DTL-81760B, others fail rapidly. • Panels that meet 14 days fail after 16 days

Conclusions • Titanate based conversion coatings hold promise as a replacement for chromates on Al2024 T3 alloys • Further investigation will be required to turn the titanate conversion process into an industrially accepted process

Acknowledgements • The financial support of NUWC from their ILIR program is gratefully acknowledged.

Setting N ew Standards

Corrosion Inhibiting Therm oplastic P olym er P rovides both passive and active protection to the entire system

Bolted system constructed from dissim ilar m etals

Enviropeelcoating prevents ingress of m oisture

Built-in inhibitors penetrate allsurfaces preventing any corrosion

Although caps have been used to protectthe bolts -they are notprotecting anyw here else. Caulking has been used to try and protectthe flange faces – butithas shrunk back,creating the perfectenvironm entfor m oisture to getin and corrosion to accelerate! Itis notpossible to fitcaps – or any other form of protection here – the caps are too close together The allen bolts atthe base have no chance. Stream ing rustfrom above,galvanic corrosion from the various m etals in the system – and a perfectcorrosion collection pointin the top of each bolt!

TH ER E IS A B ETTER W A Y!

A STM B II7 TESTIN G 3000 hour hot salt fog testing results

M ild steeltestpiece

U nprotected surfaces have exfoliating rust

Protected surfaces have no corrosion – nuts and bolts turn freely by hand

N A V SEA 23236C LIST O F R EQ U IR ED A STM TESTS

A pplications on U S Coastguard vessels

Pr ot ect i on f or w h ol e sy st em Pa ssi ve b a r r i er p r ot ect i on Act i ve cor r osi on i n h i b i t i on Ea sy a ccess t o su b st r a t e Ea sy a ccess t o su b st r a t e Re- usa b l e n on - t ox i c m a t er i a l

At m osp h er i c t est sa m p l es i n N or w a y a n d sp l a sh zon e t est s i n N or t h Ca r ol i n a

ASTM a n d ot h er t est i n g of q u a l i t i es a n d f u n ct i on s Lon g - t er m t est i n g w i t h l ea d i n g a ut h or i t i es su ch a s DN V Lon g - t er m a p p l i ca t i on p r og r a m m es Pr og r a m m e d ev el op m en t f or r ecog n i sed CI STP st a n d a r d Ap p l i ca t i on d ev el op m en t w i t h m a j or com p a n i es

Lon g - t er m a p p l i ca t i on p r og r a m m es

A t y p i ca l col l ect i on of f l a n g es a n d va l v es f r om t h e Br i t a n n i a

Coa t ed w i t h En vi r op eel , t h e Br i t a n n i a ’ s f l a n g es r em a i n r ust f r ee

Ap p l i ca t i on d ev el op m en t w i t h m a j or com p a n i es

AM AZI N G RESULTS On stored conveyor pulleys • Return for replacement without Enviropeel • Failure rate with full Enviropeel protection

4 4 .5 % 0%

On operational pulleys – No bearing has failed since Enviropeel applied, previous average only 9 months! • Average bearing life in original trial location: 9 m on t h s • Current bearing life in original trial location with Enviropeel applied: 3 6 + m on t h s • Resulting component life increase: 400+ % • Resulting saving in pulley changeout costs: 63+ % • Reduction in maintenance costs: 95 % • Percentage of Enviropeel costs to rebuild costs: 10-15 % • Percentage of Enviropeel costs to pulley change out costs: 5-7 % • Resulting percentage reduction in risk exposure: 63+ % • Anticipated increase in component lifetime: 600 %

The innova tion a nd cost- sa vings a chieved by the use of Enviropeel for BHPB a nd Rio Tinto won the WA Engineers Austra lia Engineering Excellence Awa rd in 2005

Th a n k You

The Alocit Group Presents Delta T Marine

Don’t confuse Ceramic Insulation Coatings (CICs) and Ceramic Coatings Ceramic Insulating Coatings (CIC’s) produce temperature differentials across the surface of the coating. Ceramic Coatings are designed to protect or change aesthetic color when applied (i.e. protective coatings for engine pistons or decorative tile coatings.) Ceramic Rooftop Coatings – elastomeric binders with glass sphere particles.

What is the track record of Delta T Marine? Delta T has been used on over 300 vessels since 1995 as a thermal insulation material. Vessels include OSV, anchor handlers, cargo ships, asphalt ships, aircraft carriers, tankers, ferries, and yachts.

Why use Delta T Marine? Provide same insulation results as conventional systems at a reduced cost. Ease of application – REDUCE MAN HOURS! – escalate time frame of the vessel. Reduces heat transfer efficiently – no hot spots or vertical heat piping. Not space constrictive on application – no need to carry insulation into area resulting in time loss. Can be sprayed with hot work ongoing! No shut down. Installs easily on irregular surfaces – stiffeners Lightweight Protects surface due to direct bonding Reduces/Eliminates condensation without worry of CUI Existing yard personnel can apply the coating with little training.

Conventional Marine Insulation Disadvantages of conventional insulation Shooting pins Priming pins Cutting insulation Providing vapor barrier Capping pins Labor intensive/space constrictive. Long-term issues with conventional insulation Vibration Repairs & maintenance Installation problems Water immersion - ruins effectiveness Promotion of fire – foam systems Difficult to see/locate problems Corrosion under insulation (CUI)

Advantages of Delta T Marine No more CUI issues Repairs easy with touch up process No regular maintenance Virtually no wear Total inspectability without any removal Protects the substrate at all times

Easy to use New method using spray-applied system Tape off any unwanted areas Apply coating Clean up Yield increase to 1000 ft²/hr (100m²/hr)

What about R-Value? If used inside a vessel and applied directly to shell, wall, stiffeners, overheads at 1.0-1.5mm an RvE value of 9-13 is achieved*. Delta T used in combination with conventional wallboard inside a vessel provides better insulation protection than if 2-3” of batt insulation was used**. Reduces airborne structural noise 50-80% depending on thickness of coating. Equates to 2” of Mineral Wool *** * Biotect Labs ** Independent study NAS *** Gieger and Hamme testing

Comparison Data 2” R13 Batt Insulation vs Delta T @ 1.0 mm

1" Measurem ent aw ay from Surface 200 190 180

Tem perature Deg F

170 160 150

Av Outside Temp

140

1" from Conv. Insulation

130

1" from Coating

120 110 100 90 80 70 60 0:00

0:30

`1:00

1:30

2:00

2:30

3:00

3:30 Tim e

4:00

4:30

5:00

5:30

6:00

6:30

7:00

Case History Condensation barrier/thermal barrier

Vessel: Laney Chouest Problem: CUI, Personnel Protection, thermal barrier Solution: Delta T stopped condensation and provided comfortable interior environment

Vessel: Kennicot Problem: Condensation Solution: Delta T stopped condensation and provided comfortable interior environment

Case History Condensation barrier/thermal barrier

Vessel: USS Constellation Problem: CUI, Personnel Protection, thermal barrier Solution: Stop condensation and provide comfortable interior environment Vessel: Barge Alvania Problem: Loss of heat and personnel protection Solution: Prevent heat loss and protect personnel from burn risk with Delta T

What about fire? Passes ASTM E84-87 Score 5 Flame Spread Score 5 Smoke Developed Passes IMO A.653.16 Passes No Toxic Effects of Smoke Non-combustible coating

What about the Environment? Water-based - Zero VOCs Non-toxic Saves energy Improves efficiency

Compliances Complies with most world-wide marine certifications

Technical Data Mechanical Properties Weight/litre (Wet) 0.57 kg/l Density 0.56 Accelerated aging 2100+ hrs Peak Operation temp. 500 deg F (260C) >350 Load Operation temp deg F (176C) PH 9.30 Radiant Heat Barrier 100% UV Reflection Moisture barrier Passed Cross Hatch Adhesion 100% Pull-Off Strength 300 psi Elongation 85% Coverage 0.5mm 1.47 m²/litre Weight/M² @ 0.5mm 0.35

Thermal Properties Thermal Cond. 0.097 w/m²/k Emmisivity 0.15 Reflectivity 0.85 Transmittance 0.00 Sound reduction 3-5 dB loss @ 60dB Flame Spread 5 Smoke developed 5 IMO A.653 (16) Passed Toxic effects heat None Heat Flux low R-Value Equivalency R9-15 USCG Certified ABS Certified DNV Certified Lloyds Certified

The Alocit Delta range also includes Delta dB Delta dB is a sound-dampening coating designed to reduce structural and mechanical noise generated through substrates and surfaces. Delta dB offers a very cost effective solution to sound damping problems in a flexible easy to use spray coating format.

Delta dB was tested to contrast the sound generation from a coated surface with that from a non-coated surface. Bells were setup in a test that applied 4.5 kg of force to ring the bell on a weighted arm. Graphs show the non-coated version of the bell vs. the coated bell. The relationship of the graph is decibels (dB) on the "Y" axes vertical area) and time of the vibration or sound on the "X" axes (horizontal area).

Thank You

The Effect of a Foul Release Coating on Propeller Noise and Cavitation Robert Mutton Mehmet Atlar, Martin Downie University of Newcastle upon Tyne, UK

Colin Anderson International Paints Ltd, Uk

Introduction • • • • • •

Why Coat Propellers? Previous Research Model Propeller Tests Noise and Cavitation Results Future Research Directions Conclusions

Why Coat Propellers?

Propellers are Usually Polished Clean

Why Coat Propellers? •When the reduction in ship performance is associated with the condition of the ship hull, the effect of the propeller surface condition is often overlooked. Nevertheless, the effect can be significant. •In absolute terms, the effect of the propeller surface condition is less important than the hull condition, but significantly more important in terms of energy loss per unit area. In economic terms, high return of a relatively cheap investment can be obtained by propeller maintenance. Mosaad, 1986

•A good propeller coating, prevents the build up of fouling and maintains the blade surface quality for extended periods

Propeller Coating Research During the 1939-45 war, limited trials were carried out using coatings to protect steel propellers, no follow up. The US Navy, Canadian Navy and UK Navy have carried out research on polymeric coatings (mainly Polyurethane and Neoprene) for both cavitation resistance and control of fouling. Low surface energy materials such as PTFE and nylons were tried, but with limited success. Recent Japanese trials with Silicone Resin coatings

Silicone Propeller Coating • Based on PDMS, molecule with a flexible backbone • Used Because: • Environmentally Benign • No biocide present

• Proven Drag Benefits • Work of Dr Candries, Newcastle University

• Proven Antifouling Capability • >100 ships hulls coated to date, now 180 propellers also

Silicone Coatings

Silicone Based Antifouling Coating

Copper Based Antifouling Coating

Application Sweep Blast with Garnet

Epoxy Anti-Corrosive

Tie Coat

Silicone Top Coat

Previous Research • Computer Simulations showed that the coating had the equivalent drag to a new or well polished propeller. • Sea trials have shown no little difference between a uncoated and coated propeller. • The coating significantly changes the roughness characteristics of the propeller blade surface, both amplitude and texture.

Bernicia Sea Trials • Trials conducted using University Research Vessel ‘Bernicia’ • Design based on a fishing boat • Two sets of controlled trials, one uncoated, one coated, over a measured mile

Final Power Curve Comparison Errors estimated at 10%

140000.00

120000.00

Corrected Shaft Power (Watts)

Uncoated Trial 100000.00

Coated Trial

80000.00

60000.00

40000.00

20000.00

0.00 6.00

6.50

7.00

7.50

8.00

8.50

Tide Corrected Speed over Ground (knots)

9.00

9.50

Bernicia Propeller after 24 months

Propeller Roughness Ra Frequency Distribution for the Uncoated Propeller, the Newly Coated Propeller and after 1yr in Service 40.00%

35.00%

New Coating Coating After 1yr in Service

30.00%

% Frequency

Uncoated Propeller 25.00%

20.00%

15.00%

10.00%

5.00%

0.00% 0

2

4

6

8

10

12

14

16

Ra value (microns)

18

20

22

24

26

28

Sm Frequency Distribution for the Uncoated propeller, Newly Applied Coating and After 1yr in Service 100.00% 90.00% 80.00%

New Coating 1yr in Service Uncoated Propeller

% Frequency

70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 0

500

1000

1500

2000

2500

Sm Value (Microns)

3000

3500

4000

Model Propeller Tests • Tests have been conducted in controlled conditions using the Emerson Cavitation Tunnel. • Test conducted for uncoated propeller and coated propeller. • Looked into Effect of Coating on propeller Performance. • Also looked at noise and cavitation effect.

The Emerson Cavitation Tunnel

The model propeller Ship type Deadweight Length Overall Max Draught Speed Power (installed) Built

Medium Tanker 96920 tonnes 243.28 metres 13.616 metres 14.86knots 9893kW 1992

Full-Scale Propeller Dimensions Diameter Mean Face Pitch Expanded Blade Area Ratio Design Advance Coefficient, J

6.85m 4.789m 0.524 0.48

Model-Scale Propeller Dimensions Diameter = 0.35 m Expanded Area Ratio = 0.524 Pitch Ratio = 0.699 Material: Aluminium Alloy

Comparison of Open Water Characteristics in Atmospheric condition (water speed 4ms-1, Confidence limits 95%) 0.8

uncoated Kt uncoated 10Kq uncoated Efficiency coated Kt coated 10Kq coated Efficiency

0.7

Kt, 10Kq, Efficiency

0.6

0.5

0.4

0.3

0.2

0.1

0 0.25

0.35

0.45

0.55 Advance Coefficient, J

0.65

0.75

0.85

Noise Measurements •Noise measurements recorded using a Bruel and Kjaer type 8103 Hydrophone at both loaded and ballast condition •Analysed using the standard method of ITTC •The coating is theorised to mostly effect the broadband noise

Noise - Results •Results Revealed that the Coating DID have an Effect on Propeller Noise Net Propeller Noise - Loaded Condition V = 4.0 m/s J = 0.75 σ = 0.498 130

Uncoated Propeller SPL (dB ; re 1mPa, 1Hz, 1m)

120

Coated Propeller

At High Frequencies and High Advance Coefficients, the Coating Provides a Reduction in Noise

110

100

90

80

Net Propeller Noise - Loaded Condition V = 4.0 m/s J = 0.45 σ = 0.498

70 10

100

1000

10000

100000

130

Centre Frequency (Hz)

Uncoated Propeller Coated Propeller

At Low Advance Coefficients, the Coating Appears to Provides an Increase in Noise

SPL (dB re 1mPa, 1Hz, 1m)

120

110

100

90

80

70 10

100

1000


10000

100000

Noise - Results Net Propeller Noise - Ballast Condition V = 4.0 m/s J = 0.65 σ = 0.320 130


110

100

90

80

70 10

100

1000

10000

100000

Centre Frequency, Hz

Net Propeller Noise - Ballast Condition V = 4.0 m/s J = 0.40 σ = 0.320

130


120

SPL (dB ; re 1µ Pa, 1Hz, 1m)

SPL (dB ; re 1mPa, 1Hz, 1m)

120

As The Cavitation Increases the Difference between the two Curves Diminishes

110

100

90

80

70 10

100

1000


10000

100000

Cavitation - Results Inception •The Results show that the Coating Effects the Cavitation Inception Point Very Little •The Ballast Condition Shows a Slight Increase in Inception Advance Coefficient, Could be Due to Damage on the Coating as much as a Fundamental Effect

LOADED CONDITION (J) Uncoated

Coated

% Change

Inception

Inception

Inception

0.517

0.505

-2.32

Dissidence

Dissidence

Dissidence

0.513

0.510

-0.58

BALLAST CONDITION (J) Uncoated

Coated

% Change

Inception

Inception

Inception

0.542

0.590

8.86

Dissidence

Dissidence

Dissidence

0.540

0.557

3.15

Cavitation - Results Developed • The developed cavitation on the uncoated propeller was more intense, extensive and stable when compared to the coated one. • The less stable nature of the developed sheet cavitation on coated blades displays some cloud cavitation along the lower boundary. • The tip vortex was less well defined on the coated propeller.

Cavitation - Results Coated

Uncoated

J=0.50

J=0.40

Model Test Conclusions • There appears to be some effect of the coating upon the noise levels. • The beneficial effect of the coating appears limited to the broadband frequencies at the higher advance coefficients. • As the cavitation increases, the difference between the noise levels of the uncoated and coated propellers at smaller advance coefficients diminishes. • The developed cavitation on the uncoated propeller was more intense, extensive and stable when compared to the coated one. • The tip vortex was less well defined on the coated propeller • Further study is needed to draw firm conclusions

Full-Scale Comparison Basis Vessel for Model Tests was coated in 2001 Pictures are after 37 months in service, without cleaning Minor Damage to coating No Fouling Present

Full-Scale Trials • Work with Major Shipping Companies • Performance Monitoring before and after propeller coating • Aim to see what effects are achievable in service conditions including long term benefits and durability

Full-Scale Trials

Full Scale Performance Monitoring • Results are promising with reports of reduced fuel consumption, less slip and reduced noise in many cases • Performance Data has been collected from a number of vessels over a period of time to look at long term effects • Long-term coating durability has been proven to 37 months

Future Research Areas • Further Detailed Noise and Cavitation Measurements • Extrapolation of Model Results to Full Scale • Development of Standard Roughness Estimation Methods for Coated Propellers • Boundary Layer Measurements on Propeller Blades • Effect of Slime Layers on Coated Propellers • Research Towards Ideal Propeller Coatings

Conclusions • The results presented here, for the first time, are the results of an initial study into the effect of a Foul Release system on propeller noise and cavitation. • The coating DOES effect noise and cavitation. • The coating appears to reduce noise at high frequency (broadband noise) when the propeller operates at high J values (lightly loaded). • The coating appeared to reduce the extent and intensity of cavitation. • Based on these results a much larger study is underway.

ANY QUESTIONS?

Vacuum Consolidation of Commingled Thermoplastic Matrix Composites for Marine Applications M. Ijaz, P.N.H Wright, A.M. Robinson, A.G. Gibson, School of Marine Science and Technology School of Mechanical and Systems Engineering University of Newcastle upon Tyne Advanced Marine Materials and Coatings February 22nd - 23rd 2006

Acknowledgements This work began life as a project involving Newcastle University and PERA, funded by the DTI LINK Scheme on Structural Composites

CCME University of Newcastle upon Tyne

Background Traditional (i.e. thermosetting) composites are achieving increasing success in large engineering structures. This is being driven by the well-known advantages of composites over metals (strength/weight, corrosion resistance, formability). Though thermoset based composites are well established, alternatives such as thermoplastic based composites are of interest, driven by environmental requirements. So- what is the way forward for thermoplastics in large structures? University of Newcastle upon

CCME Tyne

Large themoset structures Hunt Class MCMV

Mirabella V Visby Corvette

Sandown Class MCMV Advanced Marine Materials and Coatings February 22nd - 23rd 2006

Large themoset structures Wind turbines

Sugar Grove bridge

Strongwell

Mudmats

Halgavor bridge


Aims of the work •To model and understand the vacuum bag consolidation process with a view to large scale marine applications, rather than smaller piece part manufacture. •To develop an understanding of the heat transfer required for the softening, impregnation and subsequent consolidation under vacuum of comingled thermoplastic reinforced composites.


Aims of the work •The effect of three critical processing parameters are investigated •Pressure •Temperature •Time at temperature


Current largest thermoplastic structures

PACIFIC 22 Mk II

Halmatic

Assault craft

Halmatic


Commingled Glassfibre-thermoplastic composites in present study Commingled Fabrics

Thermoplastic matrix

Process temp. oc

Comfil®

PET homopolymer

280-300

(Johns Manville)

(Semi-crystalline)

PET copolymer

210-230

(Amorphous)

Twintex™ (Vetrotex International)

PP

180-200

(Semi-crystalline)


CoPET Tg ~ 60oC (Amorphous)

PET Tg ~ 87 oC (Semi-crystalline)

Log E

PET Tm ~ 265oC PP Tg ~ 0 oC (Semi-crystalline) PP Tm ~ 165 oC

TemperatureoC Schematic plot of log modulus vs. temperature for amorphous and semicrystalline polymers CCME University of Newcastle upon Tyne

The Process Breather Bagging film

Release film

Mould tool

Sealant tape


Bagged laminate pack


Process modeling and characterisation Heating Consolidation

Advanced Marine Materials and Coatings February 22nd - 23rd 2006

Experimental Setup for process monitoring Oven

LVDT Thermocouples

Vac. Port

Vac. Bag

Release Film Breather Fabric Sealant

Commingled Fabric

Tooling

Data Acquisition Unit


Consolidation Test Rig

Displacement LVDT

Rubber Membrane

Vacuum

Insulation Sample Heater

Temperature


PET copolymer: heating 250 Oven air Breather

Temperature

o

C

190 Lam . Top Lam . Mid. 130

70 Lam . Bot.

10 0

700

1400

2100

2800

Tim e sec. CCME University of Newcastle upon Tyne

Mid. Tem perature

% Consolidation


% Consolidation

M id. Tem perature


IMPREGNATION MODELING

Governing law for consolidation rate  −B     T 

dX = F ( X )e dt Since B dX  T  F(X ) = e dt

A ‘master curve’ of F(X) vs. X can be derived from results by varying B. Model two stages separately


A simplified version of the Kamal equation is used to describe F(X):

dX = Ae dt

 −B     T 

X (1 − X ) n

m


This can be simplified further, and applied to the two processes:  − BA     T 

dX A mA ( ) = AAe 1− X A dt  − BB    dX B mB  T  (1 − X B ) = AB e dt Where:

X = fX A + (1 − f )X B


calculated m easured


100 calculated

% Consolidation

75

m e asured

50

25

0 0

700

1400

2100

2800

Tim e sec. CCME University of Newcastle upon Tyne

Conclusions z Consolidation occurs in two stages {A low temperature solid state de-bulking near to Tg {A Full melt impregnation at a higher temperature (at and above Tm in the case of semi-crystalline matrices) z The experimental rig described has proved useful in determining consolidation characteristics z It is hoped that this understanding of the consolidation process will allow further exploitation of these materials in the marine sector by allowing for consistent and repeatable production CCME University of Newcastle upon Tyne

Please address questions to: Dr P Wright: [email protected] Prof A Gibson [email protected] Advanced Marine Materials and Coatings February 22nd - 23rd 2006

ADVANCED MARINE MATERIALS & COATINGS

Recommend Documents