116

J.Nav.Eng. 45(1). 2009

"This paper was presented at the "Warship 2009 – Airpower At Sea" Conference, Organised by the Royal Institution of Naval Architects".

THE HYDRODYNAMIC DESIGN OF THE QUEEN ELIZABETH CLASS AIRCRAFT

CARRIERS

BY

A. HARRIS, & T. DINHAM-PEREN BMT DEFENCE SERVICES LIMITED, UK

L. SEARS THALES NAVAL LIMITED, UK

N. IRELAND QINETIQ, UK

ABSTRACT

The hydrodynamic development of the Queen Elizabeth Class Aircraft Carrier is described. The initial development of the ship design focussed on a podded propulsion hullform prior to the adoption of the twin shaft arrangement.

Suitable hullform parameters were obtained from a database of previous tank tests indexed by non-dimensionalised resistance, followed by station fitting and hull fairing. An extensive programme of work to achieve acceptable propulsive efficiency, propeller induced vibrations and cavitation performance was undertaken. Numerical tools, towing tanks and cavitation tunnels were used. The manoeuvring and seakeeping performance of the final design have also been examined in a series of seakeeping and manoeuvring tests undertaken in the Ocean Basin at Haslar.

NOMENCLATURE

ABP Adjustable Bolted Propellers

ACA Aircraft Carrier Alliance

B Beam

CAD Computer Aided Design

CB Block Coefficient

CFD Computational Fluid Dynamics

CTOL Conventional Take Off and Landing

DOF Degree of Freedom

117

J.Nav.Eng. 45(1). 2009

ESTD Electric Ship Technology Centre

FEA Finite Element Analysis

FPP Fixed Pitch Propeller

GRP Glass Reinforced Plastic

IEP Integrated Electric Propulsion

IMO International Maritime Organisation

ITTC International Towing Tank Conference

JCA Joint Combat Aircraft

L Length

MCR Maximum Continuous Rating

MoD Ministry of Defence

MSCA Maritime Strategic Capability Agreement

MSI Motion Sickness Incidence

PID Proportional Integrative Derivative

PU Polyurethane

QMCS QUALISYS Motion Capture System

QE Class Queen Elizabeth Class

RAO Response Amplitude Operators

RLG Ring Laser Gyroscope

RN Royal Navy

RPM Revolutions Per Minute T Draught

URD User Requirements Document

WoD Wind over Deck

INTRODUCTION

The hydrodynamic design of the Queen Elizabeth Class Aircraft Carrier (QE Class) can be divided into four phases:

• Definition of user requirements and constraints; • Initial hullform design and testing; • Propulsion system interaction and optimisation; • Final testing for client acceptance.

118

J.Nav.Eng. 45(1). 2009

This paper aims to give the reader an oversight of the complete process in chronological order.

HYDRODNAMIC DESIGN REQUIREMENTS AND CONSTRAINTS

The design of the QE Class has been developed to meet an extensive set of user requirements. The hydrodynamic requirements were chosen to ensure adequate aviation operability, ability to deploy to an area of operations worldwide and ability to manoeuvre in company with other fleet assets.

Ship Speed Requirements

The maximum speed of the ship is not specifically declared in the user requirements. QE Class is however required to generate sufficient Wind over Deck (WoD) to facilitate launch and recovery of aircraft.

The ability to transit to an area of operations at a mean speed of advance is also required, which must include periods of concurrency flying for the air group when the ship is forced to adopt a course favourable to aircraft launch and recovery. The ship must then sprint to return to the task group.

The final requirement relating to ship speed is to operate aircraft in a virtual box defining an area of operations. A high intensity flying programme must be maintained. The ship cannot proceed continuously into the wind due to the constraint of the box dimensions so it must turn downwind to regain position between periods of aircraft launches and recoveries. Depending on the intensity of operations a high sprint speed may be required on the downwind leg.

The WoD requirements for the Joint Combat Aircraft (JCA) with varying payloads were researched. The worst case for the generation of WoD was taken to be calm conditions. An acceptable ship speed / payload trade-off was agreed.

It was found that a ship generating sufficient WoD would also be able to sprint to return to the task group on transit and also to regain position within the operating box.

The probability of calm or very light winds occurring in head seas of varying sea state was explored to ensure an adequate allowance was made for head seas added resistance in powering calculations.

Seakeeping Requirements

The user requirement references the criteria set within 1 for seakeeping assessment. The JCA is also designed to be operated within the criteria of 1. Analysis has been undertaken against the catapult launch / arrested recovery and Short Take-Off and Vertical Landing (STOVL) criteria as the design is to be adaptable to either configuration.

Analysis of the criteria set showed that lower accelerations on the flight deck were beneficial to overall aircraft operability. A long natural roll period was desirable

1 NATO Standardization Agreement 4154 Common Procedures for Seakeeping Assessment in the Ship Design Process

119

J.Nav.Eng. 45(1). 2009

to reduce vertical velocities and accelerations experienced at the flight deck, especially near the deck edge some 35m off the ship’s centreline. The roll amplitude requirements of 1 could be met in bow quartering seas in the highest operating sea state. The possibility of non-coincident wind and wave direction for aircraft launch and recovery plus benefits to aircraft handling in non-head seas directions led the customer to specify two pairs of folding fin stabilisers. Habitability and human factors related seakeeping performance is also improved.

The driving criterion for aircraft operability was identified as pitch, in terms of both global pitch amplitude and absolute vertical displacement of the aft most point of the flight deck. The importance of minimising vertical motion at the aft point of the flight deck is to ensure adequate clearance during the approach to an arrested recovery. Length proved to be the most significant variable in improving pitch performance. Ships of less than 250m showed a marked degradation in performance. Initially a target length of 290m overall was set for QE Class, with the current ‘Delta’ design being slightly smaller at 288m.

Other mission-based criteria defined in 1 were assessed, such as Transit and Patrol and Replenishment At Sea. None of these criteria influenced overall ship seakeeping performance as naval air operations required the lowest level of motion.

Manoeuvring Requirements

Manoeuvring requirements are based on achieving IMO Circular MSC.137 criteria 2 with additional military requirements. These include a reduced turning diameter, a minimum rate of heading change requirement and a requirement that QE Class is to be directionally stable. A specific emergency turn requirement was also specified to ensure it is possible to turn into the wind and launch aircraft under threat conditions within a given period of time.

The manoeuvring requirements in combination with considerations of controllability in waves led to the specification of unusually large rudders for QE Class. Rudder size was set based on non-dimensional rudder area of past designs known to have good manoeuvring performance.

The emergency turn requirement was particularly challenging. The trade off between deliberately low transverse metacentric height and a strict heel on turn limit for flight deck operations resulted in investigations into fast acting fluid or moving weight heel correction compensation systems. Further definition of the emergency turn scenario allowed the designers to demonstrate compliance without the need for a fast acting heel correction system.

Hydrodynamic Design Constraints

In addition to the performance requirements a number of constraints have influenced hydrodynamic design.

2 IMO Circular MSC.137(76)

120

J.Nav.Eng. 45(1). 2009

Static stability was of concern as the vessel must comply with Defence Standard 02-109 3. The minimum beam and transverse waterplane inertia was dictated by the projected end of life displacement and vertical centre of gravity estimate. Checks were performed and a minimum value of transverse metacentric height was set. Low initial stability in the damaged condition was acceptable due to the presence of buoyant spaces within the sponsons providing a margin of large angle stability. This permitted the long natural roll period required for low levels of acceleration on the flight deck.

Draught was limited to allow entry into the home port of Portsmouth as well as the lock gates of certain dry docks.

Internal volume and deck area requirements led to the adoption of the high freeboard design. Much volume is contained in the deep sponsons. Freeboard to the underside of sponsons and the side aircraft lifts has been examined during seakeeping testing.

Radar Cross Section considerations led to the adoption of flared topsides of constant angle (sometimes referred to as batter).

It should be noted that the only propulsion constraint imposed by the customer in the User Requirements Document (URD) was the preclusion of nuclear propulsion.

No specific Underwater Radiated Noise (URN) signature requirement was placed on the design of the propeller.

Balance of Requirements

With these constraints in place the midship section of the ship was well defined at an early stage. Attention turned to the selection of appropriate hydrodynamic parameters and the drafting of initial hull lines to provide a resistance optimised hull form. Assessment of powering would follow. Seakeeping and manoeuvring test programmes would be undertaken only once the ship’s speed performance had been proven.

HULLFORM AND APPENDAGE DEVELOPMENT

During the development of the hydrodynamic design of the hull form and appendages the target design point went through four major changes and many minor revisions between the start of the initial work in 2001 and the completion of the near final design by mid 2006. This occasioned a number of hull form design, propeller and appendage design studies and included about 17 major sets of models tests on 8 models covering resistance, propulsion, wake survey, seakeeping, manoeuvring, appendage-optimisation, rudder and skeg size optimisation, flap-optimisation, bulb-optimisation and cavitation tests.

The detailed description of this design and test process is outside the scope of this present paper, but the main issues influencing the selection of appropriate studies are described below.

3 Defence Standard 02-109 Stability Standards for Surface Ships Part 1 – Conventional Ships

121

J.Nav.Eng. 45(1). 2009

The Challenge

From the outset of this study, the combination of target length, breadth, draught, displacement and design speed put this design well outside of the range of data for past Naval Hull forms (including Aircraft Carriers). The target speed length ratio was lower, and the fullness of the vessel greater than normal Naval Practice.

Another requirement was that the midship section should have 7º batter throughout the depth of the vessel and this put additional constraints on the hull form. For finer and higher speed forms batter can be beneficial, but for forms such as this, batter has an adverse effect as it increases the effective block coefficient.

The overall result of these factors was that the resistance characteristic of this vessel would differ from that of a normal Naval Hull Form.

0

0.5

1

1.5

2

2.5

3

3.5

4

0.4 0.5 0.6 0.7 0.8 0.9 1

Speed Coefficient

Res

ista

nce

Coe

ffici

ent

Past Aircraft Carrier Form Present Form Slightly Fuller Form

Approx Design Speed for Present Form

FIG.1 – EFFECT OF FORM ON RESISTANCE CHARACTERISTICS

Figure 1 shows the difference between the characteristics, the speed and resistance coefficients being calculated on common denominators so that comparisons can be made. Three characteristics can be noted:

• The larger size and fullness means that the resistance is higher for the present form;

• The resistance increases very rapidly at speeds higher than the ‘design’

speed; • A slight further increase in fullness results in the resistance curve starting

to rise at a lower speed (i.e. ‘design’ speed is lower).

The position of the ‘design’ speed point depends very much on the fullness and details of the design of the vessel (such as bulb shape, length of parallel middle body and position of the fore and aft shoulders) and it will be seen that the present design is close to a limiting case – if the vessel were even slightly fuller the

122

J.Nav.Eng. 45(1). 2009

resistance and hence power would increase rapidly. Thus achieving the required powering performance was going to be a challenge.

Resources and Tools

A number of resources and tools were available for the design development work. At the initial design stage these consisted of:

• A database which contains around 5000 vessels of various merchant and naval vessels covering a wide range of parameters;

• A suite of programs for hull form manipulation, including the addition of

batter to otherwise vertically sided hull forms; • A set of in-house procedures for making estimates for resistance and

propulsion of new hull forms based on past data for similar vessels, including corrections for hull shape and effect of batter;

• Gaddwave, a CFD potential flow code for wave profile and resistance

evaluation; • An initial propeller design program; • A strip theory program for seakeeping evaluation; • A manoeuvring estimation program.

For more detailed evaluation a Navier-Stokes CFD program and model testing were available.

A number of hull form design evolutions were required during the course of this project and the general process was as follows:

• The database was interrogated for forms within a specified parameter range (allowing for the effect of batter) to obtain a selection of candidate vessels close to the design target;

• For these vessels, corrections were made to the resistance for the changes

in geometry required to match the target vessel (i.e. corrections for L, B, T, midship batter and CB);

• The vessels were sorted on the basis of corrected resistance. The form with

lowest resistance at target speed was selected and the lines examined. Froude Number is a major factor in the selection of the basis design and even small changes in this can result is a different basis vessel being selected;

• If this vessel was not suitable for some reason, the next best vessel was

examined until a suitable basis form was found. For example, the data base was indexed by LPP whereas a design like this is driven by LWL

123

J.Nav.Eng. 45(1). 2009

requirements. It took several iterations of the procedure to refine search parameters to find a suitable basis vessel;

• The area curve was extracted and used as basis for new vessel design; • New hull sections for new design were drafted and these differ

significantly from the original. Other good forms in the database, past experience and also the particular requirements of the new design all affected the section and waterline shapes used in the new design.

This process resulted in an initial design which had good powering qualities and provided a basis for further design development. An example of how good qualities from the basis hull are carried through is that it is the author’s experience that for even quite small changes in the required block coefficient or target speed, it is better to go back to the data base to get a good basis area curve rather then rely on standard area curve transformation techniques if minimum resistance characteristics are required.

This initial design was then further elaborated through a number of sequential design changes to explore the effect of variations in factors such as area curve shape, LCB position, hull section shapes and bulb shapes using GaddWave to evaluate the forms. This process resulted in the design being tuned to the exact design requirements and took the design process further than is possible using past data alone.

The process followed can lead to the development of new hulls which in some cases are more efficient than the ‘basis’ design and apart from a fairly esoteric relationship with the basis hull, are effectively completely new designs.

Most of the candidate basis hull forms were past ocean liners. These vessels are the result of over a hundred years of intensive form development. The powering performance of liners was sufficiently important for multiple form variants to be tested on a regular basis right up to the late 1960’s, at which point true liners were no longer built. After this period the exact combination of requirements that led to the development of the liner hull form no longer existed, but with the advent of slightly slower and fuller form naval vessels (such as QE Class) the requirement has now re-emerged and the data that exists for past liner forms is an invaluable resource for projects such as this.

Initial Design Work

Initial design work was undertaken by both BAE Systems and Thales during the competitive phase of the project. This section describes work undertaken on the Thales design which was eventually developed into the current design by the ACA.

The first few designs were initial hull investigations. These led to two variants, one a conventional twin screw design and the second a twin podded design. Both forms had a swan neck bulbous bow. A desk study report was produced giving initial powering estimates and looking at issues whether the large side sponsons would be acceptable from a seakeeping point of view.

124

J.Nav.Eng. 45(1). 2009

From these preliminary studies a further design was evolved. This design was about 290m length overall and had four pods arranged in a manner similar to QUEEN MARY 2 with two outboard fixed pods forward of two inboard azimuthing pods. In order to accommodate these four pods the form had a wide transom stern with shallow ‘V’ shaped sections. The deadrise angle of the lower portion of these aft sections was kept in excess of a recommended minimum value so as to avoid problems with stern slamming. This was a major point of concern as at this point in time reports were emerging from the cruise liner industry of slamming problems on podded vessels with flat sterns.

A set of calculations to give wetting height probability curves were performed and a set of model tests were performed at Haslar to look at resistance, propulsion, seakeeping and manoeuvring of this design. Of particular interest was a wake survey test to examine the nominal wake field for the aft pods when the forward pods were operating. This showed that the aft pods would not be unduly affected by the propeller race from the forward pods when travelling in a straight line.

This work showed the design to be broadly viable and was the basis of the design submitted by Thales at the close of the competition phase at the end of 2002.

265m Length Overall Design

As will be discussed in section 4, after the end of the competitive phase the ship design was revised. Initial guidance was provided on the effect on powering of changing the hull form and giving input on the effect of pods versus twin or quad propellers. The outcome was that the target design was revised to a vessel of about 265m overall with two conventional propellers.

This revision to the design parameters necessitated a re-run through the hull design process and in addition to length and breadth requirements, additional design constraints were given in that for three target displacement values the vessel had to meet stability and draught requirements. In order to evaluate this, a design tool was prepared which could introduce geometric variations in a basis design and then evaluate the requirements at the three required displacements. The tool was able to search the valid range of form distortion and find forms which met these requirements. These were then used as a basis for further design work during July to September 2003 which led to the definitive 265m overall design. During this process the weight of the design increased which made the design requirements more difficult to fulfil.

Model tests were undertaken at SSPA in Sweden over the period Nov 2003 to Jan 2004. The tests covered resistance, propulsion, wake survey and free running seakeeping and manoeuvring tests. These tests showed that the manoeuvring and seakeeping was generally acceptable and that there were low levels of stern, sponson and side lift slamming at all headings. Although seakeeping and manoeuvring performance was generally acceptable the shaft power requirement exceeded available propulsive power and the wake quality was poor due to the restriction on the stern design imposed by the shorter design length and the fullness of the form. Also the tested appendages had yet to undergo sufficient detail design contributing to the poor wake field at the propellers.

125

J.Nav.Eng. 45(1). 2009

Non-hydrodynamic design issues led to the reversion to larger ship dimensions. This afforded the opportunity to improve powering performance and continue appendage detail design.

288m Length Overall Design

A period of propulsion optimisation studies took place over the period January to May 2004 which is fully described in Section 4. A range of B/T variations of a hullform of 288m overall were developed to support these powering investigations. In the chosen design the rudder was moved as far aft as possible which allowed the fitting of larger diameter propellers with acceptable propeller hull clearances. Moving the rudders aft raised the issue of possible rudder ventilation when put hard over.

In the period August to October 2004 a set of resistance, propulsion, wake survey, free running seakeeping and manoeuvring and cavitation tests was undertaken at SSPA in Sweden using the ‘Near Design’ propeller. These tests showed the design to be viable with acceptable powering and cavitation characteristics.

In January 2005 a number of additional tests were also performed to look at the effect of varying the rudder and skeg size on manoeuvring control, course stability and heel in turn. As a result of these tests a middle rudder size and larger skeg were adopted.

Finalised Hullform

During Late 2005 and early 2006 the option of fitting a stern flap and a revised bulbous bow shape was investigated using both CFD and model tests. For the CFD studies a range of flap angle were investigated. Flap angle is defined as the angle the underside of the flap makes to the local buttock line angle of the parent hull. As a result of the CFD study, two angles (2º and 8º) of flaps were model tested. Results for a 0º duck tail were also available from earlier tests. A 2º flap was selected for further testing.

A parametric study of length, breadth and height variations of bulbous bows was undertaken using CFD in both the UK and France. These included a design prepared by Bassin d’Essais et Carènes (BEC). The BEC design had a central axis line more parallel to the keel line than the baseline QE Class design. The most promising design was tank tested and showed a significant reduction in shaft power requirement.

126

J.Nav.Eng. 45(1). 2009

FIG.2 – BULBOUS BOW DEVELOPMENT (COURTESY OF DELEGATION GENERALE POUR L'ARMAMENT / BASSIN D'ESSAIS

DES CARÉNES)

The finalised hull form thus included the skeg and bulb changes, but apart from a slight revision to the forefoot design for docking purposes, the form remained virtually identical to the initial 288m overall form.

In July to November 2006 a model of the finalised form with the ‘Design’ propellers was tested. The model was tested for resistance, propulsion, wake survey and cavitation tests and the performance was acceptable. The resistance of the form was assessed and at the intermediate to deep draughts was found to be better then recent best practice.

During these tests an oblique wake survey test was also performed on the model with yaw angles of ±20º, ±10º, ±5º. These showed some interesting effects, including a ‘wake reversal’ effect whereby the A-bracket on the ‘down stream’ diverts the oblique flow into the upper propeller disc, effectively reversing the normal wake peak found here.

Concluding this section, by December 2006 the hullform design had effectively been consolidated and validated for powering, seakeeping, manoeuvring and propeller cavitation performance. It was this design which was accepted for construction.

127

J.Nav.Eng. 45(1). 2009

POWERING THE QUEEN ELIZABETH CLASS

During the competition phase between BAE Systems and Thales culminating in January 2003, both teams had evaluated various propulsion system concepts for QE Class. Various mechanical and Integrated Electric Propulsion (IEP) system designs were evaluated. Propulsors studied included waterjets, propellers and podded propulsors.

Propulsion System Design Concept

The various propulsion system options evaluated during the competition phase had to address the differing requirements of a Conventional Take Off & Landing (CTOL) or STOVL aircraft in the ship design. During the competition phase, both teams were progressing two ship designs to satisfy either aircraft option. The power to be delivered into the water to achieve the performance requirements for WoD was significant and the operating profile, coupled with challenging range requirements required a high part load efficiency from the propulsion system across the entire speed range.

At the close of the competition phase, both teams had selected IEP systems as their baseline; the Thales design comprising four commercial pods (2 fixed, 2 azimuthing) and the BAE Systems design comprised of two shafts and a single azimuthing pod.

The MoD selected the Thales design to be taken forward into the demonstration phase of the project.

Propulsor Options

The selection of pods was predicated on the technical and commercial advantages offered; most notably reduced integration effort (compared with shafts) during the ship build programme and hence the removal of the propulsion system from the critical path. The pod arrangement was similar to that of the emerging QUEEN MARY 2 design.

Pods Versus Shafts

Following the down-selection, the project entered the demonstration phase, the purpose of which was to demonstrate to the MoD that the design is sufficiently mature and de-risked to provide confidence that the programme will be delivered to time within the budget. With the focus now firmly shifted to demonstration of a de-risked baseline design, a number of concerns related to pods were to be investigated in more detail.

At this time there were no pods in service at the required ratings for the QE Class and some of the pods in service at the time, with ratings significantly less than that required, were suffering from reliability issues in service (most notably within the cruise industry).

In addition, there was very little understanding of the capability of pods to meet specific military requirements related to Shock, URN and Electro-Magnetic Interference. On this basis, the MoD arranged a stakeholder day on 18th July 2003 to review the risks and agree the way forwards with respect to podded propulsion.

128

J.Nav.Eng. 45(1). 2009

The outcome of this stakeholder day was that notwithstanding the obvious benefits of pods, the risks were considered too high. The principal concerns were in delivering the required ratings, in service reliability, and the technical and programme risk in designing a pod to meet military requirements. This led to the decision to progress the ship design with a conventional propulsion solution comprising between two and four Fixed Pitch Propellers (FPP) driven by conventional shaftlines and electric motors.

The Change From Pods To Shafts

Following the decision to proceed with conventional shafts, the number of shafts/propellers and the arrangement and matching of the prime movers to drive them had to be evaluated against the customer requirements. The extant aft hull lines were also specifically designed for pods and hence a redesign would now be required once the number of shafts had been selected.

Number Of Propellers/Shafts

The powering requirement for QE Class is significant, hence the evaluation of the number of shafts/propellers required to deliver power into the water. The balance of ship design considerations, performance and technical risks as well as procurement, ship build and through life costs was one of the principal design considerations.

A working group comprising a broad range of industry experts was established by the project to evaluate the design options, which comprised two, three or four shaft designs. Due to the asymmetry of the arrangement and impact on the hull and machinery design to incorporate a centreline shaft for the triple shaft design this was dismissed quite early in the evaluation process. The four shaft design offered a more optimal thrust loading for the propeller, however it introduced greater complexity and exhibited significant issues in arranging machinery to meet survivability requirements. The procurement and through life costs were also not attractive for this solution. This left the twin shaft solution as the preferred way forwards, however concerns over propeller thrust loading, tip speed, pressure pulses and the delivery of the required power down two shafts still caused concern, hence the same working group was tasked with de-risking the twin shaft design.

TWIN SHAFT DESIGN

Prime Mover/Propeller Matching

There was a strong desire to pull-through the IEP technology being de-risked at the Electric Ship Technology Centre (ESTD) in Whetstone for the Type 45 Destroyer programme. However, whilst Rolls-Royce had evaluated a propeller design for this same rated rpm, its higher tip speed, reduced efficiency and poor pressure pulse predictions highlighted that the rated rpm of the Type 45 was too high to permit a balanced naval propeller to be produced at the larger diameter selected for QE Class.

129

J.Nav.Eng. 45(1). 2009

Following various technical discussions with the motor supplier, the risk in delivering the increased torque required to provide rated power at a reduced shaft speed was considered to be acceptable on the basis of similar motor and converter designs already tested for naval application. Having secured a reduction in the shaft speed and with a need to minimise shaft rake due to arrangement constraints, the propeller diameter was fixed at a maximum of 6.7m, giving a 30%D tip clearance against the hull.

The Propeller Wake Field

Wake survey and cavitation tunnel tests of the initial twin shaft hull form (a shorter 265m overall design) showed that there was a great deal of optimisation required for the shafting appendages in order to improve an obviously poor wake field. The large shaft out bossings where the shafts exited the hull and the bulky intermediate A-brackets were observed to drive the high wake peak. It was further observed that large wake shadows were also apparent due to the design and the poor alignment of the A-bracket arms.

In support of the working group Rolls Royce had already conducted a series of initial propeller designs for a two, three or four shaft design against a nominal assumed wake field.

These propeller designs were all seen to exhibit low open water efficiency compared to that typically expected and exhibited higher than desired pressure pulse predictions. Pressure pulses are a cause for concern with regard to local and wholeship hull borne vibration. Further, from the marine engineering perspective, a poor wake field gives rise to higher propeller induced excitation of the shafting system, particularly axial and torsional vibration as the blades pass through the non-uniform wake field. The high propeller loading and poor wake field would have been likely to result in erosive cavitation and low capitation inception speed.

It was decided at this point to remove the shaft out bossing completely, adopt a shaft tunnel bearing to support the forward tube shafts and embark on a detailed appendage design programme. This coincided with the final change in ship dimensions to the 288m length overall design.

Appendage Design

Optimisation of the appendage design was a principal package of work in the de-risking programme for the twin shaft design. Due to the investment and timescales in further cavitation tunnel testing, a partial hull and appendage model was built to enable initial de-risking of the appendages design in a wind tunnel. Through flow visualisation undertaken using CFD and wind tunnel, modifications were made to optimise the appendages prior to building a partial model for proof of concept cavitation tunnel tests at a cavitation tunnel in Portchester, UK. Due to the size limitations of the tunnel, wake screens were fitted to produce a wake field comparable to that produced from CFD.

In parallel and following consultation with Lloyds Register, revised A-bracket designs were also developed by the structures team to reduce the arm sections and bearing boss diameters. The connection of the arms to the bearing boss was also subject to detailed review to optimise wake and reduce any cavitation erosion in

130

J.Nav.Eng. 45(1). 2009

service. As this design was a departure from the Lloyds rule requirements, Finite Element Analysis (FEA) was utilised to support a first principles approach to the design.

A 1:30 scale towing tank model was constructed. This was appended and a wake survey undertaken at the propeller disc location. The resulting wake plots were compared to CFD results and supplied to Rolls-Royce for design of the propeller. Good correlation between computed and measured wake was demonstrated. Plots of axial wake before and after shaftline and hullform modifications are given in Figure 3 (red = low velocity, blue = high velocity).

FIG.3 – IMPROVEMENT IN AXIAL WAKE IN PLANE OF PROPELLER DISC

(COURTESY QINETIQ) AND CAVITATION COMPARISON

It was desirable not to perform cavitation testing with a stock propeller; hence Rolls Royce were contracted to supply a model propeller termed the ‘near-design’ in order to get the most value out of the tests. These tests were attended by the working group and included representation from a wide range of industry and MoD experts.

The group reported back to the project and the decision was taken that the risk of the twin shaft design was acceptable and further mitigations to the remaining technical risks could be progressed.

TESTING THE QUEEN ELIZABETH CLASS

QinetiQ were contracted in late 2007 to manufacture a model of the final hull form and carry out the final hydrodynamic testing programme involving an extensive series of seakeeping and manoeuvring experiments.

131

J.Nav.Eng. 45(1). 2009

Model Build

A new model of QE Class was manufactured to accurately represent the final hydrodynamic hull form. The model was manufactured to a sufficient standard of robustness to match the life of the class.

The size of the new model balanced the desire to be as large as possible to minimise scale effects, whilst being of a practical size to install the necessary instrumentation and propulsion. This resulted in a 1:44 scale model, the overall beam becoming the limiting factor in order to fit within the dock in the test facility. This yielded a model of about 7 metres in length and 1.6 metres in breadth.

The model was constructed of a central canoe body and five sponson components. The canoe body was constructed from GRP, the sponsons from high density PU foam coated with epoxy resin. The aircraft lifts were made from timber and could be fitted in their raised or lowered positions. The model was manufactured to ITTC tolerances 4. The model included all appropriate underwater appendages, shaft brackets, shafts, couplings, fairings, rudders, bilge keels, active fin stabilisers and transom platform extension. Model stock propellers were used to match the parent propellers diameter and pitch. The hubs of the model propellers were modified to mimic that of the ABP design to be fitted to the ships.

The accuracy and detail specified for the model required QinetiQ to develop a 3D CAD model of the QE Class, Figure 4.

FIG.4 – IMAGE OF 3D CAD DEFINITION OF QE CLASS MODEL

This process allowed an independent prototype of the final hydrodynamic model to be developed. In particular, the vessel shaftlines were constructed from a number of independent production drawings and assembled to ensure a cohesive design.

Where possible, the model was assembled from components manufactured by numerically machined processes, for example the skeg hung rudders, Figure 5.

4 International Towing Tank Conference, Recommended Procedures, Model Manufacture, 7.5-01-01-01, 2002, Revision 01.

132

J.Nav.Eng. 45(1). 2009

FIG.5 – MODEL SHAFT BRACKET AND RUBBER MANUFACTURE

The generation of the 3D CAD model allowed the flight deck islands and side sponsons to be numerically cut thus ensuring a high fidelity representation of the hull form, Figure 6.

FIG.6 – COMPLETED MODEL OF QE CLASS

The model was marked up with a grid and significant openings and intakes highlighted for video filming in the Ocean Basin.

Facility

The model tests were conducted in the Ocean Basin facility situated at Haslar, England, a large indoor model basin of dimensions (metres) 120 (L) by 60 (W) by 5.5 (D). The basin is fitted with a 5 segment wedge type wavemaker capable of generating regular waves of maximum height of 0.5 metres.

Prior to these tests, the MoD, as part of the MSCA contract awarded to QinetiQ, funded the provision of a new QUALISYS motion capture system. This effectively records track and speed of the model to a positional accuracy of 2mm, and allows the 6 DOF motions to be calculated, for any point on the model.

Instrumentation

The QE Class model included a significant amount of onboard instrumentation, measurement transducers, data logging, propulsion and control systems and

133

J.Nav.Eng. 45(1). 2009

battery power. In addition, it also included strain gauged rudder stocks and a four fin active stabiliser system. The onboard instrumentation included:

• Ring Laser Gyroscope (RLG); • QMCS infra red light emitters; • Rudder actuation and helm angle potentiometer; • Strain-gauged rudder stocks; • Active fin stabilisers; • Propulsion motor, gearbox and controller; • Propeller shaft tachometers; • Propeller shaft thrust and torque dynamometers; • Relative motion wave probes.

The RLG was installed primarily for course control by the autopilot, albeit that it's 6 DOF motion data was used for verification of the QMCS data. An autopilot course keeping algorithm of the proportional-plus-look-ahead type was used during the testing incorporating representative rudder rates. The fin stabiliser controller was based on a simple PID form which was tuned during forced roll tests to give an appropriate response.

The model utilised modern data acquisition techniques, where data and control was transferred between the model and the shore station by wi-fi. The data was filtered to remove high frequency noise. Data were sampled at 50Hz model scale and analysed directly after acquisition to ensure consistency, then stored in calibrated engineering units. A significant number of analysis computer programs were utilised within the MATLABTM environment, the data control and capture systems produced within a LabVIEWTM environment. Ultimately, up to 30 different data channels were recorded at any one time.

Test Programme

The model test programme comprised of seakeeping and manoeuvring tests. Specific ship conditions of varying displacement, trim, GMT and roll period were tested. The conditions were selected to cover the operating range of the ship from its ‘new’ condition on initial trials to its ‘end of life’ condition after a number of years operation.

The seakeeping tests comprised motion measurements for a range of ship speeds covering slow cruise to maximum speed. The tests were conducted for sea states 6 and 7, more extreme wave conditions had been evaluated during previous model test campaigns. Both long crested irregular and regular wave tests were conducted, the latter to support measurements for wave headings from stern aspects. Measurements and observations were conducted with the aircraft hangar lifts in

134

J.Nav.Eng. 45(1). 2009

their ‘up’ and ‘down’ positions and with the fin stabiliser system both active and passive.

The programme included a series of roll decrement and forced roll testing for a range of ship speeds. The results of the forced roll tests, using the model active fin stabilisers, provided support to the control engineering of the ship fin stabiliser system.

The manoeuvring tests were conducted in calm water only, with no provision for current and wind action on the model and comprised:

• IMO manoeuvring tests including turning circles, zig-zag manoeuvres and astern stopping tests at MCR and maximum helm angles;

• Supplementary tests comprising turning circles with pull-out manoeuvres,

zig-zag manoeuvres, low angle weave, yaw checking and limited control; • Emergency manoeuvres comprising combined accelerating and turning

tests.

The overall test programme, including all necessary calibration data, required in excess of 800 individual runs.

Seakeeping Tests

The model was conditioned to the required displacement, trim, roll, pitch and yaw gyradii, GMT and roll period and calm water speed calibrations conducted to determine the shaft RPM versus speed relationship.

Initial forced roll tests, at a single ship speed and mean condition, were conducted to tune the model active roll stabilisers. Validation of the achieved performance was carried out in stern quartering seas where the system gave an overall roll reduction of close to 90%. These control gains were used throughout the testing, accepting that the measured roll reduction performance would be unlikely to be repeated across all speed and heading combinations. The forced roll tests provided forced roll gains (Roll/Fin angle) for a range of ship speeds and fin driving frequencies, the roll decrement tests provided damping coefficients for the hull form with and without the fins deployed.

Ship RAO were produced from the motion responses for 6 DOF, total motions at different positions onboard the ship were synthesised in terms of displacements, velocities and accelerations using MATLABTM routines applied to the model responses measured by the QMCS system. MSI values were calculated at 1 and 4 hour intervals. A photograph of the QE Class model in head seas is shown in Figure 7.

135

J.Nav.Eng. 45(1). 2009

FIG.7 – QE CLASS CARRIER MODEL IN HEAD SEAS

Video records and data from relative wave height measurements, were provided for a review of wetness. A comprehensive dataset was provided to the ACA for analysis and comparison against performance requirements.

Manoeuvring Tests

The IMO tests comprised turning circles to both port and starboard, to verify turning symmetry, at close to the maximum speed of the vessel and at maximum helm angle. Zig-zag manoeuvres of 10°/10° and 20°/20° were also conducted to an agreed rudder rate. A requirement of IMO is the conduct of a full astern stopping test. This was simulated at model scale to provide estimates of the ship’s reach during stopping as follows:

• Shafts locked at full speed; • Windmilling shafts from full speed; • Shafts locked at full speed then astern RPM applied when about half full

speed reached; • Windmilling shafts from full speed then astern RPM applied when about

half full speed reached.

Some of these scenarios are not practical on the QE Class and required significant effort to simulate with the model propulsion arrangement, which differs from that on the ship. However, the tests did provide a measure of the shaft torques experienced during stopping and highlighted the need for astern RPM to bring QE Class to rest within a reasonable distance. The stopping manoeuvres were conducted using pre-programmed automatic control to ensure that consistency was achieved between runs.

The testing also included runs to assess the low speed control and yaw checking ability for both ahead and astern operation and control with one shaft locked and the opposite rudder only operational.

A comprehensive series of turning circles and pull-outs were conducted for a range of speeds and rudder angles, Figure 8.

136

J.Nav.Eng. 45(1). 2009

FIG.8 – EXAMPLE TURNING CIRCLE AND ZIG-ZAG PLOTS

In addition to the 10°/10° and 20°/20° zig-zags, much smaller zig-zag angles down to 3° were conducted to examine the ability of the model to manoeuvre under low rudder angles at low speed. Plots of rate of turn versus rudder angles were produced to determine the degree of directional stability.

Emergency Manoeuvres

Part of the manoeuvring tests was the simulation of an ‘Emergency’ manoeuvre related to launching aircraft from the QE Class. Prior to these tests a number of related studies had been conducted, including numerical simulation and physical model tests. These formed the basis of establishing the key parameters and operational envelope for such a manoeuvre:

• Propeller shaft RPM ramp rates;

137

J.Nav.Eng. 45(1). 2009

• Maximum heel angle in turn; • Maximum heel angle in last phase of manoeuvre; • Helm angles; • Initial and final ship speeds; • Shaft torque limits.

Initial predictions were made, prior to testing, to reduce the actual number of scenarios to be tested. Ultimately, these predictions highlighted the trade off between turning ability, acceleration and the heel limits set for such a manoeuvre. The tests showed that the heel angle requirement in the last phase of such a manoeuvre drives the requirement for the QE Class to have completed its turn in the shortest possible time. The requirement to both accelerate and turn in a short space of time drives the shaft RPM ramp rate but is conditioned by the shaft torque to remain within the overall maximum limits for the ship’s propulsion system.

A number of manoeuvres were conducted to determine the minimum time for such a manoeuvre whilst still meeting the constraints imposed. Satisfactory scenarios were those categorised by moderate ramp rates and helm angles which better allowed the manoeuvre to be carried out within the allowable heel angle and shaft torque limits. An example of a typical manoeuvre is shown in Figure 9.

FIG.9 – EXAMPLE OF 'EMERGENCY' MANOEUVRE PLOT

CONCLUSIONS

A hullform based on commercial practice has proved to be a suitable basis for a large aircraft carrier design. Acceptable powering performance has been achieved. Risks associated with the interaction of the hullform and propulsion systems of

138

J.Nav.Eng. 45(1). 2009

previous highly powered naval vessels have been addressed by physical and numerical modelling.

A comprehensive series of seakeeping and manoeuvring experiments on a large scale model of the QE Class aircraft carrier design has successfully been conducted. This provided the ACA with the dataset from which compliance with specified performance can be determined.

ACKNOWLEDGEMENTS

The authors would like to formally acknowledge the excellent relationship and cooperation that existed between the ACA, BMT and QinetiQ during the development of the hydrodynamic design of the QE Class aircraft carrier.

The authors also wish to acknowledge the contribution of the staff of Lloyd’s Register Technical Investigation Department, SSPA Gothenburg, the Rolls-Royce Hydrodynamic Research Centre and Bassin d’Essais des Carènes.

Any views expressed are those of the authors and do not necessarily represent those of the ACA (and its member companies), BMT Defence Services and QinetiQ.

AUTHORS BIOGRAPHIES

Andrew Harris graduated in Ship Science from the University of Southampton in 1995. Joining .BMT Defence Services Limited he undertook a range of submarine and surface ship in-service support tasks. Since 2002 he has been seconded to the Aircraft Carrier Alliance, initially responsible for the management of hydrodynamic design and more recently as the project’s Platform Architecture Manager.

Tom Dinham-Peren is the Chief Hydrodynamicist at BMT Defence Services Limited. He has over 25 years experience of ship hydrodynamics with particular expertise in the fields of hydrodynamic hull design, resistance, powering and seakeeping. He has been responsible for the initial Type 45 Hull form studies carried out by BMT SeaTech for the then VT Group Ltd and for the large model test programme for the Type 45 AAW Destroyer which was carried out by BMT. More recently, he has been responsible for the hull design of the QE Class for the Aircraft Carrier Alliance and model testing programme. He has a BSc in Naval Architecture from the University of Newcastle upon Tyne.

Leon Sears is a Senior Marine Engineer at Thales Naval UK Ltd. Since 2003 he has been working as a technical lead in the Aircraft Carrier Alliance Power & Propulsion Team, responsible for the propulsion, steering and stabiliser systems. He is a qualified marine engineer and has worked on the design of propulsion and auxiliary systems across various defence projects for UK and foreign navies over the past 10 years. Prior to working in the defence industry he worked in the commercial shipbuilding industry, responsible for engineering systems designs for fast ferries, coastguard cutters, fast attack craft and fleet auxiliary vessels.

Nick Ireland is a Principal Hydrodynamic Consultant working for the Maritime Platforms Practice at QinetiQ Haslar. He is a qualified naval architect working on a range of projects relating to propulsor design and general hydrodynamics. He has

139

J.Nav.Eng. 45(1). 2009

previously worked in the hydrodynamic research field and UK warship building industry, specifically ferry safety and the design and construction of high speed naval vessels. He provides support to both the UK MoD and Industry on a range of propeller and hydrodynamic related consultancy. He also represents QinetiQ and the UK in a number of European hydrodynamic forums.

"The Royal Institution of Naval Architects is an international professional society whose members are involved world-wide in the design, construction and maintenance of military, commercial and recreation marine vessels and structures, The Institution publishes a range of leading technical journals and organises an extensive programme of conferences and training courses, covering all aspects of the global maritime industry. Details of membership, publications and events can be found on the Institution website at www.rina.org.uk "