Proceedings of the 4th Ship Control Systems Symposium, Den Helder, The Netherlands, Volume 1

2

DEC. 1984

ARcHERROCEEDINGS

Lab. y. ScheepsbouwIwn

Technische Hogeschool

Delft

P1975-7

Volume 1

THE SYMPOSIUM WILL BE HELD IN THE NETHERLANDS, THE HAGUE CONGRESS CENTRE 27-31 OCTOBER 1975

Statements and opinions expressed in the papers are those of the authors, and do not necessarily represent the views of the Royal Netherlands Navy.

The papers have been reproduced exactly as they were received from the authors.

VOLUME 1

CONTENTS

Introduction by W.J. Blumberg. SESSION A:

Chairman: B. Veldkamp

Vice-Admiral R. Neth.N., Commander in Chief Conning a ship with different steering systems.

I. Oldenkamp and P.J. Paymans

The impact of automation upon warship design - the "Total System" concept.

C.G.W. Marsh and A.J. Stafford.

A system for ship handling in rough weather. K. Lindemann and N. Nordenström.

Multi-purpose marine simulator system and its applications. _1-69 S. Okana, S. Nagata and H. Matsunobu.

Bridge control by computer. Esteve and F. Bouthelier.

SESSION B:

Chairman: A. Chaikin

Research and development program manager Naval Sea Systems Command

Washington

Proving a propulsion control system for a modern warship.

P. Chadwick.

Dynamics of naval propulsion gas turbine. _1-113 Venturini.

The development of a ship control system from the initial see Vol.6 concept through to the final ship sea trials.

P. Mason, G.B. Coventry, A.M. Dorrian.

SESSION C:

Chairman: H.R. van Nauta Lemke

Professor in control engineering, Department of Electrical Engineering, Delft University of Technology.

The accuracy requirements of automatic path guidance. W.H.P. Canner.

Experiment with a new adaptive autopilot intended for controlled turns as well as for straight course steering.

J. Oldenburg.

Experience of ship steering automation in the USSR merchant marine.

A.A. Yakushenkov.

An application of Kalman filter to the discrete time route tracking of ships.

1-141 1-152 17167 17170/184 Page 1-5 1-28 1-45 1-83 1-98

INTRODUCTION

BY

WALTER J. BLUMBERG

Automation and Control Division

David W. Taylor Naval Ship Research and Development _Center Annapolis Laboratory

Annapolis, Maryland, U.S.A.

It is a great honor and privilege for me to make the introductory remarks for the Fourth Ship Control Systems _{Symposium and to provide a brief}

summary of the four symposia to date.

These symposia, originated and sponsored by what was then the Marine Engineering Laboratory, have

provided an international forum for the presentation and exchange of

scientific information in a field which has grown increasingly important to the naval community.

The international aspects of these

symposia are evidenced in this fourth symposium by the contributions of authors _{from 11 countries; by the location of} the symposium in the international city of _{The Hague; and by the hosting and} organizing by the Royal Netherlands

Naval College. When the concept of the first ship control

systems symposium was first dis-cussed with Naval representatives of

Canada and the United Kingdom in Annapolis in 1964 and 1965, we were entering

an era of more advanced ships with higher speeds and maneuverability; and the need for increased study, development, _and application of control technology,

from a systems point of view, was apparent. Since World War II, development of

sensors and control instrumentation had been gradual, and mostly found in

connection with specific items of hardware. By the mid

60's

_{however, applications of limited automatic}

control began to appear on more and more ships (even though

autopilots had been known and used since the time of Minorsky in 1922)

-- and gas turbines began to be used along with steam and diesel engines

as main propulsion systems for large _{ships. In} 1966, the introduction of the _{EO Classification by Det Norske}

Ventas, permitt-ing 24-hour operation solely

on instrumentation -- without manual _{attendance in} the engine room, _{-- gave further impetus to the manufacture of}

more reliable equipment and the introduction of

increased automation and automatic _control. Ship control has been a concern from the time of the _{very earliest of boats} and ships. _{When we began our work on}

integrated ship control in the early _1960's (as reported in the First

Ship Control Systems Symposium in _{1966), we were ever} reminded of the quotation

attributed to the Athenian

scientific historian (and fleet commander) Thucydides

-- who, about 425 BC,said:"A collision _{at sea can} ruin your entire day." _{Now, 2400 years later, the}

same problems, more compli-cated, are still with _{us; but we finally have the tools}

to cope with them. It is particularly appropriate

to note the increasing scope of ship control

systems. In the past, the ship control

systems designer could feel relatively secure in a field which changed

very slowly over time. _{Ship speeds, sizes, and} accelerations were relatively stable and slow to change.

But today, there is an

increase in the types of new vessels. _Their charac-teristics and dynamics are far _{different and involve scientific}

disciplines that were not traditionally associated

with ship control systems _designs. It is

appropriate, therefore, that in this fourth symposium, a number of the sessions and panels be devoted to these new disciplines; to unconventional ships; and to new types of control systems. This continues a practice started in the first

symposium -- of discussing the problems as well as the successes of new technology and applications.

The first symposium, held in Annapolis in 1966, included 400-500 delegates from the three invited countries of Canada, the United Kingdom and the United

States. It had as its theme "Ship Control Systems" emphasizing the integrated systems concept, and the influence of missions on the analysis and synthesis of ship control systems. It brought together, for the first time, those technical personnel who were engaged in the development of ship control systems to meet these broad new challenges, and those ship builders and operators who had to live with the new systems. In total, thirty-three landmark papers were presented. Covered subject areas included automation, human factors and simulation.

The symposium was notable for contributions such as that of C. M. B. Read on gas turbines and controllable pitch propellers, especially his prediction of dynamic performance at the design stage; of R. E. F. Lewis on sea trials of direct bridge control of engine and helm; of W. W. Rosenberry on cost-effectiveness for a ship control system; and of others.

The first symposium can be summed up in the words of the Keynote Speaker, RADM H. G. Bowen, USN, then Deputy Chief of Naval Operations (Development), who put ship control in the broad overall setting of total ship design. He empha-sized the need for greater reliability, maintainability, fault detection, cost-effectiveness, and automation -- and concluded by saying "we are on the eve of great developments".

The success of the first symposium, and the accelerated developments in ship control that followed, provided great confidence in planning for the next

symposium.

The second symposium -- which was also attended by nearly 500 delegates from the three invited countries of Canada, the United Kingdom, and the United States, -- was held by the Naval Ship Research and Development Center at Annap-olis in 1969. It had as its theme "Propulsion and Maneuvering Control of Sur-face Ships", with emphasis on ship control automation. In addition to three panel sessions, thirty-one papers were presented.

Special highlights of the symposium were the lively panel discussions on. Ship Automation; Control Aspects of Hydrofoils; and Control Aspects of Surface Effect Ships. The results of full scale tests, theory, and simulation were presented by the panels and discussed with the attendees. Of particular interest were papers on gas turbine control systems and dynamics as covered by researchers such as A. Sunley on the DDH 280 control system; F. W. Johnson and W. H. Wiley on control of gas turbines and diesel engines with controllable pitch propellers; and C. J. Rubis on braking and reversing ship dynamics. Also, several important papers were presented on ship automation, maneuvering, and man-machine relation-ships among which were those by T. S. Mather on radar ship separation measurement; D. W. Norman on jet flaps and controls; K. J. Mitchell and D. J. Strong on replen-ishment at sea; T. D. Mara, R. C. Cooper and C. G. Kurz on advanced conning system for operator control; P. R. Wyke on advanced automatic pilot; and G. B. Rozran on manning/automation study. In addition, two other papers of special note were Y. K. Ameen and H. A. Barker on simulation of random sea waves using pseudo-random signals; and D. W. Baker and C. L. Patterson on a new method of represent-ing propeller characteristics. These are only a sampling of the technology of ship control covered by the second symposium.

The first symposium provided the stimulus and interest for further ship control work and the second symposium explored, in more detail, some of the

specific areas of importance. The third symposium was held in 1972 in the United Kingdom. It was most ably organized and hosted by the Ship Department, Ministry of Defense, by Captain G. A. Thwaites RN, J. B. Spencer, and D. J. Strong, and held in the city of Bath at the University of Bath.

The third symposium was expanded considerably, with representatives being invited from more than 12 countries. Attendance was again excellent, with over seventy papers presented at the technical sessions. The interest and needs of the ship control community had grown rapidly in three years, and, as you can see, resulted in double the number of papers presented in the past.

The theme of the third symposium was "The Applications of Control Engineer-ing to Surface Ships and their Systems". Papers presented showed the continuing interest in control of gas turbines and areas of automatic control. Presentations on unmanned engine spaces were given -- with special attention to monitoring and maintenance systems. Also, a number of papers were presented on autopilots,

related automatic maneuvering and positioning systems,and on precise navigation for survey and other types of vessels. For the first time, superconducting pro-pulsion systems were discussed. The design and arrangements of the Netherlands' guided weapons frigate was of particular interest. There were also some papers on prediction and simulation studies for use in developing control systems. The third symposium can be summarized as an excellent compilation of design and appli-cation results which were useful baselines for future ship control designs.

We now come to the fourth ship control systems symposium being held here in the city of The Hague. This promises to be a stimulating program with over eighty papers to be presented covering a broad range of subjects. The theme of this fourth symposium is "Research in and Application of Control Engineering for Sur-face Ships and their Systems".

During the past few years, ship control designers have been asked to solve problems -- which although not all new -- have grown in importance. The papers in this Symposium, as you will see, go deep in reflecting current trends in ship control designs. In some of the symposium sessions, in-depth discussions will explore significant problems and results among specialists in each field. _In summary, the fourth symposium copes with these problems from the point of view of both theory and application.

Throughout these symposia it is of considerable interest to note that the line of research undertaken by one laboratory is sometimes paralleled by similar research in another laboratory thousands of miles away. In this instance both parties can benefit from an exchange of information which can be obtained _{at these} meetings.

That the subject of ship control systems has become _{more important can be} seen from the increase in published articles and reports in both technical and general publications, and recent presentations at meetings of _{technical societies.}

From this review of the first four symposia it _{appears that the research and} development reported to date have resulted in significant _{applications to ship} control problems. This is as it should be. All of us here have the opportunity to take advantage of the ideas presented and the personal _{contacts made. One of} the goals of these symposia is that these ideas will be put to use in future endeavors.

It is gratifying that this symposium is being hosted and organized by the Royal Netherlands Naval College. I would like to especially thank Commodore J. G. C. Van de Linde, Head of the Royal Netherlands Naval College for his

enthusiastic encouragement and approval throughout the planning of this symposium. And to Commander A. C. Pijcke and Lt. W. Verhage -- who had the primary manage-ment responsibility for the symposium, and who worked tirelessly carrying out

all of the myriad of detailed organizational tasks and day to day problems re-quired to make this symposium happen today -- go our boundless appreciation and admiration. I know, from working on previous symposia, the tremendous task it

is to organize such a meeting, involving scheduling, logistics, selecting and getting authors papers in on time, preparing and sending invitations, arranging a fine social program, as well as the many other tasks that had to be done. All indications from these superb arrangements are that they have been extremely successful. We are very appreciative of their efforts and are proud to share in their final product.

On behalf of the David W. Taylor Naval Ship Research and Development Center and the United States Navy, I extend best wishes for a successful fourth symposium

in this international city of The Hague and hope that this will be followed by many more.

CONNING A SHIP WITH DIFFERENT

STEERING SYSTEMS

BY

P.J. PAYMANS AND

I. OLDENKAMP

(NETHERLANDS SHIP MODEL BASIN)

SUMMARY

On a real time ship manoeuvring simulator

_{an experiment has been}

performed in order to investigate four alternative steering

_systems,

including the conventional rudder control, a rudder control with

feed

back from the rate of turn, a rate control and

_{a conventional rudder}

control combined with information about the future path.

Four experienced pilots alternately steered a simulated 450,000

DWT

tanker along

_{pre-seth path: Resulte showed a superior performence}

of the control system with a rate-of-turn feedback.

_{An additional}

seriesof runs on the simulator revealed some interesting data on the

human skills involved in the controlling of large

_ships.

I. INTRODUCTION

A long time ago it was recognized already, that a remarkable increase

of the controllability of a ship could result

_{from the availability}

of a movable part on the rear end of the vessel. Moving that part

-known as the rudder - enables the mariner

_{to make the Chip turn in}

a desired direction and to maintain a desired heading. The

_physical

forces to be applied by the mariner in order to move the rudder,

gradually diminished by the installation of steering devices, from

early tiller to modern elektremechanical

_{and hydraulic systems.}

No doubt that from a human point of view

this is

a beneficial

While the output side of the control system was continuously

improved, much less attention was given to the input side of the

system. In almost all cases the input of the ship control systems

is

in the form of

a desired rudder angle, basically reflecting the

historical way of steering: direct manipulation of the rudder by a

tiller. The only feed-back for the mariner resulte froM the

rudder angle indicator and the observed changes of heading (compasses

and outside view) of the ship. This way of steering is presented

schematically in Fig.

I. desird manoeuvre DISPLAYS percieved inputs, outputS and errors HUMAN OPERATOR CONTROLS motion feedback desired rudder angle

FIG.1 Block diagram of the convntional strIng of a ship.

This figure shows that the system can be considered a closed loop

system, with the elements: the mariner (human operator)

,

the system

to be steered, in this case the ship, via steering engine and rudder

unit, and the feed-back of the actual manoeuvre to the bridge

displays.

It can be questioned, however, to which extent the conventional way

of steering ships is the best way, considering for example the

increasing ship's sizes and speeds. This and related problems have

been recognized and investigated (1),

(2), (3) severa.1 times.

Especially important with regard to the steering system, three factors

actual

STEERMG rudder actual

ENGINE angla _{SHIP manoeuvre}

AND RUDDER

are to be mentioned:

1.

large ships have a long delay (in the order

of 1

minute) between a rudder call and a

resulting observable change of heading,

due to the lags of the steering machine

and the large mass of the ship.

large ships may have directional instability,

causing a range of rudder angles (around the

midship's position) for which the rate of

turn is not defined.

The direct concern with the actual rudder

deflection is a rather complicated manner

of control. One is hardly aware of this fact,

because man is

so acquainted with the

conventional steering of ships. Yet a change

of heading at the very least requires three

rudder calls: one to start the ship turning,

one to counteract the turning and one to lay

the rudder midships again. For comparison: a

car needs only two steering actions for

changing course.

Thinking of possible solutions for the problems mentioned above, some

can be deduced easily, at least, theoretically.

As

well as

improve-ments are made in the steering engine itself, improveimprove-ments can be made

on the input and output sides of the control system (4). The first

category requires some technical solution for reduction of the

time-lags and slowness of the steering engine. On the output side also

mainly technical solutions come up for improvement of the

rudder-effectiveness, for example a larger rudder. However, on the input

side of the control system one finds the human operator. Because of

the very special nature of this human operator, this opens the

possibilities to search for solutions in the information processing

and control facilities on the bridge of the ship.

In the study described in this paper it

is tried to compare some

control facilities, that theoretically might be improvements.

More or less arbitrarily chosen, besides the conventional rudder

control, three alternative steering systems were tested:

I. A "rate-aided" steering system.

In this type of steering system the mariner chooses

a desired rudder angle as he normally does. A control

system calculates the rate of turn according to the

desired rudder angle and adjusts the actual rudder in

such a position that the ship indeed turns as desired.

One can imagine, that the steering task in this case

is simplified at least with respect to the instability

of the ship's behaviour.

2

A "rate control" steering system.

With this system the mariner steers the ship's rate of

turn directly by turning a wheel on the bridge. A control

system chooses the rudder angles necessary for giving

the ship the rate of turn wanted. It will be clear that

this system, although uncommon, requires less actions

for changing course compared to the conventional use of

the rudder. Starting a turn requires only one control action:

turning the wheel to the desired rate of turn.

Stopping the turn after completion of the desired change

of heading, can be done with one action also: turning the

wheel back in the zero turning rate position.

3

A device for better anticipation of te ship's path.

This third method for steering the ship was designed to

improve the mariner's information about future movements

of the ship. Superimposed on the bridge radar two spots

were shown, that indicated the ship's position 2.5 and

5 inutes respectively in the future,

if no further

actions were taken. This information was hypothesized to

reduce the mariner's uncertainty caused by the inertia of

large ships and the time lags in the effectuation of orders.

Fig.

2 summarizes the principal differences between the systems

(ef Fig. ; also).

The comparison of the steering systems was done by the simulation of

both the ship's behaviour and the steering system;

characteristics

on a ship manoeuvring simulator. It has to be emphasized that aspects

related to the practical implications of the systems studied were

not considered and that the choice of the systems was arbitrarily

desired _manoeuvr ro r desired

error

MarlOeuVre desired er manoeuvre AN

a. conventional steering system.

MAN

desired rate

_{of turn}

0_

desired rudder

angle

motion feedback

corrective _{rudder angle}

motion feedback

b. " rate - aided control S' steering system.

RATE

CONTROL SYSTEM

motion feedback

C. "rate control" steerin9 systen,

1

information future Path

PREDICTIVE SYSTEM motion feedback I - 9 CONSTANT FACTOR

rate of turn

SHIP

POsition and

hading

d. Steering system with information about futur path.

FIG. 2

Simplified block diagrams of th four systms studid.

rate

of

turn

actual

manoeuvre actual manoeuvre actual manoeuvre actual manoeuvre

rudder _{2.912L._ SHIP}

dsired rudder angle

HI desired actual rudder rudder desired ,-,t./error MAN angle angle SHIP manoeuvre

-to some extent, but also suggested by considerations from control

and system theories.

After completion of the investigation with the different steering

systems, some additional runs on the ship manoeuvring simulator

were performed in order to get insight in the way different mariners

understand the behaviour of very large ships. Section 3

of this

paper deals with this experiment in more detail.

2. EXPERIMENT

I

2.1 Apparatus

This part

I

of the study consisted of 64 runs on a ship manoeuvring

simulator which includes a ship's bridge with the instruments,

usually found on board modern ships, such as compasses, rudder and

engine indicators, and a radar display. Fig.

3

shows the interior

of the bridge.

It is situated in the middle of a cylindrical screen with a diameter

of 20 metres, on which the outside view is projected (Fig. 4).

Fig.

4 Crossectional view of the ship control simulator.

The ship's

characteristics as well as the influences of wind,

currnt and for example bank suction are programmed am a hybrid

computer so that the ship's reactions on for instance a rudder call

are the same as the reactions of the real ship in a corresponding

situation (5).

The pictural block diagram of Fig. 5 explains the way the simulator

works as a closed loop system with the human operator as

_{a part of}

Fig.

5 Pictural block diagram of the simulation process.

2.2 Experimental conditions

For this study the simulator represented a fully loaded oil tanker

of 450,000 DWT with a length of 346 metres, a beam of 85.5 metres

and a draft of 23 metres. The ship was sailing

7

kn. in open sea

without wind and current and under zero visibility conditions. The

subjects had to steer the ship along a

line that was displayed on

the radar screen, which, of course,

is not possible in reality.

This line represented the track that the ship in this experiment

would follow if three rudder angles were applied, each of a

pre-determined number of degrees and at pre-pre-determined time intervals.

The first rudder angle was 7.5 degrees to starboard, applied

after 56 seconds, the second was 15 degrees to port after either

midships again, was applied at the moment of zero rate of turn

after the change of heading.

This sequence of three rudder movements generated a change of

heading of either 30.5 degrees or 61.3 degrees to the right. The

path of the ship during this manoeuvre, represented by a line on

the radar display, could be followed exactly by well steering

subjects, although theoretically, as the subjects did not know,

how the line was constructed actually. Fig. 6 shows the radar

picture with the line to be followed.

Fig.

6 The radar display with the line to be followed

(the change of heading of 61.3 deg. is shown).

In order to manoeuvre the ship the subjects were on the bridge of

the simulator with the availability of the radar display, the

tachometer, the compass and the steering device only. The last one

had a form, shown in Fig. 7. Dependent on the kind of system in

operation this wheel was controlling either the rudder angle or

the rate of turn, as such connected with rudder indicator respectively

rate of turn indicator in the bridge console.

Four experienced ex-pilots, aged between 55 and 61 years sailed

each of the four steering systems four times, making 16 runs per

subject.

Fig.

7 The steering device available to the subjects.

(see text).

The experimental design determined the conditions and the

order

of the manoeuvres based on the principle of repeated measures

on

the variables: steering system, change of heading and sequence

effects (6).

Registrations were made, as

a function of time, of the path of the

ship's centre of gravity and the use of the control

wheel. After

each run the subjects had to judge the steering system with regard

to its simplicity, accuracy and safety

in manoeuvring very large

ships.

2.3 Results

The results of the experiment show important

influences of the

steering systems on the behaviour of both man and

ship.Looking at

the path of the ship's centre of gravity during the manoeuvres it

has to be remembered from the former

section

that the deviations

from the track to be followed ("reference

line") can be considered

measures of the performance of the

different steering systems. An

analysis of variance (7) suggests a

significantly

(a -

.05) worse

the future track information) compared to the "rate aided" and

"rate control" steering systems.

Fig. 8 shows how that effect developed.

E 50

at

So Rudder control aided control path information

/'

....:--- -____

----

--- Rate

---- Rate

Predicted

--Change of heading 51.30 1 Curved parts of the reference

° Change of heading 30.5 line i 1000 2000 3000 Distance travelled (m)

FIG.8

The mean trajectories of the ship's centre of gravity as influenced by

the steering system.

For reasons of simplicity in that figure the curved part of the

reference line is stretched, while x-axis and y-axis have different

scales. The conventional rudder control actions prove to be not

early or drastic enough to follow the reference line as accurately

as with the other ystems.The spread of the individual runs around

A

the mean trajectory does not differ so much for the steering

systems. Indicated by the standard deviation only the system with

future track information shows a somewhat larger value (see Fig.

9).

The better performance of the "rate (-aided) control" steering

systems can be explained by their more effective choice of the

rudder angles. Fig.

10 presents the average rudder angles the

subjects asked for during the manoeuvres, as a function of the

distance travelled.

-50

C

m

50

20 -2 Rudder control Rate aided Rate control

Predicted path information

Change of heading 61.30

Curved parts of the reference line

Change of heading 30.5'

Rudder control

--- Rate aided

Rate control

Predicted path information

,. N. s

'

"SS ,

/

Change of eading 61.30 ..,

Curved parts of th reference line .1/4

\Chang ot headi g 30.5°

\

I

i

\

/

'I---,

1 t i

--

I o 1000 Distance travelled (m)

FIG.10 The average rudder angle as influenced by the available steering system

1000 2000 3000

Distance travelled (m)

FIG. 9

The standard deviation of the individual runs around the mean trajectories,

The essential difference between conventional and semi-automatic

rudder control proves to be the counter rudder movements for

diminishing the rate of turn of the ship when the heading had

changed enough. The estimation of thebest moment to lay the rudder

in the opposite direction in order to slow down the ship's turning

can be considered difficult, except for the "rate-aided'control"

steering system. With that system the subjects do not need to

worry about counter rudder actions:

a simple laying the rudder

midships generates rudder movements to reach the desired zero rate

of turn. A clear indication of the superiority of the "rate-aided

control" steering system is found in Fig. 11,

in which some measure

of the number of ship's oscillations is presented. Only this system

together with the conventional rudder, which is highly familiar to

the subjects, resulted in a very quiet ship's behaviour.

14 12 10 c 8

06

E4

o 2

Rudder Rate-aided Rata _Predicted path

FIG.11

The average number of times the rate Of turn _{changad sign.}

This very accurate and quiet ship's behaviour did not require

more effort of the mariner3.

On the contrary, the "rate-aided control"

_{steering system had the}

least number of control actions,

_{as presented in Fig.}

_{12. The}

subjects.

t, 60 50

5

40 an

Z30

o

E 70 20 10

Negative

judgement No decision

ei-e Simplicity of the steering

o

o Safety of the manoeuvre

a a

Precision of the track

FI0.13

The judgement of the subjects concerning the simplicity of

the steering system and the safety and accuracy

of the

manoeuvres in relation to the available steering

system.

Rudder Rate-aided Rate Predicted

path

FIG.12 The number of orders to the control wheel , dependent of the

steering system.

Additional information about the future track was not highly

appreciated by the subjects. Compared te the other

steering systems

it was cons id er d more complex, less accurate and less safe under

the conditions studied.

Fig. 13 summarizes these subjective ratings

of the systems by the

Positive

judgement Rudder Rate- aided Rate

Preditted

3.

p(PERIMENT II

A second series of simulator runs aimed at collecting some

fundamental data about the intuitive comprehensiveness of the ship's

behaviour. Again te ship to be

sailed wise

the fully loaded 450,000

DWT tanker, Manoeuvring in open sea, without wind, current and with

zero outside visibility. The starting speed was 7

knots. Also the

same manoeuvre had to be performed, although, this time, only the

change of heading

of

61.3 degrees to the right had to be performed.

The control facilities to steer the ship along the predetermined

reference line, as shown on the radar display, were restricted to

three orders only.

By pressing a pushbutton (see Fig.

7) preprogrammed rudder movements

were initiated. The first push caused a rudder angle of 7.5 degrees

to the right, the second a rudder angle of 15 degrees to the left

and the third order moved the rudder to the midship's pesition asain.

When the moments of pushing were chosen enactly on the right time,

that means after 56, 333 and 596 seconds respectively, the ship's

centre of gravity would have followed the reference line exactly.

E -10 s. t.-15 172 o ú 5 6 0

-5

o -5 Experienced SS Less experienced ss Inexperienced ss _ I N'

\

\

...7\

\

ur\

N..'

I

\

\

`,..

\

I I

r ,

,

I /-

/

,

I o 500 1000 Dtstance travelled frn)

FIG.14

The mean trajectories of the ship in relation to the experince

Three groups of four subjects were involved in the experiment.

The first group consisted of the same subjects - experienced

ex-pilots - as those involved in the first experiment. Four naval

architects were considered a less experienced group, at least in

manoeuvring skill, while four draftsmen were invited as representing

inexperienced mariners. Each subject performed the manoeuvre on the

ship simulator three times, after an ample explanation of their

task, completed with one or more training runs.

Registrations were made of the ship's trajectories and the errors,

made in the pressing of the pushbutton at the right moment.

The results of this experiment show, how different people have

different ideas about the behaviour of very large ships.Eome subjects

under-estimate the inertia of the ship, others in fact are too afraid

of the ship's inertia. Looking at the path of the ship's centre of

gravity during the first trial of the subjects the differences in

experience are very clear. Fig.

14 (upper part) shows how orily the

experienced group succeeded in following the reference line (drawn

as in a straight line for setter interpretation).

t.1 e .co o o 40 -

40-u

30-

s; 30 -a, -₀ o

20-FI0.15 The average timing errors ol the subjects , in relation to their

experience level. o '1+ E .E 20 -a o L s 10 E i-g,so 50

-But the other subjects learned from their first trial. This can be

concluded from their last trial, which is given in Fig. 14 (lower

part). The deviations now are smaller and more convergent to the

reference line.

The timing errors in pushing the button were surprisingly enough

-not very much influenced by the subject's experience level.

In Fig. 15 it

is seen, that the first rudder call was wrong by

about 20 seconds (average and irrespective of sign), while the

second call, implying 15 degrees counter rudder, was estimated with

a mean error of about half a minute.

A-52 n 400 o u 350 A To-15' o g'300 _ E 2 3 Trial 2 3 Trial

FIG 16 _{Timing errors of the experiencd ss f each column represents one run}

of ene subject )

The larger errors of the group with the experienced subjects are

caused mainly by one subject, who heavily under-estimated the ship's

inertia. In Fig.

_{16 this subject is referred to as "A".}

It has to be mentioned, that the moment of laying the rudder

midships again, i.e. the third push of the button, was chosen

erratically to such an amount that an appropriate interpretation

has to be Omitted. For this reason only results concerning the first

and the second action by the subjects are taken into

account.

A

C D

4. DISCUSSION

With regard to the first experiment, discussed in section 2,

it

can be concluded, that in the conditionm

tested, the behaviour of

a very large ship is highly

influenced by the kind of steering

system available to the mariner. Compared to

the conventional

rudder control, which means that the mariner asks for any

rudder

angle and via quartermaster and steering engine the

rudder is moved'

accordingly, the "rate-aided control" steering systems seems

to have

advantages. Without changing the asking for rudder

angles - a very

familiar control behaviour of mariners - the feed-back

of the

resulting rate of the ship compensated for instability

phenomena

and facilitated the reduction of the

turning speed after the change

of heading was completed.

In the second experiment it is just

demonstrated, how difficult

the timing is of the rigbt moment for

reducing the rate of turn of

the ship in order to prevent either

cutting the corner or overshoot.

(Fig.

15)-Besides, the timing errors tend to cumulate, as

shown in Fig.

17,

in which te relationship is given

between the errors of the first

call and those of the second one.

FIG.17 The relation between the timing error of the first rudder call (horizontal axis) and the timing error of the second

rudder call (vertical axis).

R= 75

.

Second call 100-1

80-too late (sec.)

60-.. 40-. . .20-, % , -60 -40 -20 20 40 60

First call First call too early (sec.)

.-30-

too late (sec.)

:

-401

The correlation coefficient of .75 suggests a conclusion, that if

a first decision has been taken late, the second decision very

often will also be late.

The superiority of "rate-aided" manoeuvring was indicated by a

better tracking of the reference line (Figs. 8 and 11) with a

minimum of control actions (Fig.

12). Also the subjective judgement

of the subjects was positive (Fig.

13). These results are compatible

with research data from the field of human tracking behaviour (8).

The performances with the "rate-control" steering system, in which

the subjects did not cope with rudder angles directly, are leas

impressive. Although the track was followed accurately (Fig. 8),

the number of control actions was higher and the rating of the

system by the subjects was somewhat less possitive. Some reasons

for this fact can be suggested, of which the more important ones

have to be found in the parameters of the control system itself.

The amplification between rate of turn asked for and rudder angles

chosen by the system proved to be too large and caused drastic

rudder movements and considerable time loss consequently. Therefore

a following experiment has to deal with a better adjustment of the

system parameters.

Another reason is the uncommon type of action on the bridge:

controlling the rate of turn

imstead of the rudder angle is quite

unusual for mariners. So it was observed that some subjects continued

to give "counter rudder" to check the turn, although rate control

makes such action superfluous. Moreover it is doubted if rate control

is the most suitable control system for the manoeuvring of ships (8).

The results of the manoeuvres with information about the future

track show a relatively lower level of performance. Compared with

the other systems the reference line was followed less accurately

(Figs.

8 and 9) with more oscillations (Fig.

11) and with less

confidence, as expressed by the subjects (Fig.

13). Definite

conclusions about the somewhat disappointing performance of the runs

made with this system have to be drawn very carefully. This is due

to several factors, of which two will be mentioned here:

I. Manoeuvring a ship with information about the future track

is

_{a very unfamiliar task, especially for experienced pilots.}

Their performance could have been negatively influenced by

the confusing type of information and the lack of opportunities

to get acquainted with such systems.

2. The way the future track was shown - by the superimposing

of two targets on the radar display, 150 and 300 seconds

ahead of the ship's bow - was chosen arbitrarily.

Other, and perhaps better ways for presentation can be

thought of easily, either on radar displays or independently.

As other systems (9), (10), (11), have demonstrated to

increase system performance to various degrees, further

study of "looking into the future" type displays should be

recommended.

As a final remark of the experiments described now, it can be

concluded, that the device with which the ship is steered, largely

influences the success of the manoeuvre. Although the conditions

were very restrietive - no wind, no current, no other traffic and

a rather elementary manoeuvre - the manoeuvring

of a very large ship

could benefit from a "rate-aided control" steering system. It

permitted in fact the mariner to ask for rudder angles in the

conventional way, and at the same time it helped prevent oscillations

and overshoots by controlling the counter rudder movements

automa-tically. The performance with the "rate control" steering system and

the future track information is not established well enough,

probably

to a large extent due to bad adjustment of the system

parameters and

the unfamiliar characteristics for the mariner

respectively.

The results of the second experiment, showing the errors

and the

large individual differences in the "timing" of the

control actions

in order to steer a very large ship along a desired

path, plead for

a very careful approach of problems

related to the control task

elements. Changes of that task, either by changing the

steering system

or by providing man displays,

including for example anti collision

systems, may so interfere with the

familiar ways of steering, that

the intended improvement is not realized, or even

worse: a

deterioration of the performance could occur.

Apart from the results of the studies

discussed here, these

experiments show

that by integral simulation of the

controlled ship

system, various problems related to

the human factors in ship

manoeuvring can be studied with relatively

little effort in cost

and time. Although, as yet,

it

is not common practice, like in other

transportation fields, to simulate technological

advance in ship

building and ship operation with research data

gathered by simulation

objective and accurate information ship simulation can give about

the performance of the ship control systems, the human operators

included.

REFERENCES

J.F. Hooft

The manoeuvrability of ships as influenced by environment

and human behaviour"

Journal of Navigation, London 1974, P. 367

J. van Dixhoorn, L.A. Koel6 and J.P. Hooft

"Feasibility and profit of navigation information and

navigational aids offshore"

Proceedings 23rd International Navigation Congress,

Ottawa 1973.

Symposium on"Ship Handling", Wageningen 1973.

Netherlands Ship Model Basin, Publication no. 451,

Wageningen

The Netherlands.

J. van Amerongen and N.D.L. Verstoep

"Improvements of ship manoeuvrability by means of automation"

Symposium on "Ship Handling", Netherlands Ship Model

Basin

Publication no. 451, Wageningen, The Netherlands.

G. Mak and G.M. van der Bend

"Hybrid computation for the simulation of the

manoeuvring

of ships"

Tenth Symposium on Naval Hydrodynamics,

Boston, 1974.

B.J. Winer

"Statistical Principles in Experimental

Design"

New York, 1962, 2nd ed.

W.L. Hays

"Statistics for psychologists"

New York,

1963.

E.C. Poulton

"Tracking Behaviour"

In E.A. Bilodeau "Acquisition

of Skill"

New York, 1966.

Bernotat, Dey and Widlok

"Die Voranzeigen als antropotechnisches

Hilfsmittel bei der

Führung vom Fahrzeugen"

Forschungsberichte des Landes Nordrhein

Westfalen, no. 1895

C.H.R. Kelley

"Predictor Instruments look into the future"

Control Engineering 3, March 1962.

W.A. Wagenaar, et al.

"Auxiliary equipment as a compensation for the effect of

course instability on the performance of helmsman"

Report no. 1971-C11, Institute for Perception,

Soesterberg, The Netherlands.

THE IMPACT OF AUTOMATION UPON WARSHIP DESIGN - THE "TOTAL SYSTEM" CONCEPT

BY

CAPTAIN G G W MARSH OBE., MA., RN COMMANDER A J STAFFORD MIEL, MIMechE., RN

SUMARY

The purpose of this paper is to discuss the possible effects of future trends in control and surveillance technology upon warship design. During the past ten years the commercial marine has seen a substantial increase in the level of automation of ship control and allied functions, and the reasons for this are examined in the warship context with particular reference to the important factor of crew size. Although the paper is written primarily from the point of view of machinery control and surveillance, it nevertheless

identifies the need for the whole ship to beconsidered as a "total system" as a conceptual approach to the problem of control and surveillance design. The paper concludes that the development of warship design now stands at a critically important crossroads, and that it is necessary to make a fundamental re-appraisal of the role of the warship and the part to be played by the crew in order to determine which technological path to follow. Eather way,

decieions taken during the next few years could have a far-reaching effect, not only upon warehip design but also upon the very nature

of naval service.

It is emphasised that this paper represents the personal views of the authors, and does not necessarily reflect either the opinion or the policy of the Ministry of Iefence (United Kingdom) or the Unip Department.

2. INTRODUCTION

During the past decade, the automation of ship control and machinery surveillance functions has become a generally accepted practice in the commercial marine industry, and has been recognised by both ship owners and classification societies as being the cost-effective way to operate merchant shipping. The influences which have brought

about this change have been economic: as the costs of sea-borne man-power have risen, and as difficulties of recruitment have increased, methods have been devised of achieving compensating reductions in crew numbers. Not surprisingly therefore, the introduction of automation in the commercial marine field has concentrated initially in providing remota control and aurveillance of ship's machinery at the one position which, under present maritime law, always has to be manned - the Bridge. Thus Bridge Control was an essential feature of the technology which made the Unattended Machinery Space

fums)

concept feasible, and this has resulted not only in a significant reduction in crew size, but also in the release of men who would otherwise be watchkeeping, to the more "productive,. taaks of main-tenance and ship husbandry.

automated merchant ships has probably been carried to its practical limit in modern control and surveillance system designs. It would seem that the only prospects of achieving farther reductions in manpower lie either in improving reliability and thus reducing the need for maintenance on passage, or by attacking such fundamental concepts as the need for bridge watch-keeping. The Partially Unmanned Bridge, or PUB concept, is a move in the latter direction which is gaining some support. In summary, the commercial marine scene is one in which great conceptual changes have been made during the past ten years. The technological and commercial viability of automated ships has been established, but the long.-term effects - particularly in terms of the effects of automation upon the social attitudes, motivation, and sense of responsibility among crew members - are only just beginning to emerge.

In contrast, and because of the lack of commercial pressures, the introduction of automation into the ship control and machinery supervision areas of warship design, has been more hesitant. Some moves have been made in the direction of Bridge Control, following commercial marine practice, but without any detailed assessment of the objectives, and some forms of remote operation and surveil-lance of machinery have been introduced to meet specific NBCD requirements for operation under closed-down conditions. However, no clear policy has emerged, though the pressure from industry to adopt increasingly sophisticated automated systems is becoming greater every year as controls technology advances. The aim of this paper therefore, is to consider the impact of further automation of ship control and machinery management functions as it is likely to be reflected in the ability of a warship to meet its operational requirements, and in its probable effects upon the size, motivation, and training needs of the crew. In pursuing this aim it is necessary first, to examine some of the possible benefits which may accrue from automation, and to consider their relevance to the warship case.

3.

THE NEED FOR AUTOMATION

The function of a merchant ship is to arrive at the port of destination safely, after an uneventful and economic voyage. At sea, the objective is to settle down to a stress-free, steady-state condition as soon as possible after leaving harbour, and to maintain that condition for as long periods as navigational conditions will permit. The function of the warship on the other hand, is to go to sea and stay at sea for the duration of a specified mission, under conditions which may be anything but uneventful. Once at sea, the objective is usually to avoid steady-state conditions as far as possible, except on passage - and even then the opportunity is usually taken to impose

a measure of operational stress on both crew and machinery by the exorcise of manoeuvres, drills and emergenciee in order to perfect the ship's response to any situation that may develop. In view of this fundamental difference in function, it cannot be surprising that the designs, methods, sud attitudes of engineers working in the naval and commercial marine fields have traditionally been reflected in two quite distinct philosophies. Thus although the moti-vation to design and implement automated ship control and machinery smrveil-lance systems in merchant ships sprang from the economic need to reduce sea-going manpower, it cannot be assumed that by following the same practices in the warship application, it will be possible to achieve similar savings.

]ring recent years, however, the essential difference between what is acceptable for a warship and what is acceptable for a merchant ship, has been obscured by economic pressures to adopt a commonality of concepts, designs and standards. Thus automation of warship machinery control and surveillance systems is already following the trends set in the merchant marine, but without any in-depth studies of the full implications of such trends in terms

urgent need to return to first principles, and to design a technical strategy for the future development and implementation of automated systems in the warship application, based equally upon trends in control and aarveil-lance technology and upon the operational requirements of the warships themselves.

The automatian of a given function is usually justified under a combination of five headings:

That a hazardous operating environment, limitations of space or other physical constraints, exclude the possibility that the task could be performed,by a man.

That, awing to the nature of the task, it can be performed more effectively by automation.

That the use of automation offers an overall economic benefit.

That the use of automation releases manpower for more important, or more rewarding tasks.

That the use of automation reduces the technical demands upon the operator, and lowers his respansibility to the point where a less-qualified grade of labaar may be used.

It is necessary to examine each of these factors in the light of warship operating experience.

3.1 Environmental and Physical Constraints

The possibility of contamination of ships' machinery spaces under nuclear fallout conditions or under chemical warfare attack led to the require-ment for occasional unmanned remote-control of these areas and the

implementation of this requirement has necessitated a measure of

automation of some control and surveillance functions. Thie requirement is evidently a continuing one but it seems likely that any new constraints of this type in the future will be in the "desirable. rather than

"essential. categroy and will be related to the working environment -noise, temperature, humidity, etc. Thus in gas-turbine and high-speed diesel ships there is an added incentive towards remote operation and

control.

One manpower-intensive task which often has to be performed in the face of a severe environmental hazard, is that of fire-fighting. On this argument alone there would seem to be a strong case for a greatly increased level of automation of fire detection and fire-fighting facilities in a warship, and it will be seen later that this case becomes even more important if a serious attempt to reduce manpower levels is to be made.

3.2 Improved Effectiveness

This in of course, the major argument which has led in recent years to the extensive automation of warship weapon systems. Here, the urgent need to reduce reaction times as a counter to high-performance aircraft

and missiles, coupled with the ever-increasing complexity of the tactical picture, has led to the development of advanced-technology data-handling and fire-control systems in which the human operator is reduced to a monitor/veto function. In the machinery controls field, the case for automation judged against this criterion is much less marked. There is

very little evidence that human operators have any serious shortcomings when controlling ships, machinery systems, and the time-constants involved even in emergency manoeuvres, or in changing from one machinery state to another, are long enough - even in gas-turbine ships - to present no difficulty to trained naval personnel.

Although there are no new factors in the operation of ships, machinery systems therefore, which necessitate a general increase in the level of automation, there are none the less a number of areas where greater automation might be expected to result in improved effectiveness, and these must be included in the overall strategy. They include:

Surveillance The human operator is notoriously ineffective at monitoring the steady-state performance of machinery because his attention wanders unless his interest is held by a changing sequence of events. Automatic alarm and warning systems, designed to alert the operator if potentially hazardous thresholds are exceeded, are therefore likely to be highly effective.

Protection Human operators are subject to errors of judgment and drill, and the risk of these errors increases as the transient response of machinery becomes faster as the inevitable concomitant of high performance specifications and high power/weight ratios. A well-engineered automatic control system on the other hand is vulnerable only to component failure, and externally-inflioted damage: it therefore offers a potentially superior performance

in the protection of machinery systems against specified and predictable events, but with the penalty of lacking any versatility to deal with the unexpected.

Ebonomic Management Automated control systems can be optimised to take account of complex parametric interactions and drifts which lie well beyond the scope of even a well-trained operator. Warships are not noted for their economoy in operation, and some potential may be assumed to exist for computer-aided systems to manage the consumption of energy and other resources to economic advantage - possibly to the extent of allowing an

increase in mission time.

Health Monitoring Most maintenance systems in use in war-ships today invoke an "Upkeep by Exchange" policy or a variant

of it. In such systems the replacement of major machinery is nearly always decided on a time-related basis - the periodicity being determined by service experience with similar equipments, or by theoretical failure-rate predictions modified by a suitable

safety factor. It is evident that aay system which enables the necessity for replacement to be determined by failure predictions based upon measured wear and parametric trends, will achieve significant economies both in monetary terma and in the critical operational factor of ship availability.

Health monitoring and trend analysis systems are already avai/able for specialised applications, and a significant growth in their warship application can be expected during the next decade.

3.3

The Economic Case

It has already been established that the strongest motivating influence in the automation of ship control and machinery surveillance _{in the}

mercantile marine, has been the need to economise in seagoing manpower. In the naval context the case for reducing ship complements is super-ficially at least even more compelling: not only is there a potential saving in direct costs, but also a saving in weight and space which can then be used to improve the weapon fit or enhance some other operational feature of the design. In monetary terms, the true cost of the service-man afloat is difficult to assess realistically because it is a function of a number of complex and

interacting

factors. These include:

Prime Costs Shipbuilding costs of accommodation, domestic and recreational facilities, plus a proportional cost of the increased size of the ship necessary to make these facilities available.

Career Costs Total pay and pension attributable to naval service, divided by total seatime.

Afloat Support Proportional cost of food, cooking, fresh water, heating, cooling,

ventilation.

Proportional cost of

administration and medical services.

Ashore Support Proportional costs of Training, Welfare, Administration etc.

A very rough estimate suggests that prime costs (a. above) for a junior rating in the Royal Navy isabout £10-15K per man per ship, and that career costs (b.) exceed MOE per man per year. Taking into account these, other potential savings under c. and d., and the general saving in weight and space, it is clear that a reduction in complement of one man at the design stage integrates into a substantial economic benefit when taken over the whole life of the ship. The conclusion must be that automation, where it genuinely replaces a man in the ship's complement at the design stage, will be a highly costeffective

invest-ment.

3.4 More Effective Use of Manpower

The use of automated datahandling systems for tracksorting, target identification, and the routine processes involved in compiling the tactical picture, has achieved significant reduction in the number of men required in the Operations Room (Cc), and at the same time has freed the Command Team for the more important, stimulating and

anthro-pomorphic tasks of threat evaluation, tactical decisionmaking and combat control. In the machinery control field on the other hand, there is a serious danger that automation of machinery management will remove the primary source of interest and motivation from the sphere of

responsibility of the more qualified and experienced engineers. In

this sense machinery automation may prove counterproductive in that it frees manpower only for the more humdrum chores of routine maintenance and ship husbandry: it will require a serious fault to add the spice of professional interest to an otherwise insipid existence.

Experience in automated merchant ships suggests that this may already be a problem. There is some evidence that a lack of specific responsibilities

such as watchkeeping coupled with an increase in leisure time and spending money, has contributed in recent years, to a significant increase at sea of drunkeness and other social problems.'

Advanced technology control and surveillance equipment is designed to make very few demands upon the technical qualities of the operator and maintainer when it is functioning correctly, but paradoxically, it often imposes a much greater strain upon technical knowledge and diagnostic capability when it does develop a fault - especially, if that fault is autside the scope of the built-in diagnostic aids. The

dilemma is especially acute in the case of a warship which has to be self-supporting and which has to respond swiftly to system failure or event action damage, under combat conditions.

The implication is, therefore, that if the dhip is to be capable of some measure of self-support - particularly in recovery from the effects of shock and minor action damge - then it is necessary to retain onboard the highest level of technical exTertise available. This is certainly borne out by experience in the weapon system field where a high degree

of automation is often accompanied by extreme difficulty in finding suitable employment for junior maintenance ratings. Thus, in the warship application, automation may ¡ncrease rather than reduce the need for highly-skilled personnel. It therefore follows that less-qualified labour can only be used as a substitute

if:-the requirement for self-support is reduced, and/or

greater dependence is placed on automatic reversionary modes (je, system redundancy).

The implications of these two conclusions are discussed later in the

paper.

In summary, this discussion has demonstrated that although there are a number of areas where the automation of machinery control and surveillance functions offers potential advantages to the warship designer in terms of operational effectiveness, none of these can be placed in the "essential" category, and some at least may have downstream effects upon manpower utilisation which may be less than desirable. The authors believe that

although a further and significant increase in the complexity of machinery control and surveillance systems may seem technologically appropriate, and superficially attractive to subjective judgement, it is unlikely to be cost-effective in ship terms unless it is accompanied by a compeneating

reduction in crew size. It is now necessary to consider in some detail, the factors influencing warahip complemente.

4.

WARSHIP MANNING CONSTRAINTS

The manpower requirement for an operational warship is determined by two task

components:-a. Scheduled Teaks These are the 'routine' or predictable teaks associated with the operational control of the ahip, its weapons, its machinery and its men. Typical examples of teaks in this category are:

Command

Ship Control (00W) and Safety

Weapon Control

Rounds and Patrols

Routine Maintenance

Routine Administration

Food Preparation, etc

b. Unscheduled Tasks These are the intermittent and unpredictable tasks that are a function of the operational use and abuse to which the ship has been subjected. The frequency with which theyOCCUTis often a measure of the cumulation stress on machinery and men. These

include:-Provision of landing and boarding pasties

Fire-fighting

Repair of damage

Fault diagnosis and rectification

Operation of manually-controlled reversionary modes

Use of sea-boats

Replenishment

Accidents,

personnel

emergencies etc.

In general it may be said that it is the performance of scheduled tasks that determines a warships operational effectiveness, whereas it is the performance of unscheduled tasks that determines the ships, ability to sustain that level of effectiveness throughout the mission. It may also be observed that whereas the scheduled tasks provide the main motivation for the crew, they are also the

tasks which are the more easily automated since the tasks themeelves, and the circumstances relating to them, can be more easily defined. In the past, the scheduled tasks have always proved to be the dominant factor in warship complementing, and this has provided a pool of reserve manpower in the Ship (watchkeepers off watch, maintenance daymen, cooks and stewards, etc) who could always be available to meet any emergency or unscheduled activity. Unfortunately this comfortable position has been steadily eroded during the past ten years, at first by the progressive automation of weapon system functions, and more recently by thm introduction of low-manpower propulsion systems (gas turbine and diesel), following the general retreat from steam.

The point has now been reached however, where the unscheduled taak load has become the dominating factor in determining a warships complement. An illustration of this has recently been provided by the issue of an instruction - following an incident in a Royal Navy ship - setting the minimumnumber of men required to be on board a conventional frigate at any time for

fire-fighting duties, as between 30 and 40. If this is to be the minimum manpower force to be available at all times under either operational conditions at sea or whilst giving shore leave in harbour, then in practical terms, it sets the minimum total complement for a 2,000 - 3,000 tonne frigate at around 150 men. Similar arguments can be advanced for other unscheduled tasks such as Damage Control, landing parties, major cleaning and painting exercises, etc. Support for the figure of about 150 men as a representative minimum for a modern frigate built to a conventional Operational Requirement, and incorporating

state-of-the-art control and surveillance technology, has also been provided by recent operating experience with HMS AMAZON (2,500 Tonnes; 160 men). This

experience has clearly demonstrated that whereas the scheduled tasks of day-to-day operational deployment present little difficulty for the crew, the

ship is nevertheless manned very close tb the minimum limit in her ability to meet the unscheduled task load, and that very little margin exists to

absorb the effects of illness, promotion, and other personnel contingencies.

The conclusion at this stage is that although the technology now exists (or if it does not exist already, it will certainly be developed in the near future) which will enable a large proportion of scheduled tasks to be automated, it would be quite wrong to assume that such an increase in automation would, by itself, bring about a compensating reduction in ships' complements. Moreover, there is a danger that the widespread and exclusive

automation of scheduled tasks will upset the delicate balance between interesting and rewarding work for the ship's company, and tedious but necessary "chores". The upsetting of this balance could lead to a lowering

of motivation and morale. Thus it is clear that in future warship designs the automation of any task must be considered in relation to the manning policy for the ship as a whole in order to achieve the right balance between effectiveness, economy and job satisfaction.

The corollary of this argument is that if it is required to seek the reduction of warship complement as a desirable objective for economic or other reasons, then the approach should be to "prepare the ground" for further automation by first reducing the unscheduled task component. The following list indicates a number of the more obvious ways by which this could be

achieved:-a. Modifications to ship operational requirements and operating characteristics,

including:-Acceptance of shorter mission times.

Acceptance

of

lower availability. Reduced flexibility in operational role.

Less emphasis an Ship survival following action damage

Greater reliance on shore support.

b. Measures to eliminate manpower - intensive unscheduled

tasks:-Automation of fire-detection and fire-fighting functions.

Elimination of manual reversionary modes of operation by implementation of greater redundancy in system design.

c. Measures to reduce the need for high-grade technical support

afloat:-Increased reliance on system redundancy in design.

More accurate prediction of system/equipment failure.

Acceptance of a higher mission abort rate.

d. General