427
1 INTRODUCTION
TheintroductionoftheConventionontheStandards
ofTraining,CertificationandWatchkeeping 1978, as
amended1995(STCW95)andtheInternationalSafety
Management(ISM)Code,implementedthroughFlag
StateImplementation(FSI)orPortStateControl(PSC)
has much increased safety standards related to the
personnel performance in maritime shipping. Safety
ManagementSystemcanbeimprovedbyident
ifying
human factors and analyzing human interactions
(Einarsson, and Brynjarsson, 2008). New accidents
show that there is still a room for development of
human factor, mainly with respect to fatigue and
stress (Berg, 2013; Fahlgren, 2007; Hejmilch, 2015;
Hetherington, 2006; Mokhtari & Khodadadi Didani,
2013).
According to The Institute of Ergonomics and
HumanFactors(UK)thedefinitionforhumanfact
or
is formulated as follows: “Ergonomics (or Human
Factors)isthescientificdisciplineconcernedwiththe
understanding of interactions among humans and
other elements of a system, and the profession that
appliestheory,principles,dataandmethodstodesign
in order to optimise human well‐being and overall
systemperformance.”EncyclopaediaBritannicasays:
“Hum
an‐factors engineeringalso called ergonomics
or human engineeringscience is dealing with the
application of information on physical and
psychologicalcharacteristics to thedesignofdevices
andsystemsforhumanuse.Thetermhuman‐fact
ors
engineering is used to designate equally a body of
knowledge, a process and profession. As a body of
knowledge,human‐factorsengineeringisacollection
of data and principles about human characteristics,
capabilities and limitations in relation to machines,
jobsandenvironment”.
There are three categories of human fact
ors
specifiedinanthropotechnicalsystems:
individualfactors(Reason,1998):
competencelevel,
stress,
motivation,
groupfactors:
managementweaknesses,
supervisionandcrewfactors,
Human Factor Modelling in the Risk Assessment of
Port Manoeuvers
T.Abramowicz‐Gerigk&A.Hejmlich
GdyniaMaritimeUniversity,Poland
ABSTRACT:Thedocumentationofhumanfactorinfluenceonthescenariodevelopmentinmaritimeaccidents
comparedwithexpertmethodsiscommonlyusedasabasisintheprocessofsettingupsafetyregulationsand
instructions. The new accidents and near misses show the necessity for further studies in det
ermining the
human factor influence on both risk acceptance criteria and development of risk control options for the
manoeuversin restrictedwaters. Thepaperpresents themodel ofhumanerrorprobability proposedfor the
assessmentofshipmastersandmarinepilotsʹerrordecisionanditsinfluenceontheriskofportmanoeuvres.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 9
Number 3
September 2015
DOI:10.12716/1001.09.03.16
428
organizationalfactors,
companypolicies,
companystandards,
systemsandprocedures.
The International Maritime Organization (IMO)
hasconsideredthehumanfactorasthereasonof80%
ofallmaritimeaccidentsmostlyconcentratedontwo
areas:
negligence‐30%ofmarineaccidents,
not adequate knowledge and practice‐20%
of
marineaccidents.
Personal characteristics of people involved in
eventsclearlyimplicatecausalfactors relating to the
workplaceandsystem(Reason,1998).
The comprehensive study on human and
organisationalfactorsinshipcollisions,presentedby
Chauvin et al., (2013) stresses the necessity to
investigatethemastersʹdecisionsincriticalconditions
concerningmainlybridgemanningandvesselspeed.
ThesedecisionsarerelatedtoSMS,ISMandcompany
safetyculture.
TheefficiencyoftheISMisalsoputintoquestion
(Battacharya,2012)asitisunderstoodbyshipmasters
mainly as a managerial tool and undermines
professionalskillsofseafarersleadingtoits
rejection.
Shipmastersresponsiblefortheshipoperationsat
the same time can have personal problems affecting
their behaviour. This creates a more stressful
environmentforamansupposedtobefitfor24hours
duty. Lots of paper work imposed on the ship staff
doesnotmakesituationany
batter.Companybudget
holders seek to minimize the manpower costs
reducing the crew number on board. Stresses
accumulate and they influence the probability of
seafarersʹ mistakes and wrong decisions. The main
objectiveofthestudypresentedinthepaperistofind
outhowdoesthestressinfluencethedecisionmaking
process and consequently the human error during
shipmanoeuvringinrestrictedarea.
2 HUMANERRORASSESSMENT
Main data for the assessment of human error in
maritimeshippingistheinformationavailableinpast
accidentinvestigationreports.TransportSafetyBoard
(TSB) of Canada and Marine Accident Investigation
Branch (MAIB) of the
UK take human and
organisational factors into account using their own
investigativemethodologies.
The full text of 44 Incident at Sea Reports was
coded and analysed using NUDIST software by
Philips (2000). This sample includes collisions and
groundings reported since 1991, where significant
human factors were contributed tothe incident. The
Incident at Sea Reports was electronically searched
forreferencetosleepandcontentwasindexedagainst
parameters such as fatigue behaviours, time of day
and contributing personnel. Incident at Sea Reports
incorporatethreelevelsofreferencetosleep,analysis
ofwhich mayassociatesleepingand sleepinesswith
accident causation. The
highest level of reference
associatesbeing asleep orbeingsleepdeprived with
accidents,butnotalwayswithfatigue.
Chauvinetal.(2013)presentedtheHumanFactor
Analysis and Classification System designed for
humanfactors accident investigationand analysis of
the underlying causes of human error. The HFACS
hasbeendeveloped on
thebasis of previous studies
andapplicationsusedinmilitaryandcivilaviation,in
the railway and mining industries for the
classification and analysis of accident data, in
shipping industry for the investigation of maritime
accidentsandmishaps(Chauvinetal.,2013).HFACS
systemdescribeshumanerroratfourlevels
offailure:
unsafeactsofoperators,
preconditionsforunsafeacts,
unsafesupervision,
organizationalinfluences.
With respect to ship collisions the unsafe acts
relatedtodecision‐makingconcern85%ofthevessels
involved in the investigated 27 collisions. The main
reasons of collisions in restricted waters involving
licensed
pilots were inter‐ship communication
problems and BRM(Bridge Resources Management)
deficiencies(Chauvinetal.,2013).
The limitation of the studies based on
investigation reports is to small number of cases in
comparison to the number of variables, which does
not support the stochastic modelling. It is also very
difficult
to investigate the influence of ship masters
andmarine pilotspersonalitycharacteristicsontheir
attitudes towards risk and probability of error in
decisionmakingprocessduetothelackofevidence.
The influence of the stressing factors on the
decision making process is important in risk
assessment of accidents related to
the navigation in
restrictedwatersandthiscanbestudiedmainlyusing
expert methods. To identify the personality
characteristicandstressorsranking,thepsychological
questionnaires,followedbystatisticalanalysiscanbe
used.
2.1 Personalitycharacteristics
Stressor is the appropriately ranked collection of
stressing factors. Every stressor creates stress and
brings
up the risk of error making (Makarowski,
2008). Human personality characteristics called “Big
FivePersonalityTraits“arethecatalystsforthestress
development. The five personality traits are as
follows:
openness,
extraversion,
conscientiousness,
neuroticism,
agreeableness.
The determination of error decision probability
allows predicting the risk of
wrong manoeuvre
performance. With respect to the definition adopted
by EMSA (Report, 2014), given by ISO 1999, risk is
definedasaproductoftheprobabilityofoccurrence
ofharmandtheseverityofthatharm.
The simplified model of the influence of human
factor,relatedtopersonalitycharacteristicof
theship
master or marine pilot on the probability of error
decision,risk of wrongmanoeuvreperformance and
numberoferrorsispresentedinfigure1.
429
Figure1Simplifiedmodeloftheinfluenceofhumanfactor
related to personality characteristic of the ship master or
marinepilotontheprobabilityoferrormanoeuvre
Theidentificationofthestressingfactorsshouldbe
basedonthewidestudyonshipmastersandmarine
pilots psychological characteristics followed by the
practicaltestsonshiphandlingsimulator.Theresults
willbeusefulforthedeterminationofweightswhich
can be attributed to the experts performing
navigational analyses or
to determine the main
stressors influencing the decision making process of
the ship master on board, assessment of risk
acceptancelevelsanddevelopmentofISMcode.
2.2 Stressasapredictoroffailure
Increasedlevelsofstress(HancockandSzalma,2008;
Sneddonetal,2012)canresultin:
reduced
workingmemorycapacity,
diminishedattention,
poorconcentration/alertness,
lowerprimaryperceptionofthesituation,
inattentiontotheavailableinformation,
narrowingoftheindividual’sattentionalfield,
incorporation of only a restricted number of core
aspects.
The levels of occupational stress (Mearns and
Hope, 2005) and associations between
stress and
accidentratesatseahavebeenmeasuredinoffshore
industry (Sutherland and Cooper, 1986, 1996;
Sneddonetal.,2012).Theresultsof amediationtest
showed that 48% of the relationship between stress
and unsafe behaviour was mediated by work
situationawareness(Sneddonetal.,2012).
The stress
and fatigue were correlated but the
multiple regression analysis showed that stress was
theonly significant predictor of situationawareness.
Stresswas foundtobe the onlysignificant predictor
for:
concentration,
projection,
attention
distraction.
The fourth examined variable was distraction.
Thesignificantpredictorfordistractionwasthe
sleep
disruption. Distraction can affect all the stages of
perception,comprehensionandanticipationtherefore
these distractive items were clustered into a single
factor.
Theprincipalcomponentsanalysisofthesituation
awarenessscaleproducedfinallyafour‐factormodel
ofthepreviouslymentionedvariables(Sneddonetal.,
2012):
FortheWork
CognitiveFailuresScale(Broadbent
et al., 1982; Wallace and Chen, 2005) a three‐factor
model has been identified using confirmatory factor
analysis:
memory,
attention,
action.
3 ASSESSMENTOFPORTMANOEUVRES
DIFFICULTY
There are several phenomena which should be
considered in the risk assessment of error during
manoeuvring in
restricted waters (Abramowicz‐
Gerigk&Burciu, 2012; Abramowicz‐Gerigk&Burciu;
2014, Gerigk, 2012; Inoue, 2013). Full Mission Ship
Handling Simulator is a tool which allows for the
determination of risk indices and arrangement of
repetitive test conditions. The risk indices are the
limiting values of three dimensional safe
manoeuvring space, ship motion parameters,
ship
propulsion and steering devices setting and output
parameters in space and time domain (Hejmlich,
2014).
Themodel manoeuvreprepared onthesimulator
by the competent expert team can be used for the
assessment of the manoeuvre performed by the
examined ship master. Comparing the errors in
dependence on human factor,
allows concluding the
humanfactor(inthe meaningofstress)influenceon
theerrorprobabilityanddefiningtheacceptablerisk
fortheparticularmanoeuvreinrestrictedarea.
The example of a model manoeuvre and risk
indicesdeterminedusingFullMissionShipHandling
Simulator are presented in figures 2 and 3.
The
assumed deviation of a particular device setting is
considered as an error in manoeuvre performance.
Five points of highest risk have been defined in the
examinedareaofPortofGdynia.
Toidentifythe attitudes of the examined pilots a
simplequestionnairetobefilledoutbeforeandafter
the
test manoeuvre performed on ship handling
simulatorhadbeen proposed.The pilotswere asked
to assess the difficulty of the test manoeuvre. An
exampleofthequestionnaireispresentedinfigure4.
The three independent variables influencing the
manoeuvre performance were taken into account:
knowledge of port area and experience‐
number of
yearsofemploymentasashipmasterandnumberof
yearsofemploymentasapilot.
430
Figure2. Example of a model manoeuvre: propulsion and
steering devices parameters, ship speed and bottom
clearanceintimedomain
Figure3. Example of the manoeuvre performance by
examinedpilots‐geometricalboundaries
‐distancesfrom
centerlineofportcanal
The preliminary set of three groups of pilots’
attitudestowardsriskhadbeenidentified:
chancer,eagertotakearisk,
neutral,conservative,
passive,reluctanttotakearisk.
Figure4. Example of a questionnaire for port pilots
performingmanoeuvreonshiphandlingsimulator
There were not clear dependencies observed
betweentheexaminedvariablesonthe tested group
of17pilots.Theconclusionwasthereforetocontinue
the research using bigger group of experts and
prepare the questionnaires to determine their
psychological characteristic and the degree of stress
impactingtheprobabilityoferrordecision.
The
ever‐present hazards associated with ship
master activity are called universals. They are
included in one of the three groups of elements
recognised in every recurrent accident scenario
(Reason,1998):
universals,
localtraps,
drivers.
In the maritime transport the universals include
(Reason,1998):
rocks,
shallow
waters,
currentsandtides,
presenceofothervessels.
The example drivers that act to propel different
peopledownthesameerror‐provokingpathwaysare
asfollows(Reason,1998):
timepressure,
cost‐cutting,
indifferencetohazards.
Thelocaltrapsarecharacteristicsofataskthat,in
combination with human error and violation
tendencies attract people into repeated scenarios of
unsafeactsorunsafeperformance(Reason,1998).
Standard occupational stress questionnaires were
foundnotsuitableformaritimeindustry.Insteadthe
list of 24 stressors was customised on the basis of
offshorestressscales(Langan‐Fox&Cooper,2011).
Thestressors
include:
workoverload,
threatofjobloss,
demandsofworkdayandnight,
conflicts on board and conflicts with shorebased
office.
Respondents rated how much stress they
perceived from each of 24 items (listed in stressors
questionnairepresentedinfigure5)ona5‐point
scale,
ranking from Definitely NO to Absolutely YES
(Sneddonetal.,2012).
The following questionnaires can be used for
personalitycharacteristicassessment:
Questionnaire of Stress Sense
(Plopa&Makarowski,2010),
QuestionaireofStimulationandInstrumentalRisk
(Makarowski,2012,
QuestionaireNEO‐FFI(Costa&McCrae,2003).
The example of a questionnaire prepared for
the
ship masters for the assessment of the degree of
stress,includingtheabovementionedelements‐the
difficultsituationatseaispresentedinfigure5.
431
Figure5. Questionnaire STRESORS – Ship Master, the
additionalinformation in thequestionnaire consists of age
andexperienceasashipmasterinyears
4 RISKASSESSMENTOFPORTMANOEUVRES
Thefurtherresearchwillbebasedontheresultsofthe
statistical studies of ship masters and pilotsʹ
personalities and probability of their influence on
stressdevelopmentunderparticularstressors.
4.1 Probabilityoferrordecision
Themainstressingsituations‐stressors–selectedfor
manoeuvring in
restricted waters and personality
characteristicshouldbedetermined.Theinfluenceof
age and professional experience together with the
influence of stress probability on the decision error
should be examined on the basis of the tests results
performedonFullMissionShipHandlingsimulator.
The transparent structure and possibility of risk
determinationinuncertaintyconditionsarethemain
reasons to implement a model based on Bayesian
network (BN). BN is a technique based on
probabilistic and uncertain knowledge, widely used
fortheriskassessment.BNisadirectedacyclicgraph,
in which the nodes represent the random variables
and arcs represent causal
relationships between the
nodes.
Theconditionalprobability tablesassignedtothe
nodesdenoteconditionaldependencies.Basedonthe
conditional independence of variables A
1,..., An and
the chain rule, BN represents the joint probability
distribution of variables {A
1,...,An} (Abramowicz‐
Gerigk&Burciu,2014).
ProbabilityofproductofeventsA
1,..,An,where:
0
121
)A...AA(P
n
(1)
isdeterminedasfollows:
)A...AA|A(P)...A|A(P)A(P
)A...AA(P
nn
n
121121
21
(2)
Theevents
n
A,...,A,A
21
formtheMarkovchainif:
)A|A(P)A|A(P)A|A(P)A(P
)A...AAA(P
nn
n
123121
321
(3)
The events can be replaced by binary values of
randomvariablesandthentheaboveequationshave
the forms of Equations 4 and 5 (Abramowicz‐
Gerigk&Burciu2014).
},{x,...,x,x
),xX,...,xX,xX|X(P...
)...xX|xX(P)xX(P
)xX,...,xX,xX(P
n
nnn
nn
10
21
112211
112211
2211
(4)
432
},{x,...,x,x
),xX|X(P...
)...xX|xX(P)xX(P
)xX,...,xX,xX(P
n
nnn
nn
10
21
11
112211
2211
(5)
Bayesian network for the assessment of stress
influenceonshipmastererrordecisionispresentedin
figure6.
Figure6. Bayesian network for the assessment of stress
influence on ship master error decision, the states of
random variableʺPersonalityʺ are assumed as: open,
extraverted,conscientious,neurotic,agreeable
4.2 Riskassessmentofportmanoeuvres
The example of Bayesian influence diagram for risk
assessment of port manoeuvres is presented in
Figure7.
Figure7. The influence diagram for the risk assessment of
harbour accidents for the vessel entering the port.
Simplified example of a ferry entering Port of Gdynia
(Abramowicz‐Gerigk&Burciu,2014)
The nature nodes in the diagram represent the
randomvariableswithtwostates:safeandALARP
(Abramowicz‐Gerigk&Burciu,2014):
entrancechannel‐ probability of accident during
navigationalongtheapproachchannel,
main entrance‐probability of accident during
entryintotheport,
docks‐probabilityof accident during navigation
insidethe
docks,
berth HelskieII‐ probability of accident during
berthingatberthHelskieII,
manoeuvres on arrival‐probability of accident
duringportmanoeuvres.
Thenaturenodes,representingexternalconditions
areasfollows(Abramowicz‐Gerigk&Burciu,2014):
sea current ‐ probability of occurrence of
dangerous windcurrent,states:yes,no,
wind velocity‐probabilityofoccurrence of
dangerouswind,states:wind<7B,wind7B,
wind direction‐probabilityofoccurrence of
dangerouswind,states:yes,no.
The description of utility nodes representing risk
ofaccidentsispresentedintable2.
Table2.Descriptionofutilitynodes.
_______________________________________________
Node Description
_______________________________________________
A/M
M/A
=RA/M‐riskofshipgroundinginchannel
A/Z
Z/A
=RA/Z‐riskofcollisioninchannel
B
B
=RB‐riskofaccidentinmainentrance
C
C
=RC‐riskofaccidentindocks
D
D
=RD‐riskofaccidentduringberthing
_______________________________________________
The decision error node represents probability of
errordecisionwithtwostates:yes,no.
The example Bayesian network and influence
diagrampresentedinthepaperweredevelopedusing
HuginResearcher7.4commercialprogramfordesign
ofthenetworkstructure,datainputandcalculations
ofthejointprobabilitydistributionfor
thenodes.
5 CONCLUSIONS
Ithasbeenstatedinseveralstudiesthatincidentand
near miss reporting is deficient in the shipping
industryandithasbeenrecognisedasthefailingpart
ofISMcodeimplementation(Lappalainen,2011).For
thisreasonthereisnosufficientdatafortheanalysis
of human
factor based on statistical methods. Ship
motionsimulationcan beused todetermine therisk
of a particular manoeuvre (Gucma, 2009), combined
withexpertsopinion.Theweightingfactorsassigned
to the experts give better understanding of the
subjectivejudgementoftheacceptablerisklevel.
Humanfactormodellingintherisk
assessmentof
port manoeuvers is very important with respect to
different attitudes towards the risk of marine pilots
andshipmasters.Thepsychologicalcharacteristicsof
ship masters and pilots, combined with identified
stressorscreatestress.Theapproachproposedinthe
paper would allow establishing susceptibility for
stress of each personality and
using statistical
methods defining susceptibility for stress for the
populationofshipmastersandpilots,withrespectto
theirageandexperienceaseitherstressmoderatoror
acceleratorrespectively.
Decision
error
433
REFERENCES
[1]Abramowicz‐Gerigk T. Burciu Z. 2011. Comprehensive
analysis of ship performance during self‐berthing.
Advancesin Engineering Research,Chapter 13,Volume
3,NovaSciencePublishersinc.,p.487‐503.
[2]Abramowicz‐Gerigk T., Burciu Z., 2014. Safety
assessmentof maritime transport‐Bayesianrisk‐based
approach in different fields of maritime transport.
ProceedingsofIMAM2013,15thInternationalCongress
of the International Maritime Association of the
Mediterranean (IMAM), Spain. Developments in
Maritime Transportation and Exploitation of Sea
resources.GuedesSoares&LopezPena(eds),Volume2,
2014Taylor&FrancisGroup,London,UK,p.699‐703.
[3]BattacharyaS.,2012.Theeffectivenessofthe
ISMCode:a
qualitativeenquiry.MarinePolicy36,528–535.
[4]Berg,H.P.,2013.HumanFactorsandSafetyCulturein
Maritime Safety (revised). TransNav, the International
Journal on Marine Navigation and Safety of Sea
Transportation,7(3),343–352.
[5]Chauvin C., Lardjane S., Morel G., Clostermann, J.‐P.,
Langard B., 2013. Human
and organisational factors in
maritime accidents: Analysis of collisions at sea using
theHFACS,AccidentAnalysisandPrevention59,p.26–
37.
[6]Corović B.,Djurovic P., 2013. Research of Marine
AccidentsthroughthePrismofHumanFactors.Promet
– Traffic&Transportation, Vol. 25, 2013, No. 4, p. 369‐
377.
[7]Costa,
P&McCrae, R., 2003. A contemplated revision of
the NEO Five‐Factor Inventory. Department of Health
andHumanServices,NationalInstituteonAging,NIH,
Baltimore,Maryland,USA
[8]Encyclopediabritannica.http://www.britannica.com/.
[9]FahlgrenG.K.,2007.WhatisHumanFactorsCompared
toCrewResourceManagement
?,1,109–112.
[10]Flin R., OʹConnor P., CrichtonM. 2008. Safety at the
SharpEnd:AGuidetoNon‐TechnicalSkillsPaperback,
AshgatepublishingLtd.,Hampshire,England.
[11]GerigkM.,2012.Assessmentofsafetyofshipsafterthe
collision and during the ship salvage using the matrix
type risk
model and uncertainties. in Sustainable
Maritime Transportation and Exploitation of Sea
Resources. Proceedings of the 14th International
Congress of the International Maritime Association of
theMediterranean(IMAM),Volume2:715‐719,London:
Balkema.
[12]Gucma, L., 2009. Wytyczne do zarządzania ryzykiem
morskim.Szczecin:AkademiaMorskawSzczecinie.
[13]Hanzu‐
Pazara, R., Barsan., E., Arsenie, P., Chiotoriou,
L.,Raicu,G.:Reducingofmaritimeaccidentscausedby
human factors using simulators in training process,
Journalof Maritime Research,Vol. V,No 1, 2008,p. 3‐
18.Hejmlich A., 2014. Czynnik ludzki w ocenie
bezpieczeństwamanewrówportowych.PraceWydziału
NawigacyjnegoAkademiiMorskiej
wGdyniz.26.
[14]Hancock P.A., Szalma, J.L., 2008. Performance under
Stress.AshgatePublishing,Aldershot,England.
[15]Hetherington C., Flin R., Mearns K., 2006. Safety in
shipping: The human element. Journal of Safety
Research37,p.401–411.
[16]Inoue,K., Okazaki, T., Murai, K., &Hayashi,Y.,2013.
Fundamental
Study of Evaluation at Berthing Training
forPilotTraineesUsingaShipManeuveringSimulator.
The International Journal on Marine Navigation and
SafetyofSeaTransportation,p.135‐141.
[17]Langan‐Fox,J.,Cooper,C.,L.,2011.HandbookofStress
in the occupation. Edward Elgar Publishing Ltd.,
Cheltenham
[11]Lappalainen, J., 2011. Incident
reporting in Finnish
ShippingCompanies.WMUJoin(Berg,2013).
[18]Makarowski, R., 2008. Granice Ryzyka. Kraków:
OficynaWydawniczaImpuls
[19]Makarowski, R. 2012, Kwesrionariusz Ryzyka
Stymulacyjnego i Instrumentalnego nie tylko dla
sportowców.PsychologicalJournalVol18No2
[20]Mearns K., Hope L., Ford M., Tetrick L., 2010.
Investment
in workforce health: exploring the
implications for workforce safety climate and
commitment. Accident Analysis and Prevention 42, p.
1445–1454.
[12]Mokhtari,A.H.,&KhodadadiDidani,H.R.,2013.An
EmpiricalSurveyontheRoleofHumanErrorinMarine
Incidents. TransNav, the International Journal on
Marine Navigation and Safety of
Sea Transportation,
7(3),p.363–367.
[21]PhillipsR.,2000.Sleep,watchkeepingandaccidents:A
content analysis of incident at sea reports,
TransportationResearch–PartF,No3,2000,229‐240
[22]Plopa M., Makarowski R., 2010. Kwestionariusz
PoczuciaStresuPodręcznik.VisiaPress&IT,Warszawa
[23]Reason J.,
1998. Achievinga safe culture. 1998: theory
andpractice.Work&Stress,VOL.12,NO.3,p.293‐306.
[16]Report No.: PP092663/1‐1/2, R. 2., 2014.
EMSA/OP/10/2013 Risk Level and Acceptance Criteria
for Passenger Ships. First interim report, part 2: Risk
AcceptanceCriteria.EuropeanMaritimeSafetyAgency.
[24]Sneddon A.,
Mearns K., Flin R. (2012). Stress, fatigue,
situation awareness and safety in offshore drilling
crews.SafetySci.Volume56,p.80‐88.
[25]SutherlandV.J.,CooperC.L.,1986.ManandAccidents
Offshore:TheCostsofStressamongWorkersonOiland
GasRigs.Lloyd’sList/Dietsmann,London.
