Kristiansen Svein. Maritime Transportation: Safety Management and Risk Analysis
Подождите немного. Документ загружается.
Mackworth (1946) made one of the early studies of the temperature effect on
monitoring tasks as shown in Figure 11.6. It can be seen that the number of monitoring
mistakes starts to increase sharply above an effective temperature of 30
C. This
corresponds to the upper limit for what is experienced as comfortable.
Obviously the question of acceptable temperature is also related to the duration of
exposure. As shown in Figure 11.7, the duration of unimpaired mental performance
decreases quickly as the effective temperature rises above 30
C (Wing, 1965).
Finally, it should be mentioned that there seems to be a special combined effect of
warmth and sleep loss (Pepler, 1959). Observed phenomena are reduced performance in
tracking tasks and gaps in serial responses. Simply stated, the following mechani sms are
seen: ‘Warmth reduces accuracy’ whereas ‘Sleep loss reduces activity’.
Ta b l e 11 . 4 . Thermal climate factors on vessels
Section Situation Solution
Living quarters Reasonable control Ventilation and air conditioning is
standard
Navigation bridge Thermal control up to 22–28
C
Heat sources: large window panes,
electronic equipment and open
doors
Ventilation and air conditioning is
standard
Galley A number of heat sources: stove, etc. Improved isolation of heat sources
Long periods above thermal comfort
criteria
Engine rooms Air cooling of engines
Temperature often 10–20
higher
than outside temperature
Heat radiation from hot surfaces
Heavy repair and maintenance work
High relative humidity in smaller,
confined spaces
Work in engine rooms will in general
take place well outside thermal
comfort criteria
Under winter conditions large
temperature variations within
the same space
Cargo tanks Inspections and final cleaning
operations may be stressing in
hot outside climate. Protective
clothing may contribute to
the stress
Cleaning by permanently installed
equipment
On deck Both high and low temperatures
depending on time of the year
and geographical latitude
Sun radiation
Clothing and protection
Effect of wind (air speed) under
winter conditions
326 CHAPTER 11 HU MAN FA CT ORS
11.7.3 Noise
The primary noise sources onboard are: main engine, propeller, auxiliary engines and
engine ventilation. The noise (or unwanted sound) is measured in terms of sound intensity
and expressed as dB(A). Table 11.5 summarizes maximum recorded and recommended
noise levels for the main ship sections.
11 .7.4 Infrasound
Infrasound is acou stic waves below 20 Hz that are not audible by the human ear. It has,
however, been found that for high intensities infrasound has negative effects on the human
in terms of reduced well-being, tiredness and increased reaction time.
The relation between infrasound and possible negative effect is still only partly
understood. It has been suspected that even higher intensities (above 100 dB) may have
adverse effect on control of balance, disturbance of vision and choking.
Figure 11.6. Monitoring mistakes as a function of effective temperature. (Source: Mackworth,1946.)
Figure 11.7. Upper limit for unimpaired mental performance. (Source: Wing,1965.)
11 .7 PHYSICAL WORK ENVIRONMENT 327
Typical ranges for infrasound measured on vessels are 4 Hz (55 dB) to 16 Hz (95 dB).
The fact that infrasound is present both on the bridge and in the engine control room
may pose a problem to safe operation.
11.7. 5 Vibr ation
Vibration is seen by a significant part of the crew as one of the most disturbing
environmental factors onboard. This is partly explained by the fact that it is experienced
both at work and off-duty. This means that persons affected are given no opportunity
for restitution during free hours. Vibration is expressed by acceleration (m/s
2
)in
selected frequency bands (Hz). Vibrations are usually measured for a range from 0.5 Hz
to 125 Hz.
The main sources of vibration are the main engine and the propeller. The induced
vibrations are further transmitted through the steel hull and deck houses. High
superstructures are further subject to resonance phenomena. It is also experienced in
the aft part as an effect of the hogging–sagging movement of the hull girder, giving very
low frequency resonance (0.5 Hz).
ISO has given norm values for exposure (time) versus vibration, both with respect
to ‘reduced work performance’ and ‘reduced comfort’. Figure 11.8 illustrates the
Ta b l e 11. 5 . Noise in vessels
Section Situation Solution
Living quarters Highest values: dB(A) ¼58–70
Depends mainly on the distance to
engine rooms
Variation with ship type
28% experienced the noise
as troublesome
Navigation bridge Highest values: dB(A) ¼65–73 Lowest noise levels on vessels with
bridge in the fore part
(passenger, Ro-Ro)
Some effect on direct communication
and use of internal communication
equipment
Galley Mean value: dB(A) ¼71–77
Recommended value 65 dB(A) is exceeded
due to the background noise
Engine rooms Highest values: dB(A) ¼93–113 Wear ear protection
Recommended value: 100 dB(A)
Factors: power, engine type
Sectioning of engine room
Risk of physiological damage
(reduced hearing)
50% of engine personnel report the
noise as troublesome
328 CHAPTER 11 HU MAN FA CT ORS
ISO norm for maintaining profiency. It can be seen that the most sensitive frequency
area is roughly 4–8 Hz. For longer durations the vertical acceleratio n should not exceed
0.31 m/s
2
.
Measurements of vibrations by Iverga
˚
rd et al. (1978) can be summarized as foll ows:
. Navigati on bridge: Most vessels have a vibration level below recommended values.
The exception is tankers. However, crew reports some discomfort.
. Engine room: Most observations lie between the comfort and performance curves.
Although the comfort limit is not exceeded, personnel reports problems due to the
reduced opportunity for restitution during the free watch.
. Cabin area: Vibration level coincides with comfort line in the critical area 8–16 Hz.
Confirms that restitution becomes a problem.
11. 8 EF F EC T OF SEA MOTION
One of the classical torments of life at sea is the so-called seasickness that may affect both
passengers an d crew. The more correct term is motion sickness as it is present in different
transport modes and in certain other situations. Motion sickness may be experienced
during anything from riding a camel, operating a microfic he reader to viewing an IMAX
movie. How strongly one may be affected differs, but it usually involves stomach
discomfort, nausea, drowsiness and vomiting, to mention a few.
The problem has recently been surveyed by Steven s and Parson s (2002) in a maritime
context. As the authors point out, one of the consequences of seasickness is less motivation
and concern for the safety critical tasks onboard. However, for the majority of persons
there will be an adaption to sea motion over time so that the sickness fades away after a
few days. There is also some comfort in the fact that susceptibility usually decreases with
increasing age. The effect of moti on appears sometimes in the form of the sopite syndrome,
which is less dramat ic and usually only experienced as drowsiness and mental depression.
Figure 11.8. Vibration tolerance limits as function of exposure time. (Adapted from ISO 2631,1985.)
11.8 EFFECT OF SEA MOTION 329
This affection may also be safety critical in the sense that it has an effect on performance
and is not always evident to the subject.
Various explanations of motion sickness have been proposed, but contemporary
theory points to the conflict or mismatch between the organs that sense motion:
. Vestibular system in the inner ear: semicircular canals that respond to rotation and
otoliths that detect translational forces.
. The eyes (vision system) that detect relative motion between the head and the
environment as the result of motion of either or both.
. The proprioceptive (somatosensory) system involves sensors in body joints and muscles
that detect movement or forces.
The theory assumes that under normal situations the three systems detect
the movements in an unambiguous way. However, under certain conditions the senses
give conflicting signals that lead to motion sickness. A typical maritime scenario is
experiencing sea motion inside the vessel without visual reference to sea, horizon or land
masses. In this case the vestibular detects motion in the absence of visual reference.
Watching the waves from the ship may give rise to conflict between vision and vestibular
system.
Extensive experiments in motion simulators found that subjects were primarily
sensitive to vertical motion (heave) and that maximum sensitivity occurred at a frequency
of 0.167 Hz (Griffin, 1990). Given the fact that the principal vertical frequency in the sea
motion spectre is 0.2 Hz, the occurrence of seasickness is understandable. There are
two methods for estimation of motion sickness: Motion Sickness Incidence (MSI) and
Vomiting Incidence (VI). Both methods are outlined by Stevens and Parsons (2002). There
are a number of effects of seasickness:
. Motivational: drowsiness and apathy
. Motion-induced fatigue (MIF): reduced mental capacity and performance
. Reduced physical capacity
. Added energy expenditure to counterbalance motion
. Sliding, stumbling and loss of balance
. Some interference with fine motor co ntrol tasks
The effect on cognitive tasks is more inconclusive as it has been difficul t to isolate the
effect of physical stress. Another problem is the fact that bridge tasks involve a number
of cognitive processes and skills which may be influenced differently by the sea motion
(Wertheim, 1998).
In order to minimize the risk of seasickness, it is necessary to establish operational
criteria. Baitis et al. (1995) point out that it is not the roll angle in itself that lim its the
operation but rather the vertical and lateral accelerations associated with them. Table 11.6
shows the criteria proposed by NATO (NATO STANAG 4154, 1997), which are based on
both earlier and recent principles. An alternative set of criteria make a distinction between
330 CHAPTER 11 HU MAN FA CT O RS
different ship types (NORDFORSK, 1987) as summarized in Table 11.7. It can, however,
be concluded that the two sets of criteria agree fairly well.
1 1.9 H UMAN RELIABILITY ASSESSMENT
Risk analysis involves making quantitative estimates of the risk associated with designs
and operations. Typical is estimation of the probability of having specified accident events
by means of fault tree analysis (FTA) or event tree analysis (ETA). In order to make
realistic estimates, human-error-related events must be incorporated. This is not a trivial
problem given our limited understanding of this phen omenon and lack of systematic data.
Despite this, various approaches have been developed and to some degree have also
been supplemented with quantitative data. The objective of human reliability assessment
(HRA) has been stated as follows by Kirwan (1992a, 1992b):
1. Human error identification: what can go wrong?
2. Human error qua ntification: how often will it occur?
3. Human error reduction: how can it be prevented or the impact reduced?
The key methodological steps are outlined in Figure 11.9.
Considerable experience has been accumulated with the so-called first-generation
HRA methods, primarily from applications in the process and nuclear industries.
However, they have come under increasing attack from cognitive scientists who point
to the fundamental lack of ability to model human behaviour in a realistic manner
(Hollnagel, 1998).
Ta b l e 11. 7. General bridge operability criteria for ships
Merchant ships Naval vessels Fast small craft
Vertical acceleration (RMS) 0.15 g 0.2 g 0.275 g
Lateral acceleration (RMS) 0.12 g 0.1 g 0.1 g
Roll (RMS) 6.0
4.0
4.0
Source: NORDFORSK (1987).
Table 11.6. Sea motion operability criteria
Motion sickness incidence (MSI) 20% of crew in 4 hours
Motion-induced interruption (MII) 1 tip per minute
Roll amplitude 4.0
RMS
Pitch amplitude 1.5
RMS
Vertical acceleration 0.2 g RMS
Lateral acceleration 0.1 g RMS
Source: NATO STANAG 4154 (1997). RMS, root mean square; g, acceleration of gravity.
11.9 H UMAN RELIABI L ITYASSESSM ENT 331
11.9.1 Th er p
The Technique for Human Error Rate Predict ion (THERP) is probably the best-known
and most widely used technique of human reliability analysis. The main objective of
THERP is to provide human reliability data for probabilistic risk and safety assessment
Figure 11.9. Human reliability assessment steps (Kirwan,1992a,1992b).
332 CHAPTER 11 HU MAN FA CT ORS
studies. The methods and underlying principles of THERP were developed by Swain and
Guttmann (1983) and are often referred to as the THERP Handbook.
The methodological steps are:
1. Identify system failures of interest: This stage involves identifying the system functions
that may be influenced by human errors and for which error probabilities are to be
estimated.
2. Analyse related human operations: This stage includes performing a detailed task
analysis and identifying all significant interactions involving personnel. The main
objective of this stage is to create a model that is appropriate for the quantification
in stage 3.
3. Estimate the human error probabilities: In this stage the human error probabilities
(HEPs) are estimated using a combination of expert judgements and available data.
4. Determine the effect on system failure events: Estimating the effect of human errors
on the system failure events is the main task of this stage. This usually involves
integration of the HRA with an overall risk/safety assessment (i.e. PRA/PSA).
5. Recommend and evaluate changes: In this stage changes to the system under
consideration are recommended and the system failure probability recalculated.
Possible solutions for various human factors problems include job redesign,
implementation of mechanical interlocks, administrative control s, and implementa-
tion of training and certification requirements.
The probability of a specific erroneous action is given by the following expression:
P
EA
¼ HEP
EA
X
m
k¼1
PSF
k
W
k
þ C
where:
P
EA
¼ Probability of an error for a specific action
HEP
EA
¼ Basic (nominal) operator error probability of a specific action
PSF
k
¼ Numerical value of kth performance shaping factor
W
k
¼ Weight of PSF
k
(numerical constant)
C ¼ Numerical constant
m ¼ Number of PSFs
The probability is a function of the error probability for a generic task modified
by relevant performance shaping factors. The basic HEPs can be looked up in 27
comprehensive tables in the THERP Handbook (Swain and Guttmann, 1983). The PSFs
are tabulated in the same fashion. The three sets of PSFs are shown in Table 11.8.
The modelling of relevant human actions in event trees is based heavily on the task
analysis performed in stage 1. In the present stage a more detailed analys is is carried out
11.9 H UMAN RELIABI L ITYASSESSM ENT 333
Ta b l e 11 . 8 . PSFs inTHERP
Situational
characteristics
Architectural features
Temperature
Humidity
Air quality
Lighting
Noise and vibration
Degree of general cleanliness
Work hours and work breaks
Availability of special equipment
Adequacy of special equipment
Shift rotation
Organizational structure
Adequacy of communication
Distribution of responsibility
Actions made by co-workers
Rewards, recognition and benefits
Job and task
instructions
Procedures required (written or not) Work methods
Cautions and warning Plant policies
Written or oral communication
Task and equipment
characteristics
Perceptual requirements Frequency of repetitiveness
Motor requirements (speed,
strength, etc.)
Control–display relationships
Anticipatory relationships
Interpretation
Decision-making
Complexity (information load)
Narrowness of task
Task criticality
Long- and short-term memory
Calculation requirements
Feedback (knowledge of results)
Dynamics vs. step-by-step activities
Team structure and communication
Man–machine interface
Psychological
stressors
Suddenness of onset Long, uneventful vigilance periods
Duration of stress Conflicts about job performance
Task speed Reinforcement absent or negative
High jeopardy risk Sensory deprivation
Threats (of failure, of losing
job, etc.)
Distractions (noise, flicker,
glare, etc.)
Monotonous and/or
meaningless work
Inconsistent cueing
Physiological
stressors
Duration of stress Atmospheric pressure extremes
Fatigue Oxygen insufficiency
Pain or discomfort Vibration
Hunger or thirst Movement constriction
Temperature extremes Lack of physical exercise
Radiation Disruption of circadian rhythm
G-force extremes
Organism factors Previous training/experience Emotional state
State of current practice or skill Sex differences
Personality and intelligence variables Physical condition
Motivation and attitudes
Knowledge required
Stress (mental or bodily tension)
Attitudes based on external
influences
Group identification
334 CHAPTER 11 HU MAN FA CT ORS
on each of the relev ant human actions identified and a comprehen sive description of the
performance characteristics is made. Each human/operator action is divided into tasks
and subtasks, and these are then represented graphically in a so-called HRA event tree
(Figure 11.10).
HRA event trees model performance in a binary fashion, i.e. as being either a success
or a failure. Branches in a HRA event tree show different human activities, and the
values assigned to all human activities represented by the branches (except those in the
first branching) are conditional probabilities. Figure 11.11 illustrates a simple HRA
event tree (Swain and Guttmann, 1983). As can be seen from this figure, it is common
to present the correct actions on the left-hand side of the tree and failures on the
right-hand side.
11.9. 2 Cri t icis m o f th e HR A A pp r o a c h e s
It was commented above that the binary categorization of human erroneous actions
used in HRA event trees may be too simple to make any claim on psychological realism
(Hollnagel, 1998). First of all there is an important difference between failing to
perform an action and failing to perform it correctly. Furtherm ore, the HRA event tree
approach fails to recognize that an action may happen in many different ways and for
different reasons.
Figur e 11.10. HRA event tree for two successive subtasks.
11.9 H UMAN RELIABI L ITYASSESSM ENT 335