Kristiansen Svein. Maritime Transportation: Safety Management and Risk Analysis
Подождите немного. Документ загружается.
Hollnagel (1998) argues that in cases where human behaviour is a cognitive function
or a mental act, rather than a manual action, the use of the event tree description does
not makes sense. This is because cognitive acts are not so easily separated into dist inct
and sequential subtasks, as is the case with manual actions. Human cognitive functions,
such as diagnosis and decision-making, and in particular the notion of a ‘cognitive error’,
play an increasingly important role in HRA. The standard approach used to model
cognitive functions in HRA event trees has been to decompose these functions into their
assumed components or subtasks, e.g. problem identification, decision-making, execution
and recovery. This may at first seem sensible, but there are a num ber of problems related
to it:
. The number of subtasks that need to be modelled in the HRA event trees in order to
incorporate the cognitive functions would easily become excessive, making the event
trees complex and difficult to model.
. It is difficult to obtain appropriate and reliable Human Error Probability (HEP) data
on each of the (assumed) components of cognition.
. There are doubts as to whether such modelling is psychologically realistic.
Below is given a comprehensive list of the shortcomings of first-generation HRA
approaches. These shortcomings should be kept in mind in the overall assessment of
quantitative risk estimates:
. Scarcity of data: W ithin HRA there is a general lack of appropriate data that can be
used for quantitative predictions of human erroneous behaviour. Much of the data
Figure 11 .11. Mode lling princi ples inTHE RP.
336 CHAPTER 11 HU MAN FA CT O RS
used comes from sim ulator studies, and there is no basic agreement on how these data
reflect the real world, or how they can be modified in order to achieve this.
. Lack of consistency in treating errors of commission: Fragola (2000) argues that this is
one of the most glaring weaknesses of conventional HRA approaches. Well-intended
actions with unintended and undesirable consequences largely remain an ‘unknown’
area in HRA.
. Inadequate proof of accuracy: Demonstrations of the accuracy of HRAs for real-world
predictions are almost non-existent. This goes particularly for non-routine tasks.
. Inadequate psychological realism: Many first-generation HRA approaches are based on
highly questionable assumptions about human behaviour.
. Insufficient treatment of performance-shaping factors: In most first-generation HRA
approaches the influence of factors such as management attitudes, safety culture,
cultural differences, etc. (i.e. organizational factors) are not treated sufficiently. There
is also little knowledge on how these factors actually affect performance.
. Inadequate treatment of dynamic situations: Conventional HRA approaches are
relatively static, meaning that they do not consider the dynamic situations under which
tasks are performed.
. A mechanical view of humans: Conventional HRAs use a decomposition appro ach
adapted from reliability analysis of mechanical systems (i.e. hardware) on human
actions. This decomposition is binary, i.e. either success or failure. In addition,
this view results in a focus on the observable (or overt) aspects of human behaviour,
instead of the more ‘internal structure’ of the problem space.
. Quantitative rather than qualitative focus: This is argued by some to be a weakness
of many conventional HRA approaches, mainly because the quan titative estimates
produced are so uncertain.
. High level of uncertainties: Different HRA methods may produce widely different
values for the Human Error Probabilities (HEPs) when used on the same tasks.
. Inadequate identification and understanding of error causes: This weakness is in
accordance with the weakness of ‘inadequate psychological realism’ described above.
Most of the first-generation HRA approaches run into problems when trying to
explain why humans make errors.
. Lack of a systematic analysis of task structure: Analysts who use HRA methods/
techniques make judgements based on the information obtained from task analysis.
Therefore, a sys tematic task analysis is essential to enhance the validity and
consistency of the HRA results. Most first-generation HRA approaches lack such a
systematic task analysis.
. Inadequate provision of error reduction strategies: Few HRA methods provide clearly
defined strategies for how the estimated HEPs may be reduced in order to enhance
safety.
It is only in the past decade or so that the criticism directed at the techniques of
conventional HRA have resulted in various efforts to resolve the problems stated above. It
should also be recogni zed that second-generation HRA approaches should be developed
from the point of view of solving the limitations of the first-generation HRAs.
11.9 H UMAN RELIABI L ITYASSESSM ENT 337
REFEREN CES
Baitis, A. E. et al., 1995, Motion Induced Interruptions (MII) and Motion Induced Fatigue (MIF)
Experiments at the Naval Biodynamics Laboratory. Technical Report CRDKNSWC-HD-1423-
01, Naval Surface Warfare Center, Carderock Division, Bethesda, MD.
Caridis, P. A. et al., 1999, State-of-the-Art in Marine Casualty Reporting, Data Processing and
Analysis in EU Member States, the IMO and the United States. CASMET Project Report
C01.D.001.1.1, National Technical University of Athens, Department of Naval Architecture
and Marine Engineering, Athens.
Elliott, E., 1960, Perception and alertness. Ergonomics, Vol. 3, October, pp. 357–364.
Fragola, J., 2000, Focus, strengths, and weaknesses of HRA. In Swedung, I. (ed.), Risk Management
and Human Reliability in Social Context. 18th ESReDA Seminar, Karlstad, Sweden.
Fukushima, H., 1976, Factors contributing to marine casualties. Journal of Navigation, Vol. 29(2),
135–140.
Griffin, M. J., 1990, Motion sickness. In Handbook of Human Vibration. Academic Press, New York.
Hollnagel, E., 1998, Cognitive Reliability and Error Analysis Method (CREAM). Elsevier Science,
Oxford, UK.
ISO 2631, 1985, Guide to the Evaluation of Human Exposure to Whole Body Vibration. International
Standardization Organization.
Iverga
˚
rd, T. et al., 1978, Work conditions in shipping – A survey. In Swedish: Arbetsmiljo
¨
innom
Sjo
¨
farten – En kartla
¨
ggning. Sjo
¨
fartens Arbetarskyddsna
¨
mnd, Stockholm.
Jayle, G. E. et al., 1959, Night Vision. Charles C. Thomas, Spingfield, FL.
Kirwan, B., 1992a, Human error identification in human reliability assessment. Part 1: Overview of
approaches. Applied Ergonomics, Vol. 23(6), 299–318.
Kirwan, B., 1992b, Human error identification in human reliability assessment. Part 2: Detailed
comparison of techniques. Applied Ergonomics, Vol. 23(6), 371–381.
Mackworth, N. H., 1946, Effects of heat on wireless telegraphy operators hearing and recording
Morse messages. British Journal of Industrial Medicine, Vol. 3, 143–158.
Mackworth, N. H., 1950, Researches on the Measurement of Human Performance. Medical Research
Council Special Report 268. HMSO, London.
Madison, R. L., 1974, Human factors in destroyer operations. M.Sc. Thesis, Naval Postgraduate
School, Monterey, CA (NTIS: AD 775 011).
McCormick, E. J., 1976, Human Factors in Engineering and Design. McGraw-Hill, New York.
McCormick, E. J. and Sanders, M. S., 1983, Human Factors in Engineering and Design. McGraw-
Hill, New York.
NAS, 1976, Human Error in Merchant Marine Safety. Maritime Transportation Research Board –
Commission on Sociotechnical Systems, National Academy of Sciences, Washington, DC.
NATO STANAG 4154, 1997, Common Procedures for Seakeeping in the Ship Design Process.
Chapter 7: Seakeeping Criteria for General Application.
338
NORDFORSK, 1987, The Nordic Cooperative Project. Seakeeping Performance of Ships. Assessment
of a Ship Performance in a Seaway. Marintek, Trondheim.
Pepler, R. D., 1959, Warmth and lack of sleep: accuracy or activity reduced. Journal of Comparative
and Physiological Psychology, Vol. 52, 446–450.
Rasmussen, J., 1982, Human reliability in risk analysis. In Green, A. E. (ed.), High Risk Safety
Technology. John Wiley, Chichester.
RSS, 1975, An Assessment of Accident Risk in US Commercial Nuclear Power Plants. WASH 1400
(NUREG 74/014), US Nuclear Regulatory Commission.
Salvendy, G. (ed.), 1987, Handbook of Human Factors. John Wiley, New York.
Stevens, S. C. and Parsons, M. G., 2002, Effects of motion at sea on crew performance: a survey.
Marine Technology, Vol. 39(1), 29–47.
Stewart, J. P., 1973, Basic causes of marine casualties. Tanker & Bulk Carrier, June, pp. 18–27.
Swain, A. D. and Guttmann, H., 1983, Handbook of Human Reliability Analysis with Emphasis on
Nuclear Power Plant Applications. NUREG/CR-1278, US Nuclear Regulatory Commission,
Washington, DC.
Teichner, W. H. (1972). Predicting Human Performance III. The Detection of Simple Visual Signal as
a Function of Time on Watch. New Mexico State University, Department of Psychology Report
NMSU-ONR-TR-72-1.
VanCott, H. P. and Kincade, R. G., 1972, Human Engineering Guide to Equipment Design.US
Superintendent of Documents, Washington, DC.
Wertheim, A. H., 1998, Mental Load is Not Affected in a Moving Environment. Report TNO-TM-98-
A068. TNO Human Factors Research Institute, Soesterberg.
Wilson, J. R. and Corlett, E. N., 1990, Evaluation of Human Work: A Practical Ergonomics
Methodology. Taylor & Francis, London.
Wing, J. F., 1965, A Review of the Effective Ambient Temperature on Mental Performance. USAF,
AMRL, TR 65-102.
REFERENCES 339
12
OCCUPATIONAL SAFETY
‘There has been an accident’, they said,
‘Your servant’s cut in half; he’s dead!’
‘Indeed!’ said Mr Jones, ‘and please,
Send me the half that’s got my keys.’
(Harry Graham: ‘‘Ruthless Rhymes for Heartless Homes’’)
12 .1 INTRODUCTIO N
One of the first systemat ic studies of the risk picture for seafarers was done by Arner
(1970). The breakdown by fatality type shown in Table 12.1 is from this study and
highlighted one important aspect, namely that the loss of seafarers co uld be attributed to
at least three different situations: loss of vessel, work-related and non-work-related.
It is interesting to observe that 11% of the fatalities are related to loss of vessel,
28% are directly work-related (explosion/fire, fall, poisoning) and the remaining 61% are
in different ways more related to social and psychological factors. This means that safety
work must address three main areas:
. The safety of personnel in situations where the survivability of the vessel is threatened:
fire, explosion, sinking, capsizing, etc.
. The safety related to the fact that the vessel is a dangerous workplace (fall, hit by
object, poisoning).
. Thirdly, the fact that seafaring is to a certain degree associated with negative social and
economical factors.
These problem areas still exist, although their relative importance has changed since
Arner’s study (1970) was undertaken.
12 .2 LIVES LOST A T SEA
Nielsen (2002) points out that although the fatality rate in shipping is relatively high
compared with other occupations, there has been a certain improvement over time as
shown in Figure 12.1. He has, however, pointed out that the figures are quite uncertain
due to inadequate statistics on a world-wide basis. Assuming that there is under-reporting,
341
one may assess the risk level to be in the order of 10 times as high as in a comparable
land-based indust ry.
Figure 12.2 shows how the accident frequency has developed during a 10-year period
in Norwegian shipping. Although we know that the safety level has improved in a
longer perspective, it is interesting to observe that there are still cyclical variations. It is
difficult to explain these shorter-term variations, but is in general it is attributed to
business conditions, competition, the labour market, and so on.
The fatality rate on a world-wide basis was investigated recently by Nielsen (1999). The
author points out that the available data are incomplete and affected by under-reporting.
Ta b l e 1 2 . 1 . Death by type of Norwegian seafarers,1957^ 64
Main fatality type No. %
As a result of ship loss 110 11
Explosion, fire 73 7
Fall 134 13
Poisoning 87 8
Drowning 288 28
Traffic accident 42 4
Found in the water 47 5
Suicide 159 15
Homicide 28 3
Other 59 6
Total 1027 100
Source: Arner (1970).
Figure 12.1. Lives lost at sea on world-wide basis. (Source: Nielsen, 2002.)
342 CHAPTER 12 OC CUPATIONAL SA FETY
However, he has given some interesting estimates which are summarized in Table 12.2.
The estimate is based on incomplete data in different death categories, referred to
as figures from ‘responding states’. These figures have then been corrected for the
known seafaring population world-wide. The method gives an annual world-wide
loss figure of nearly 2600. In this estimate the losses related to ship accidents
represent 47%, occupational accidents (work-related) 16%, illnesses 21% and off-duty
accidents 6%.
In other words, it can be concluded that roughly 50% of fatalities are related to
phenomena other than ship accidents. Occupational accidents represent 15% whereas the
rest is related to illness, suicide, missing and events off-duty. This indicates that there are
many negative aspects of the sea-going occupation.
Table12.2. A world-wide estimate for the number of lives lost at sea
Death category Average annual
fatalities, 1990–94,
responding states
Population at risk,
responding states
Fatality rate
per 1000
work-years
Annual
fatalities
world-wide
Relative
distribution
world-wide
Ship accident 84.4 1,148,822 36.7 1229 47.4
Occupational 71.4 911,140 39.2 404 15.6
illness 62.2 582,728 53.3 550 21.2
Missing at sea
Homicide,
11.2
7.4
569,728
526,539
9.8
7.0
101
72
3.9
suicide,
unexplained 2.8
Off-duty 4.4 11.8 122 4.7
Unclear cause 4.4 11.4 117 4.5
Total 245.4 — — 2595 100.0
Source: Nielsen (1999).
Figure 12 .2. Reported work accidents in the Norwegian merchant fleet: accidents per 1000 work-years.
(Source: NMD, 1 999. )
12 . 3 LI V E S LOS T AT SE A 343
12.3 OCCUPATIONAL ACCIDENTS
12.3.1 United Kingdom
Li and Shiping (2002) have analysed accident data for the UK for the years 1989–97.
As shown in Table 12.3, the dominating occupational accident form is slips and falls,
constituting 46% of the total number. The next largest groups are manual handling
and machine operation, both accounting for 19%. It is also impor tant to notice that the
relative number of accidents related to ship casualties is only 15% compared to 47% on a
worldwide basis.
12.3.2 Norway
Eriksen (2000 ) has compared the risk level in Norwegian shipping w ith similar
occupations. As indicated in Table 12.4, the fatality rate is more than 10 times as high
as in land-based industry. A breakdown by occupational accident category is given in
Table 12.5. Data for both fatal and non-fatal accidents are shown. It confirms the picture
from the UK that falls and manual handling are primary sources.
1 2.3.3 Acci dents by ShipType
There is limited information on how the accident rate varies among ship types. A notab le
exception is the recent study of the Danish foreign trading fleet (Hansen and Vinter, 2004).
The key results are shown in Table 12.6. Looking at fatal accidents, it is clear that the
rate for coasters is much higher than for the rest of the fleet. With respect to accidents
leading to permanent injury, container ships, passenger ships and gas tankers perform
Ta b l e 1 2 . 3 . Fatal and non-fatal accidents on UK-registered ships
Year Slip/
fall
Manual
handling
Machine
operation
Rope/
hawse
Hit by
object
Noxious
substance
Electric
shock
Ship
casualty
Total Population Rate
(%)
1989 212 105 104 30 32 13 2 82 500 20,958 2.39
1990 239 106 93 41 20 8 4 79 516 18,289 2.82
1991 239 86 90 31 21 4 6 87 479 16,005 2.99
1992 181 89 72 24 13 4 8 57 392 14,118 2.78
1993 154 54 75 18 23 12 3 67 341 13,351 2.55
1994 158 77 70 16 18 3 4 58 346 12,837 2.70
1995 142 54 53 12 17 12 9 40 301 12,175 2.47
1996 166 69 59 25 30 4 9 33 364 11,043 3.30
1997 153 64 57 16 19 6 3 23 320 11,044 2.90
Total 1644 704 673 213 193 66 48 526 3559 129,820 24.89
Mean 183 78 75 24 21 7 5 58 395 14,424 2.77
% 46.19 19.78 18.91 5.98 5.42 1.85 1.35 14.78 100
Source: based on Li and Shiping (2002) and data from MAIB.
344 CHAPTER 12 OC CUPATIONAL SA FETY
best. Gas tankers have an exceptionally low rate taking all reported accidents into
consideration.
It is necessary to stress that the Danish data may not be representative for the world
fleet as a whole.
Ta b l e 1 2 . 4 . Fatality rate in Norwegian industry,1999
Sector Shipping and
fishing
Offshore Land-based
industry
Number of fatalities 31 1 57
At risk (workforce) 58,000 13,600 2,000,000
Fatality rate ¼fatalities
per 1000 employees
0.53 0.07 0.03
Source: Eriksen (2000).
Ta b l e 1 2 . 5 . Work accident types in Norwegian merchant shipping
Work accident type Non-fatal in 1998 Fatal 1994–98
No. % No. %
Fall, jump 289 30.3 35 44.9
Blow, squeeze, stumble 256 26.9 10 12.8
Lift, handling 107 11.2 —
Hurt by tool or machine 86 9 2 2.6
Contact with hot/cold substance 72 7.6 —
Hit by object 72 7.6 4 5.1
Poisonous substance 36 3.8 —
Violence, fight 6 0.6 1 1.3
Explosion, fire 5 0.5 3 3.8
Sport, play 5 0.5 —
Suffocating media 4 0.4 3 3.8
Traffic accident ashore 2 0.2 1 1.3
Contact with electricity 1 0.1 1 1.3
Assault 1 0.1 1 1.3
Terrorism, piracy 1 0.1 —
Other 10 1 5 6.4
Missing — 12 15.4
Total 953 100 78 100
Source: NMD (1999).
12.3 OCCUPATIONAL ACCIDENTS 345