| dc.description.abstract | In recent years, the management of risk has developed towards a risk based approach. Thus,
industries need to have a comprehensive understanding of why accidents occur and how they
develop. In order to act successfully in the prevention of accidents, the areas in which risk reduction is most beneficial need to be identified. For this reason, quantitative risk analysis has developed onwards. One method, which will be considered in detail to support these investigation, is the Bayesian Belief Network.
A Bayesian Belief Network (BBN) is a graphical model which visualises the causal relationship
between different factors and the final outcome. It represents a flexible approach which
can be used qualitatively and quantitatively.
Current research on the quantification of these factors in BBN models are mainly based on
statistical approaches evaluating incident and accident reports. This approach however comprises some problems related to a lack of data, data overload and underreporting. Thus, it needs to be decided on a correction factors, a safety margin or to rely on expert judgement. For that reason another approach based on a framework of risk influencing factors (RIFs) and risk indicators can be used to measure the effect on risk covering the overall socio-technical system.
This master thesis focusses on the identification of suitable RIFs and indicators for technical, human and organisational factors of a BBN ship grounding model. Based on a literature review the influence of BBN nodes on the occurrence of grounding accidents is investigated.
The analyses show that 80% of causes are human and organisational related, whereas only
20%represent technical causes. Often a range of causes need to be considered in order to understand the whole complexity behind grounding accidents. The introduction of new technology and automation does not always benefit marine navigation. It results in a polarised workload structure, the reduction of task-related communication and a decrease in situation awareness. Furthermore, the individual risk perception was found out to influence peoples behaviour.
This explains the occurrence of groundings also in good weather and good technical conditions
of ships. For that reason, one should improve the human-machine interface rather than
adding new technology. One beneficial method could be user-centred design combined with
regular and better trainings of personnel.
In summary, the identification of indicators represents a complex process that due to various
approaches and context-specific understanding cannot give one ultimate outcome. The
implementation of risk indicators in an operational context, represents a beneficial tool for performance surveillance and risk control. | |
