Risk-based ship security analysis – an approach based on civilian and military methods

The demands on maritime operations today are increasingly higher in terms of control, efficiency and cost. The margins for accidents and security incidents are therefore decreasing. In the area of ship safety the regulations, guidelines and methods have a history and culture of systematic research, development and implementation. In contrast, international security is highly politicized and therefore not as transparent. The result is that a tradition of ship security is not as well established. The overall aim of this thesis is to propose a method for ship security analysis that increases the overall safety of the crew and the ship. The objective is to develop a method that is systematic in order to ensure that assessment and response are complete and effective, and that the process is documented to provide evidence of decision-making. The method used is probabilistic risk assessment where quantitative analysis is central. The proposed approach is consistent with the requirements of maritime safety work. However, in the work here, the proposed methods are specifically tested for security cases. This is because hazards (without intent) and threats (with intent) evolve in different ways into risk. Therefore, they must be analysed differently in order to capture the causal relationship. The proposed approach consists of three steps: the first step consists of a threat description that documents qualitative and quantitative aspects that together describe how the threat most likely will act in relation to the ship’s vulnerability; the second step uses the threat description to define the system studied as well as the scenarios that collectively describe the harmful consequences; the third step evaluates the risk with tools from probabilistic risk assessment. The overall conclusion is that the proposed method brings the procedure and results of ship security analysis into the open and therefore allows for criticism, improvements and shared risk knowledge, not possible with less structured methods. The results also show that the calculated probabilities agree with available statistics, which indicates that the analysis succeeds in describing the central causal relationships of the scenarios modelled