National Ship Risk Model

The Norwegian Maritime Authority (NMA) and the Norwegian Coastal Administration (NCA) must prioritize the resources available for accident prevention. Their efforts should be directed to areas where the risk is high. NSRM has developed a risk model that can be used to monitor and communicate risk in a better way than today. This will contribute to a better targeting of the accident prevention efforts in the maritime area. The models and our methods for normalization are currently being used by our project partners in NMA (primarily) and to a some extent by the NCA. We have developed influence diagrams mapping the main risk influencing factors. This is a systematic way of mapping causes and influences for these accident types. Not only the immediate causes, but also more remote or underlying causes but also (eg.) human, organizational and regulatory factors that influence the risk of accidents. These have been disseminated and discussed with NMD and NCA, and can be used in communication about risk and risk management. These models can be used as a basis for Bayesian Belief network models in order to conduct quantitative analysis of risk influencing factors, providing us the opportunity to better understand which factors are most important for different accident types. This is currently being tested for some applications in Master theses. A part of this development process we have employed data from the incident data base managed by the NMD. However, for data on accidents to be useful in risk analysis, they must be seen against the amount of activity that does not lead to accidents. For example we must investigate whether changes in the number of accidents are caused by changes in traffic or whether there are other causes. To this end we have processed the data gathered from AIS transponders of ships to get data on the movements of ships in Norwegian waters generally. AIS is an abbreviation of Automatic Identification System. It is mandatory for all vessels above a certain size, and the system gives extensive information concerning the maritime traffic along the Norwegian coastline. These data can, among other things, inform us where the safety measures are most needed. The project has developed methods for harvesting AIS data and for linking these data to accident databases. This enables us to see possible causes and factors that are associated with different maritime accident types, and the condition of factors that are significant for risk. We have developed a system for harvesting and managing AIS-data and tested their application for normalization of accident data. Our approach has been adopted by the NMD for their monitoring of risk in Norwegian waters. We have also generated normalized statistics for several types of incidents and vessel categories. We have tested different methods of normalization, and can conclude that this improves the ability to analyze accident frequencies and construct more nuanced measures of risk. It also has methodological consequences that can be relevant for other transport sectors. Normalized frequencies, based on our methods are currently used by the NMD in reports etc. The project was completed in two years (autumn 2014-autumn 2016). The executive partners were NTNU Social Research - Studio Apertura, Safetec Nordic and NTNU. The NMA and NCA both contribute financially and will participate actively in the project. The project has already published several papers in peer reviewed proceedings. At least two journal papers and several Master theses are expected. In addition the methods developed for incident statistics are currently being used by the involved public agencies, and our modeling framework (the diagrams of risk influencing factors) are being used for different purposes by researchers, students as well as NMD/NCA.

Safe transport at sea is an important premise for moving more transport from road to sea. This project contributes to improved control with major accidents at sea and to an enhanced safety level. We will develop the framework of a national ship risk model to be used by the regulatory authorities to control maritime risk. The project is based on a stepwise methodical approach. A crucial step in the development of the risk model is to normalize existing incident data. They will be combined with activity d ata harvested and aggregated from AIS-system. Normalization will enable us to conduct far more sophisticated analyses of incidents involving vessels in Norwegian waters than currently possible. By isolating traffic as separate factors the precision and ut ility of the accident statistics for research and regulation is strengthened. The AIS data can give rich information regarding conditions that may have contributed to past incidents. Moreover, the combination of incident data and AIS data can form a basis for identifying Risk Influencing Factors. These can also be used for assessing current risk levels, and as input to the National Ship Risk Model. We will study historical incidents and AIS data in combination in order to identify Risk Influencing Fact ors. These are crucial input to the risk model, but some may also be monitored (near) real-time based on new AIS data. The National Ship Risk Model will aggregate input and data from the previous work packages and integrate it into a framework that can be used to monitor the risk levels at sea. It can also inform inspections and regulatory measures. The project will produce technical reports documenting each step of the work, for other researchers to pick up. In particular the scripts for harvesting and a ggregating AIS-data may be useful beyond the application it is put to here. The academic output of the project is one journal article on normalized incident data and one on the development of NSRM.