Risk Modelling of Collisions between Passing Vessels and Fish Farm Facilities - A Development of a Frequency Model
MetadataShow full item record
- Institutt for marin teknikk 
The aim of this master's thesis is to develop a model that is tailored to estimate the frequency of a collision between a powered passing vessel and a fish farm facility. The aquaculture industry is expected to grow, which forms the need for improved methods for planning and approval of new locations for fish farm facilities. The vessel-fish farm collision frequency model developed, is based on the collision candidate-causation probability approach. The basic principles for this approach were presented by Macduff in 1974, and has formed the basis for later frequency modeling. In recent years, there has been an increased emphasis on identifying risk influencing factors, RIFs, in frequency models, which can be used to identify risk-reducing measures. Through a literature survey on existing frequency models, important RIFs were extracted and divided into the categories; human and organizational factors, vessel characteristics, technical factors, route characteristics, external factors and traffic density. In the model developed in this thesis, AIS data containing information about actual traffic patterns is used to estimate the probability of being a collision candidate. The causation factor is estimated from a quantitative Bayesian belief network (BBN), consisting of the RIFs identified in the literature survey. A frequency model will never give a holistic picture of the reality and there will always be uncertainties attached to such models. The model developed for estimating the frequency of a vessel-fish farm collision has many similarities with existing frequency models. An advantage with the model is that it distinguishes between vessel types and sizes, which constitute a different risk of a collision. In addition, it distinguishes between facilities that are located in protected and in exposed areas, which makes it applicable for virtually all fish farms. A case study was conducted to show how the model is used in practice. The facility Ocean Farm 1 was assessed, which is the first "offshore" fish farm in Norway. The results for the causation factor was congruent with findings from existing literature. However, the estimated frequency of having a vessel collision with Ocean Farm 1 was estimated lower compared to grounding frequency estimates found in existing literature. This may be related to the probability of being a collision candidate, where the normal probability distribution was used to describe the traffic patterns. A more conservative approach, which is recommended for further work, would be to combine the normal distribution with a uniform distribution.