SAMSON

The risk assessment model SAMSON has been developed for the Dutch government since the mid-80s. The development originally started as an EU project (COST301) and has also known a number of other working titles in the past.

The model uses a voyage database which contains all ship movements over a period of time (usually one year), specifying ship type, ship size, port of departure and port of destination. For each voyage a route is constructed on the virtual network of shipping links, based on shortest route but obeying restrictions on e.g. draft and cargo, and with a preference to use a nearby TSS. This yields a number of ship movements (per type, size and direction) for each individual link.

The expected number of ship encounters is calculated from the number of ship movements. If the number of meeting encounters is compared with the observed average number of head-on collisions per year a so-called casualty rate may be derived - which has the same function as the causation factor used by IWRAP. The same process is followed for other types of accidents by comparing the exposure to the number of accidents.

The casualty rates are dependent on ship type and size, on weather conditions and probably also many more factors. The availability of accident data usually limits the number of factors that may be used.

SAMSON also contains models for the damage in case of a collision or contact. Together with typical bow forms and tank layouts for the different ship types and sizes, an estimate of the outflow of cargo and fuel oil may be made.

More details

The parts constituting SAMSON are described in detail in separate pdf files.

The models used to calculate the exposure are described for ship - ship collisons for ships underway and for a collision with an anchored vessel separately. The latter has a strong resemblance with the contact - a vessel versus an object or stranding line.

For contacts there are different models for a powered collision, when a vessel gets off course and does not correct her course in time, and a drifting collision which occurs when a vessel is not under control and drifts with wind and current onto an object. Only if the crew succeeds to anchor in time or to regain control the contact can be avoided.

When the exposure for each of the accident types has been calculated and the accident records of the same area are available, the casualty rates can be determined. With world-wide accident statistics the correlation between casualty rate and ship type and size may be added. Also historical trends may be derived from the world-wide data, allowing for the estimation of accident probabilities in future years. The process of determining the casualty rates, as it has been done in 2004 for the Dutch North Sea area is described. The casualty data as may be purchased from Lloyd's List Intelligence are specified in the separate files field specification and coding tables.

SAMSON has been used on many occasions to determine risks of oil outflow. A collision may or may not result in oil outflow, depending on the extent of the damage, the location of bunker and cargo tanks and whether the tanks were full or not. A large amount of statistical data, both of accidents and of common tank layouts depending on ship type and size, was used to constitute the outflow model.

Whereas IWRAP is based on observed (AIS) tracks of vessels, SAMSON uses the ports of departure and destination of voyages. These voyages have to be assigned to a network of available traffic links, to result in a traffic database which finally gives the traffic density and composition for each link. The way the network of imaginary links is built up and the way the voyages are routed is explained in"creating a traffic database". It should be noted that this process was designed long before AIS came into existence. Instead data of aerial observations have been used to verify the routing process. In the near future research will be undertaken to use AIS data for deriving voyage data instead of purchasing voyage data.