The Collision Risk Analysis Of Very Large Ships In Singapore Strait

Singapore Strait is one of the busiest straits in the world. There are more than one hundred thousand vessels carrying important cargoes through Singapore Strait. Because of the large ship traffic volume and the narrow channel, marine accidents occur frequently in the strait. Besides the ship’s development tendency to large-size is obvious, many very large ships enter the team of carriage of goods by sea. Once the very large ship is involved in a collision accident, it will cause heavy casualties and property losses, even the irreversible environmental pollution. Therefore, it’s extraordinarily significant to analyze the very large ship collision risk in Singapore Strait.This thesis summarizes the existing collision risk analysis methods and typical collision models, and analyzes their respective advantages and disadvantages. This thesis introduces the situations of route and traffic separation scheme in Singapore Strait. This thesis collects datas of traffic volume and maritime traffic accident, and analyzes these datas. This thesis defines very large ship as ships of more than 100,000 dwt, including very large bulk carrier, very large container ship and very large oil tanker. This thesis devides Singapore Strait into 6 waters according to traffic separation scheme characteristics, and creates different COWI models in each water. This thesis makes the very large ship collision probability calculation system in Singapore Strait using MATLAB. Through the system, we can calculate the collision probability between bulk carrier, container ship and oil tanker of very large size and 4 levels of ships of 20 thousands tons below,20 to 50 thousand tons,50 to 100 thousand tons and more than 100 thousand tons within the waters in 6 waters. This thesis analyzes the results of probability calculation and gives the ranking of collision probability in 6 waters. This thesis concludes that ship’s average speed within the water is the safest speed.Based on BP neural network, this thesis sets up the very large ship collision consequence assessment model in Singapore Strait using MATLAB. This thesis selects 6 characteristic indexes, including ship type, ship tonnage, ship age, ship speed, collision position and collision angle. This thesis sorts accidents into three categories: ordinary accident, serious accident and very serious accident. This thesis designs a neural network of 10 inputs and 3 outputs, and divides existing accident datas into a training group and a validation group. This thesis validates the model using the validation group, and recognition rate can be stabilized at 85%. Through the consequence assessment model, we can predict the consequence of a collision accident inputing the relevant characteristic indexes, which can provide maritime administration authorities with some references.Finally, this thesis obtains very large ship collision risk in Singapore Strait multiplying collision probability with consequence probability, and corresponding risk control measures are proposed.