The Evaluation Safety Of Passenger Ferry Based On Dynamic Risk Prediction

Passenger ferries play an important role in the field of carriage of passengers by sea which gradually replace short distance passenger ships.This traffic mode shows an incomparable superiority in the transportation near shore,over strait or between islands.But because of the bad condition of ships,low comprehensive quality of crew members,and little management efficiency of companies,a large number of life and property loss comes up from the marine perils.Thus,it is of vital importance to analyze and evaluate the risk and safety of the ferries transportation.The paper uses a risk prediction technique of Dynamic Bayesian Network(DBN)to forecast and evaluate the ferries transportation.The Dynamic Bayesian Network is based on a probability network which adds time information to the Static Bayesian Network,where the dynamic mode has an ability to dispose the time series data.The probabilities of every grades of accidences will be eventually obtained by establishing the DBN predicting modes,structuring the DBN network structure and forecasting the risk.Moreover,the obtained variation trends of the probabilities of each grades of accidences occurrence by time can be analyzed to clarify the variation trend of the whole risks.In the case analysis,the value of risk in every time slice and every grades of accidence probability can be obtained by analyzing real cases.It shows that the established mode can nicely reflect the dynamics of the ferries navigation risk.According to the forecasted results,the managers and crew members can get a better reference to forecast the navigation risk and take appropriate safety measures.