A Big Data Analytics Method for the Evaluation of Ship - Ship Collision Risk reflecting Hydrometeorological Conditions

Abstract

This paper presents a big data analytics method for the evaluation of ship-ship collision risk in real operational conditions. The approach makes use of big data from Automatic Identification System (AIS) and nowcast data corresponding to time-dependent traffic situations and hydro-meteorological conditions respectively. An Avoidance Behavior-based Collision Detection Model (ABCD-M) is introduced to identify potential collision scenarios and Collision Risk Indices (CRIs) are quantified when evasive actions are taken for each detected collision scenario in various voyages. The method is applied on Ro-Pax ships operating over 13 months of the ice-free period in the Gulf of Finland. Results indicate that collision risk estimates may be extremely diverse among voyages, and in 97.5% of potential collision scenarios the evasive actions are triggered only when risk is at 45% or more of its maximum value. The overall CRI for ships operating over the given area tends to be lower for adverse hydro-meteorological conditions. It is therefore concluded that the proposed method may assist with the (1) identification of critical scenarios in various voyages not currently accounted for by existing accident databases, (2) definition of commonly agreed risk criteria to set off alarms, (3) the estimation of risk profile over the life cycle of fleet operations.