Threat Level Forecast for Ship´s Oil Spill - Based on BP Neural Network Model

It´s very important to assess the threat level in time when the ship´s oil spill occurred, because the threat level forecast will help to come to a decision when dealing with the accident. BP neural network model is proposed in this paper to build a thread level forecast method for ship´s oil spill accident. Train the BP neural network first and then make a simulation.Considering the spilled oil amount, spilled location or integrated sensitivity, the weather when the accident happened etc., the BP neural network can output the thread level of the accident. In this paper, we select the recent oil spill accident data of Guangzhou Xiaohu waters which is representative, and through the simulation results to find out that the forecast result is close to the experts´ estimate. Hope that the method proposed in this paper would provide guidance for relevant departments when dealing with emergency incidents.