Shanghai Port Waters Ship Oil Spill Analysis And Development Trend Forecast

Along with the speedy development of the economy and the foreign trade in china, especially in Shanghai and its surrounding, the Shanghai port becomes an international port stage by stage; meanwhile, it attracts more and more cargos and vessels. According to the statistics from the Ministry of communications, the throughout increases greatly from 13,960 million tons in 1990 to 44320 million tons in 2005 with the annual growth of 8.8%. As a result, the traffic environment has been changed and the ships'oil spillage has become more serious with larger and faster vessels; besides, the average spilled-oil increases fastly. The ships'oil spill accident is paroxysmal and uncertain, but they jeopardize the ocean calamitously.To minimize the loss, we must complete beforehand prevents and afterwards emergency handles. Afterwards emergency handles refer to emergency rescue and clear dirt action after the accidents occur. Afterwards the emergency handling result mainly is decided by the establishment situation of the emergency reaction system. Medium-term & long-term trend forecast are important bases to determine the quantity of emergency resources the emergency reaction system needs.This paper mainly establishes Support Vector Machine (SVM) forecast model of vessels oil spill accidents in Shanghai port. Besides, this paper also established Grey forecast model and Projection Pursuit Regression (PPR) forecast model in contrast against the SVM model.The research content summary is as follows:The first chapter mainly elaborated the present situation of research in the ship's oil spill events forecast and its correlation domain, and summarized this paper's research content. The second chapter analyzed the law and the influence factors of the ships' oil spillage in Shanghai port, and forecasted the influence factors with time series model. The third chapter expatiated Statistical Learning Theory (SLT) and the basic principle and algorithm of Support Vector Machine (SVM). Then, the process of SVR algorithm and its application in Shanghai port were revealed in the fourth chapter. The SVR model is rather helpful for the forecast of oil spillage trend. Besides, Grey forecast model and Projection Pursuit Regression (PPR) forecast model were established in contrast against the SVM model in the fifth chapter. Through experiments, we discovered that SVM model's forecast precision was higher. The fifth chapter summarized research's results.