基于时间权重的船舶事故组合预测分析

发布时间：2018-03-04 14:32

本文选题：福建沿海　切入点：船舶事故　出处：《集美大学》2017年硕士论文　论文类型：学位论文

【摘要】：如今科技的不断更新,船舶的设备越来越先进了,理论上讲,船舶发生事故的可能性应该越来越低才是,但是近几年船舶事故还是保持在一个较高的趋势水平,通过研究发现,导致事故居高不下的原因是由于各方面的管理不善导致的,所以加强船舶的安全管理是很有必要的。福建沿海水域地处台湾海峡,是国际船舶航行的一条重要通道,保证福建沿海水域的交通安全至关重要,所以对该水域的船舶事故调查研究就是一项必不可少的环节。通过分析过去几年内该水域的船舶事故发展规律,分析预测未来某一时间该水域船舶事故的趋势及发生事故的主要因素,对引起事故的原因积极采取相应的措施,避免同样的原因导致多起事故的发生,尽可能的降低船舶事故的发生量,提供过往船舶一条安全的海上航道。由于船舶事故的随机性、复杂性等特点交织在一起,所以本文以定性分析和定量分析相结合的分析方法。从福建省各个地方海事局收集整理船舶事故资料进行统计,通过分析福建沿海的水域情况和事故的统计规律,结合典型事故分析,归纳出导致福建沿海船舶事故的主要因素。然后分析福建沿海2001~2014年船舶事故数据特点,总结福建沿海船舶事故的特性和规律,借鉴国内外学者在船舶事故预测上的研究方法,发现灰色理论的GM(1,1)预测模型、ARIMA时间序列预测模型及BP神经网络预测模型在船舶事故的预测上具有很强的适用性。并对福建沿海船舶事故数据进行单一预测,运用其结果可以有效的分析事故数据的变化。为了进一步提高模型的预测准确性,针对福建沿海的船舶事故数据建立时间权重的组合预测模型,给出时间权重应该满足的条件,建立时间权重函数,并且推导时间组合权重的计算公式。通过验证表明,以时间权重为基础的组合预测模型在其预测精度方面较单一的预测方法不但精度高,而且具有较好的普遍适用性。最后,提出加强福建沿海水上安全管理的措施,为海事管理部门及时的控制船舶交通事故的发生量提供科学有效的决策依据。
[Abstract]:With the development of science and technology, the equipment of ships is getting more and more advanced. Theoretically, the possibility of ship accidents should be lower and lower. However, in recent years, ship accidents have remained at a high trend level. The reasons for the high level of accidents are due to mismanagement in various aspects, so it is very necessary to strengthen the safety management of ships. The coastal waters of Fujian are located in the Taiwan Strait and are an important passage for international ship navigation. It is very important to ensure the traffic safety in the coastal waters of Fujian, so the investigation and study of ship accidents in this area is an essential link. By analyzing the development law of ship accidents in this area in the past few years, This paper analyzes and predicts the trend and main factors of ship accidents in this water area in a certain time in the future, and takes corresponding measures to prevent the same causes from causing many accidents. Reduce the number of ship accidents as far as possible, and provide a safe sea channel for passing ships. Because of the randomness and complexity of ship accidents, So this paper combines qualitative analysis with quantitative analysis, collects and collates ship accident data from various local maritime authorities in Fujian Province, and analyzes the situation of Fujian coastal waters and the statistical law of accidents. Combined with the analysis of typical accidents, the main factors leading to the marine accidents along the coast of Fujian are summed up, and then the characteristics of the data of ship accidents in the coastal areas of Fujian from 2001 to 2014 are analyzed, and the characteristics and rules of the accidents in the coastal areas of Fujian are summarized. Drawing lessons from the research methods of domestic and foreign scholars on ship accident prediction, It is found that Arima time series prediction model and BP neural network prediction model based on grey theory have strong applicability in ship accident prediction. In order to improve the prediction accuracy of the model, the combined forecasting model of the time weight is established for the ship accident data in Fujian coastal area, and the conditions that the time weight should be satisfied are given. The time weight function is established, and the calculation formula of time combined weight is deduced. The verification shows that the combined prediction model based on time weight is more accurate than the single prediction method. Finally, the measures to strengthen the management of water safety along the coast of Fujian Province are put forward to provide a scientific and effective decision basis for the maritime management department to control the occurrence of ship traffic accidents in a timely manner.
【学位授予单位】：集美大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U698.6

【参考文献】