海上风险评估决策系统外文翻译资料

2023-01-14 02:01

Ocean Engineering 38 (2011) 171–176

Contents lists available at ScienceDirect

Ocean Engineering

journal homepage: www.elsevier.com/locate/oceaneng

A decision-making system to maritime risk assessment

Jean-Franc- ois Balmat a,n, Freacute; deacute; ric Lafont a, Robert Maifret b, Nathalie Pessel a

a LSIS, UMR CNRS 6168, University of South-Toulon-Var, B.P. 20132, F-83957 La Garde Cedex, France

b DCNS, Division SIS, B.P. 403, F-83055 Toulon Cedex, France

a r t i c l e i n f o

Article history:

Received 25 May 2010

Accepted 17 October 2010 Editor-in-Chief: A.I. Incecik

Available online 11 November 2010

Keywords:

Maritime risk assessment Maritime safety

Fuzzy risk factor

a b s t r a c t

In this study, we propose a fuzzy approach in order to evaluate the maritime risk assessment applied to safety at sea and more particularly, the pollution prevention on the open sea. The work is based on the decision-making system, named MARISA, presented in Balmat et al. (2009). This system allowed deﬁning a risk factor for each ship according to shiprsquo;s characteristics and weather conditions. In this novel paper, the proposed system takes into account the ship speed evolution and the ship position with respect to maritime shipping lanes is developed. To validate the method, we present an example of results with real data.

Introduction

Nowadays, the maritime risk assessment is an important research theme. Like this, many studies have been realised to identify high risk ships (Degreacute; , 2003; Glansdorp, 2004; Regelink et al., 2004; Van der Heijden et al., 2004; Wang et al., 2004; Sage, 2005; Haj-Salem et al., 2006). The maritime risk modelling is a subject who takes into account several notions such as maritime safety relative to the trafﬁc or the environment protection. The objective of this present work is to evaluate the maritime risk assessment within the framework of the environment protection and more precisely, to prevent the oil pollution.

This paper presents a study about a fuzzy approach for the MAritime RISk Assessment (MARISA) applied to safety at sea which is an extension of previous article proposed by Balmat et al. (2009). In this ﬁrst study, the authors deﬁned an individual risk factor for each ship determined from a fuzzy approach. This factor has been obtained from several static data relative to the ships (age, ﬂag, gross tonnage, number of companies, duration of detention and type) and by considering the meteorological conditions (sea state, wind speed and visibility). To design a ﬂexible decision-making tool, the authors have designed a modular and hierarchical structure. Furthermore, to evaluate the approach, a simulator has been developed and some tests of several ships have been presented. In this context, a risk assessment study of various kinds of ships (bulk carrier, passenger, cargo and oil tanker) has demonstrated the efﬁciency of the approach in risk analysis.

However, to improve this system, it is necessary to take into account some other dynamic parameters such as ships speed and

n Corresponding author. Tel.: 330494142039.

E-mail address: balmat@univ-tln.fr (J.-F. Balmat).

their positions compared with shipping lanes. This is the purpose of this novel article. The paper is organized as follows. Firstly, a brief introduction of risk factor deﬁnition is proposed. In this part, the choice of relevant data, and the proposed fuzzy approach are presented. In Section 3, the architecture of the new MARISA system is described and the three risk factors (static, meteorological and dynamic) allowing to compute a global risk factor are presented. Section 4 details more precisely the design of the fuzzy classiﬁer about the shiprsquo;s speed evolution. In the last section, a simulation from a scenario on a passenger ship which navigates in Mediterranean Sea on the shipping lane Fos-Napoli is depicted. The obtained results have been validated by a human expert and show the interest of this system.

Risk factor deﬁnition

The aim of this work is to design a decision-making system enable to evaluate an individual maritime risk factor in the oil pollution prevention context. For this, in a ﬁrst stage, by performing an expert analysis of the problem, the relevant input data have been selected. In fact, the knowledge acquisition is based on statistical data and information analysis (Degreacute; and Benabbou, 2005) mixed with human expert experience. Therefore, a decision-making system based on fuzzy classiﬁers has been developed to deﬁne the risk factor.

1. Choice of the relevant data

In several papers, T. Degreacute; et al. (2003–2005) deﬁned an individual ship risk index for safety at sea (IRIS). In the papers, the authors analysed the data about accidents listed by the

International Maritime Organization (IMO) for many years. From these works, a decision-making system for the Maritime Risk Assessment has been developed (海上风险评估决策系统

关键词:海上风险评估海上安全，模糊风险因素

目前，海上风险评估是一个重要的研究课题。像这样，许多研究已经实现了识别高风险船舶的目标(Degreacute; , 2003; Glansdorp, 2004; Regelink et al., 2004; Van der Heijden et al., 2004; Wang et al., 2004; Sage, 2005; Haj-Salem et al., 2006)。海上风险评估是一门综合考虑海上交通安全、环境保护等因素的学科。本工作的目的是在环境保护的前提下对海洋风险进行评估，更准确地说，是为了防止石油污染。

然而，为了改进这一系统，需要考虑其他一些动态参数，如船舶速度和它们的位置与航道，这就是这篇文章的目的。论文组织如下，首先，简要介绍了风险因素的定义，在这一部分中，给出了相关数据的选择以及所提出的模糊方法。在第3节中，描述了新的MARISA系统的体系结构，并介绍了计算全球风险因素的三个风险因素(静态、气象和动态)。第四部分详细介绍了船舶航速演化模糊分类器的设计。在最后一节中，模拟了一艘在地中海航行的客轮在foss - napoli航线上的场景。所得结果经人工专家验证，表明了该系统的有效性。

这项工作的目的是设计一个决策系统，能够在防止石油污染的背景下评估一个单独的海上风险因素。为此，在第一个阶段，通过对问题进行专家分析，选择了相关的输入数据。事实上，输入数据是基于统计数据和信息分析(Degre and Benabbou, 2005)，并混合了人类专家的经验。为此，提出了一种基于模糊分类器的风险因素确定方法。

在其他几篇论文中，T. Degreacute; et al. (2003–2005)界定了个别船舶海上安全风险指数(IRIS)。在论文中，作者分析了交通事故的数据国际海事组织(IMO)。在这些工作的基础上，建立了海上风险评估决策系统(Balmat等)（2009）。海洋环境安全取决于船舶的特性、船舶的历史要素、船舶的航行轨迹和气象条件等诸多因素和标准，有关这些标准的数据对于设计一个有效的决策系统是非常重要的。因此，我们列出了可以在我们系统中使用的相关数据。这些输入数据允许计算每艘船的单独风险因素。数据集可从一套数据库(劳埃德船级社、国际海事组织、EQUASIS、巴黎MOU)获得。这样,船舶上的数据特征(类型、国旗、年建造,总吨位,简单或双壳),船的历史元素(公司数量,拘留期间),轨道(船的位置和速度,最后端口和目的地)和气象参数(海洋状态,风速、可见性、晚上或一天)选择(巴黎备忘录,2006;指导海洋,2007)。此外，根据事故率研究(Degre ， 2004)，还考虑了不同类型船舶的环境风险评估。

模糊理论是由Zadeh(1965)提出的。本文定义了几个模糊分类器来设计决策系统。当训练数据集不可用时，可以根据先验知识和专业知识设计分类器。专家能够使用if-then规则、它们的成员函数和规则库为分类提供标签。

2.3.海上风险评估的模糊方法

因此，Liu等人(2005)回顾了决策的不确定性推理方法。讨论了不同的形式技术，并阐述了它们在海事风险评估中的可能应用。他们提出了概率的贝叶斯理论、证据的登普斯特-沙弗理论和模糊集理论。他们比较了这三种方法，并表达了各自的优缺点。实际上，方法的选择取决于几个因素，包括可用性数据(定性和定性信息)、相互关系的复杂程度、“不确定性”的原因以及操作员所需的语言。根据应用程序的类型，可能需要合并这些方法。

其他学者利用模糊逻辑方法研究了海事领域的安全评估问题(Sii et al., 2001, 2004; Liu et al., 2005; Hu et al., 2007; Eleye-Datubo et al., 2008; Yang et al., 2008).在这些文章中，所有作者都证实了模糊逻辑利益对海事风险评估的重要性。

最后，由于大量的输入变量和专家系统的复杂性，有必要定义一个能够简化系统结构的体系结构。为此，定义了模块化和层次化的体系结构。

该体系结构是模块化的，因为规则修改必须易于实现;它是分层的，目的是简化模糊块的开发。该系统由12个简单模块和一个决策逻辑模块组成，如图所示。1.简单模块包含两个输入和一个输出。从静态特性、气象条件、航速演变及其相对航道的位置四个风险因素出发，对每艘船舶进行海事风险评估。

在静态部分，（Degreacute; rsquo;s analysis of the maritime accidents data listed by the International Maritime Organization (IMO) for several years is used (Degreacute; , 2004)。在此基础上，设计了一个模糊结构来确定静态风险因素(Balmat等)（2009）。关于这一因素的输入数据是船旗号、建造年份、公司数量、总吨位、拘留期限、船舶类型和船体类型。静态风险因子由四个模糊分类器估计，两个输入连接在两个加权块上。如Balmat等人所述(2009)，利用模糊分类器船舶的特性，可以对船舶的风险进行第一次估计。这个值取决于建造年份(最近的或旧的)和国旗(使用Paris MOUs的超额因子)。总吨参数是由Degre公司对海上事故的研究确定的。模糊分类器船舶的能力将总吨位加入到船舶特性的评价中。为了完成这项风险研究，有必要评估船舶的历史，考虑到公司的数量(小、中、大)和拘留的时间(无效、中、非常大)。第四个模糊分类器(船舶的潜力)允许直接评估与船舶相对应的风险。最后，用两个不模糊的权重块(船舶势1和船舶势2)考虑了船舶的另外两个特性(船舶类型和船体类型)。根据Degre(2004)，可以根据船舶类型确定事故率值。通过将客轮运价与客轮运价的比值归一化，定义了修正函数。结果如下: 散货船的风险是客船的1.97倍，货船的1.69倍，集装箱船的1.28倍，油轮的1.03倍。因此，块输出船舶的电势1的值乘以这些系数。最后，为了得到静态风险因子，考虑了船体的类型来加权输出船舶的潜力2。双壳油轮设计既有优点也有缺点。在我们的研究中，我们认为简单船体的静态风险系数更高(例如:高1.20倍)。