Collaborative Research: Speaking the Truth in Maritime Risk Assessment

在过去6年中，主要调查人员制定了一种评估与海运有关的风险的方法。该方法利用实时交通系统的动态模拟、专家判断启发模型和概率风险评估(PRA)技术。事实证明，该方法在美国三个主要的海运港口很有用，即新奥尔良港(路易斯安那州的密西西比河)、瓦尔迪兹港(阿拉斯加州的普林斯维利安湾)和西雅图港(华盛顿州的普吉特湾)。这些研究中有两项关注乘客安全，另一项关注石油运输风险。1998年，国家研究理事会对该方法进行了评估，并得出结论认为，海事风险评估方法缺乏不确定性。威廉王子湾风险评估是在使用概率风险评估方法来评估PWS大型油轮运输石油的安全性方面向前迈出的重要一步。由于数据非常有限，分析结果和结果结论并不可靠，必然是不确定的[国家研究委员会，1998]。该NSF项目将通过开发一个连贯的理论框架来解决不确定性，使海洋风险评估方法实现其承诺。事实是，我们并不确定。因此，说实话意味着我们用概率曲线而不是固定点估计来表达我们的分析结果。[例如，见卡普兰，1997年]。这项研究的学术价值来自于开发了一个总体贝叶斯框架，用于处理系统状态模拟与可用数据和专家判断相结合以评估风险和风险干预有效性时的不确定性。拟议工作的更广泛影响主要来自其对海上事故风险以外的领域的适用性，例如海上安全风险(故意事件而不是事故)。将制定的框架和方法将适用于其他运输方式，如航空或道路安全。除了航空安全和事故风险，这项技术将直接转移到由于我们国家机场日益严重的交通拥堵而导致的日益严重的跑道入侵问题上。

Over the past 6 years the principal investigators have developed a methodology for assessing the risk associated with maritimetransportation. The methodology utilizes dynamic simulation of themaritime transportation system, expert judgment elicitation models andprobabilistic risk assessment (PRA) techniques. The approach has provenuseful in three major maritime ports in the United States i.e. Port ofNew Orleans (Mississippi River in Louisiana), Port Valdez (PrinceWilliam Sound in Alaska) and Port Seattle (Puget Sound in WashingtonState). Two of these studies focused on passenger safety, while theother focused on oil transportation risk. The National Research Council in1998 assessed the methodology and concluded that representation of uncertainty in the maritime risk assessment method is lacking. "The Prince William Sound Risk Assessment is an important step forward in using probabilistic risk assessment methods to assess the safety of transporting oil in large tankersin PWS. Because the data were very limited, the analysis results and the resultingconclusions are not robust and are necessarily uncertain [National ResearchCouncil, 1998]."This NSF project will allow the maritime risk assessment method to liveup to its promise by developing a coherent theoretical framework toaddress uncertainty. The truth is that we are uncertain. Therefore,speaking the truth implies that we express our analysis results in termsof probability curves rather than fixed points estimates. [See, e.g.,Kaplan, 1997]. The intellectual merit of the research stems from thedevelopment of an overarching Bayesian framework for addressinguncertainty when simulation of systems states is combined with availabledata and expert judgment to assess risk and risk interventioneffectiveness. The broader impact of the proposed work is primarilydrawn from its applicability to areas other than maritime accident risksuch as e.g. maritime security risk (intentional events as opposed toaccidents). The framework and methodologies to be developed will beapplicable to other transportation modes, such as aviation or roadsafety. Aside from aviation security and accident risk, the techniquewill be directly transferable to the ever-increasing problem of runwayincursions as a result of increased traffic congestion at our nationalairports.