Dataset for the identification of the classes of events most relevant to the occurrence of individual phases of selected 30 maritime disasters occurred between 1912 and 2019

This dataset supports a new approach to modelling and causal analysis of the course of disasters using maritime events as an example. The proposed approach is based on the integration of network and probabilistic methods. The developed research method is used to confirm the thesis that in the studied population of disasters, in each phase of the process during the course of a disaster, which is distinguished by means of an event network, there are dominant classes of basic events (sets of events with similar characteristics). A proprietary research method, based on an integrated approach using event network analysis and failure tree analysis, is used to prove the thesis. This research makes it possible to distinguish, in each disaster, the different phases of its course: the latent, initiating, escalating, critical, and energy-release parts. The course of 30 maritime disasters with severe consequences is modelled in detail. The events occurred between 1912 and 2019. A total of 608 basic events are identified in the analysed population, enabling the identification and characterisation of 44 classes of events covering human error, inadequacy of equipment and environmental impact. The significance of basic events within each phase is then determined. For this purpose, Birnbaum’s reliability measure, Birnbaum’s structural measure, criticality measure, Veseley-Fussell measure, Lambert’s measure and improvement potential are used. The calculated measures enable the building of rankings on the importance of events in each of the phases of each of the analysed disasters. This makes it possible to draw quantitative and qualitative conclusions on each analysed disaster and each population studied. The dataset consist of files: 1. The development of sea disasters from 1912 to 2019 – dataset. 2. FTA models for disasters at sea for the period 1912-2019 – dataset. 3. Process models for disasters at sea for the period 1912-2019 – dataset.本数据集支持以海事事件为范例开展灾害进程建模与因果分析的全新研究范式。所提出的研究方法基于网络方法与概率方法的融合框架。所开发的研究方法用于验证如下核心论点：在所研究的灾害群体中，灾害进程的每一阶段（通过事件网络进行划分）均存在主导性基础事件类别——即具备相似特征的事件集合。本研究采用自主研发的集成研究方法，结合事件网络分析与故障树分析（Failure Tree Analysis），用以证明该核心论点。 本研究可实现对每一场灾害的进程阶段进行精准划分，包括潜伏阶段、触发阶段、升级阶段、临界阶段与能量释放阶段。本研究对30起后果严重的海事灾害进行了详细建模，其发生时间跨度为1912年至2019年。在所分析的灾害群体中，共识别出608项基础事件，由此可归纳出涵盖人为失误、设备缺陷与环境影响三大范畴的44类事件，并完成其特征刻画。 随后，本研究将确定各阶段内基础事件的重要性权重。为此，将采用伯恩鲍姆可靠性测度、伯恩鲍姆结构测度、临界性测度、韦斯利-富塞尔（Veseley-Fussell）测度、兰伯特（Lambert）测度以及改进潜力指标。通过计算上述指标，可针对每一场灾害的各阶段生成事件重要性排序，进而可为每一场被分析的灾害以及整体研究群体得出定量与定性层面的研究结论。 本数据集包含以下文件： 1. 1912年至2019年海事灾害发展进程数据集 2. 1912年至2019年海事灾害故障树分析模型数据集 3. 1912年至2019年海事灾害进程模型数据集