Experience with an expert system for automated HAZOP analysis
Experience with an expert system for automated HAZOP analysis
复制标题
DOI：
10.1016/0098-1354(96)00270-0
复制
发表时间：
1996
影响因子：
4.3
通讯作者：
Ramesh Vaidhyanathan;V. Venkatasubramanian
中科院分区：
文献类型：
--
作者：
Ramesh Vaidhyanathan;V. Venkatasubramanian
危险与可操作性（HAZOP）分析是化学过程工业中使用最广泛和公认的首选安全分析方法。一个基于模型的框架和一个专家系统称为HAZOPExpert最近已经开发了自动化分析。在酸性水汽提装置和加氢处理装置的案例研究中对HAZOPExpert的性能进行了评估。专家系统成功地模仿人类专家的推理和识别危险类似的HAZOP团队。但是，由于实施了严格的定性推理方法，与团队相比，它产生了大量无法实现的危险后果。为了使用类似于团队的额外定量知识对HAZOPExpert的结果进行过滤和排名，使用过程单元的定量设计和操作规范以及过程材料属性值开发了半定量推理方法。在乙烯厂的案例研究中，使用这种方法获得的后果的数量显着减少。
Hazard and Operability (HAZOP) analysis is the most widely used and recognized as the preferred safety analysis approach in the chemical process industry. A model-based framework and an expert system called HAZOPExpert has been developed recently for automating this analysis. The performance of HAZOPExpert was evaluated on a sour water stripper plant and a hydrotreator plant case studies. The expert system was found to successfully emulate the human experts reasoning and identify the hazards similar to the HAZOP team. But, it generated a large number of unrealizable hazardous consequences compared to the team, due to the strict qualitative reasoning approach implemented. In order to filter and rank the results from HAZOPExpert using additional quantitative knowledge similar to the team, a semi-quantitative reasoning methodology is developed using the quantitative design and operating specifications of the process units, and process material property values. Significant reduction in the number of consequences was obtained using this approach on an ethylene plant case study.