Computer-aided Operability Study

在本研究中，于1997年至1999年与芬兰的VTT自动化公司合作开发了计算机辅助HAZOP的面向对象方法。HAZOP系统主要由知识库（工厂专用知识库、通用知识库）和推理引擎组成。HAZOP系统中的通用知识包括部件故障模型和建议行动知识。在HAZOP系统中，变量偏差与部件失效模式之间的因果关系被建模为部件故障模型。所需的行动分为起因行动和结果行动。对起因和后果的行动的认识分别包括预防和发现、缓解和发现。这些知识以独立于工艺的方式发展，适用于各种工艺流程图。植物特异性知识包括有关材料、反应和植物结构（P和id）的信息。植物特定知识因植物而异，必须由用户提供。用户通过图形用户界面与系统交互。用户可以在HAZOP系统图形化编辑器中方便地绘制进程排列组件图标的P和id。一旦将P和id输入到系统中，相应的组件故障将以适当的方式在内部自动连接。HAZOP分析是通过在P和id中指定过程单元或管道中的过程变量偏差来启动的。通过搜索知识库，对设备的各个过程变量进行顺序研究，用关键词描述偏差。HAZOP系统采用面向对象的体系结构，使用专家系统外壳G2实现。

In this research, an object-oriented approach for computer-aided HAZOP has been developed by collaboration with VTT Automation in Finland from 1997 to 1999.The HAZOP System consists basically of the knowledge bases (the plant sepcific knowledge base, the generic knowledge base ) and the inference engine. The generic knowledge in the HAZOP System consists of components malfunction models and suggested action knowledge. Causal relationships between variable deviations and failure modes for components are modeled as component malfunction models in the HAZOP System. Actions required are divided into actions for the cause and actions for the consequence. Knowledge of actions for the cause and the consequence consist of prevention and detection, mitigation and detection respectively. This knowledge is developed in a process-independent manner and is applicable to wide variety of process flow sheets.The plant specific knowledge consists of information about materials, reactions, and the plant structure (P and IDs). The plant specific knowledge varies from plant to plant and has to be provided by the user. The user interacts with the system through the graphical user interface. The user can easily draw the P and IDs of the process arranging component icons in the HAZOP System graphical editor. Once the P and IDs is input into the system, the corresponding component malfunction gets connected automatically internally in the appropriate manner.HAZOP analysis is initiated by specifying a process variable deviation in a process unit or pipeline in the P and IDs. Each process variable of equipment is studied in sequence by searching the knowledge base, using key words to describe deviations.The HAZOP System for this work is implemented in an object-oriented architecture using expert system shell G2.