基于广义SDG的化工过程故障模拟分析

Chemical process fault simulation analyses based on an extended SDG

安全稳定的运行是化工过程的首要目标,因此快速、准确地发现故障成为化工生产企业的一项重要工作。符号有向图(signed directed graph,SDG)能够简洁、有效地描述复杂的因果关系,但传统SDG模型存在一些不足。该文在传统SDG模型的基础上,提出了广义SDG模型,扩展了节点(种类和状态)和边,使其包含的信息更加丰富,基于广义SDG的故障模拟不仅可以定性地给出故障程度,而且通过边的关系减少了虚假解。实例分析表明,基于广义SDG的故障模拟可以提高分析效率,改善过程安全以及提高操作人员判断故障源的能力。

Safe, stable production is the first aim of chemical processing, so fast and accurate fault detection is important in chemical plants. The signed directed graph （SDG） effectively describes complex causal-effect relationships. To overcome the shortages in the traditional SDG model, this paper presents an extended SDG model that extends nodes to kinds, states, and edge relationships with more information. Fault simulations based on the extended SDG give a qualitative expression for the fault extent and reduce spurious solutions through the edge propagation relationships. A case study shows that the extended SDG enhances the analysis efficiency, improves process safety, and improves the operators＇ abilities to identify faults.