复杂过程报警系统建模与优化

报警是及时发现生产过程中异常情况的直接手段，对保障生产安全至关重要。针对过程工业报警系统存在的报警泛滥、处置低效等问题，本课题研究报警系统涉及的报警数据、过程数据和工艺知识统一的描述方法，建立报警系统的理论框架，构建复杂过程系统的层次报警因果链图模型；建立故障的可检测性和可分辨性准则，结合过程变量的分层分组结构和过程故障与报警信息的传播概率，研究报警变量选择、单变量和多变量报警阈值的优化设计方法以及报警信息高效的层次可视化的方法；以TE过程和典型石化系统乙烯生产过程为研究应用对象，形成综合过程知识和数据驱动的复杂过程报警系统建模与优化的新理论与新方法，研发报警系统层次因果链图模型生成和报警系统信息可视化的软件原型，为工业应用提供基础。本项目研究成果具有重要的理论意义和实用价值，具有广阔的应用前景。

Alarming is a direct way of promptly detecting abnormal situations in production processes, it is crucial to ensure safety in production. In view of the existing problems of alarm flood and the low disposal efficiency in the alarm systems of industrial processes, this project will study on the unified description method for alarm systems involving alarm data, process data and process knowledge. The theoretical framework of alarm system will be established and the model of hierarchical alarm cause-effect chain for complex processes will be proposed. The criterions of detectability and resolution are given for fault. Combining hierarchical structures of process variables with transmission probability of process fault and alarm information, the choice of alarm variables, the optimal design methods of single and multi-variables alarm threshold as well as the methods of efficient hierarchical visualization for alarm information will be studied. The TE process and the ethylene production process are selected to test the validity of proposed algorithms using process system simulation. The final aim is to form a new theory and methods of system modelling and optimization for alarm system of complex industrial process based on the integration of process knowledge and data-driven. The software prototype systems of hierarchical alarm cause-effect chain model and information visualization of alarm systems will be developed. The research achievements of this project will provide the basis for industrial applications, which will be of great theoretical significance and practical value with broad application prospects.