石化装置异常工况监测预警系统研究

Abnormal Situations Monitoring and Early Warning System of Petrochemical Process

针对石化装置的工艺生产特点,以生产过程的"数据"与"知识"为基础,进行多源信息融合,研究了用于信号处理的基于提升格式的数据滤波技术、基于灰色时序模型的参数预警技术、基于平行坐标的报警优化技术;用于异常工况监测预警的基于专家系统与人工神经网络的异常工况识别技术、基于机理模型的装置物料与能量平衡分析技术、基于状态空间分析的运行状态监测预警技术;搭建了异常工况分析测试与试验平台,用于异常工况监测预警系统的可靠性与完备性分析。工业应用表明,该系统对减少误操作具有良好的效果。

According to the characteristics of petrochem- ical plant production process in the production process based on ＂data＂ and ＂knowledge＂, through the multi - source information fusion, this paper studied data filter technology in signal processing based on lifting scheme, parameter early -warning technology based on gray se- quence simulation, alarm optimization technology based on parallel coordinates; early warning system for abnor- mal condition monitoring based on expert system and ar- tificial neural network identification technology, plant material and energy balance analysis technology based on the mechanism model, early warning technology based on state space analysis of running state monito- ring. And set up a platform of abnormal condition anal- ysis and test for reliability and completeness in monito- ring and early warning system for abnormal conditions a- nalysis. Industrial application shows that the system has good effect on reducing wrong operation.