基于模拟疫苗的人工免疫系统在氯乙烯聚合间歇过程故障诊断中的应用

摘要

According to the process information, operation procedure and historical data of polyvinylchloride (PVC) industrial process, a dynamic simulation model of vinyl chloride (VCM) suspension polymerization batch process was built with Aspen. Normal simulated vaccine (SV) and fault simulated vaccine which generated by various simulated samples, combining industrial practical data, constructed normal antibody library and fault antibody libraries. This method could deal with the limitation of historical fault samples in fault diagnosis of vinyl chloride polymerization process. Artificial immune system (AIS) based on dynamic time warping (DTW) algorithm was applied to fault diagnosis of this polymerization batch process, which shows great performance.%根据聚氯乙烯工业生产过程的工艺信息、操作规程和历史数据，利用Aspen软件建立了氯乙烯悬浮聚合间歇过程的动态仿真模型，由模拟得到的多个状态的样本数据生成正常模拟疫苗和故障模拟疫苗，结合工业数据建立正常抗体库和故障抗体库，解决了氯乙烯聚合过程故障诊断中故障样本数据缺乏的问题。利用动态时间弯曲算法和人工免疫系统对该聚合反应间歇过程进行故障诊断，具有较好的诊断效果。