论文部分内容阅读
针对石化生产过程的高危性,开发了石化过程在线故障监测系统。通过OPC(OLE for process control)接口从生产现场采集实时数据,采用BP神经网络(back-propagation artificial neural network,BPNN)的模式识别方法,对生产过程进行实时故障监测,及时发现故障工况并提示操作人员采取相应措施,以减小系统运行的风险。BP神经网络的训练数据来自历史数据库,用户根据已发生过的故障工况确定训练数据的时间范围。BP网络模型的各项参数根据多次试验得到。对某工段的10个故障,其故障诊断准确率达到90%以上,具有较高的实时性和准确性。
In view of the high risk of petrochemical production process, an on-line fault monitoring system for petrochemical process has been developed. Real-time data is collected from the production site through the OPC (OLE for process control) interface and real-time fault monitoring of the production process is performed by using a back-propagation artificial neural network (BPNN) pattern recognition method to promptly discover fault conditions and prompt Operators to take appropriate measures to reduce the risk of system operation. BP neural network training data from the historical database, the user according to the fault conditions have occurred to determine the time range of training data. The parameters of BP network model are obtained according to several experiments. For a section of 10 faults, the fault diagnosis accuracy rate of more than 90%, with high real-time and accuracy.