摘要
储罐区在线监测可以有效反映储罐区作业的运行状态,但储罐区作业的过程变量通常具有较强的相关性,针对单变量的阈值监测不能够体现储罐区运行状态的问题,提出了一种基于无监督学习方法对储罐区多变量进行分析,采用主成分分析法对维度进行归约,基于统计量参数方法进行异常监测。实验结果表明,通过主成分分析使苯物料流程图中原有7维参数信息降到3维,并保留了原有数据中85%以上的参数信息。该方法在储罐区异常运行状态的检测方面表现良好,成功实现了储罐区运行状态的异常监测,研究结果对罐区监测异常有参考价值。
Online monitoring in the storage tank area can effectively reflect the operating status of the operation in the storage tank area,but the process variables of the operation in the storage tank area usually have strong correlation.In view of the problem that the threshold monitoring of a single variable cannot reflect the operating status of the storage tank area,this paper presents the unsupervised learning method to analyze the multiple variables of the storage tank area,and adopts the principal component analysis method to reduce the dimensions.Then anomaly monitoring is carried out based on statistical parameter method.The experimental results show that the original 7-dimensional parameter information in benzene material flow chart is reduced to 3-dimensional by principal component analysis,and more than 85%parameter information in the original data is retained.This method performs well in the detection of abnormal operating status in the storage tank area,and successfully realizes the abnormal monitoring of the operating status of the storage tank area.
出处
《工业控制计算机》
2024年第4期21-22,25,共3页
Industrial Control Computer
关键词
储罐
主成分分析
异常识别
在线监测
storage tank
principal component analysis
anomaly recognition
online monitoring