基于PCA权重的化工报警阈值优化  

Thresholds Optimization of Chemical Alarms Based on PCA Weight

作　　者：张桂鑫[1] 田文德[1] 靳满满 

机构地区：[1]青岛科技大学化工学院,山东青岛266042

出　　处：《过程工程学报》2017年第3期588-593,共6页The Chinese Journal of Process Engineering

基　　金：国家自然科学基金资助项目(编号:21576143);山东省自然科学基金资助项目(编号:ZR2013BL008)

摘　　要：为避免海量报警影响正常操作,对报警阈值进行了优化.考虑了变量相关性,利用主元分析法(PCA)分析变量的重要程度,并求出变量权重;结合国际标准,根据变量权重分配变量报警数,优化报警阈值;在平行坐标图上将变量数据及报警阈值可视化,查看正常和异常工况区域,考察各变量之间的变化趋势,提取有效报警值;对某工业原油常减压操作数据进行验证,结果表明优化方法能有效减少报警数.Optimization of alarm threshold to avoid interference to operator due to the massive alarm was studied. Variable weight is firstly addressed by the method of principal component analysis （PCA） considering the relevance and importance degree. Then, the thresholds of alarms are optimized according to variable weight and alarms in combination with the international standards. Finally, variable data and alarm thresholds on the parallel coordinate is visualized and the areas of normal and abnormal operation and the change trend among the variables can be monitored to extract effective alarm values. Case study of crude atmospheric and vacuum operation verification showed that the proposed method can effectively reduce the number of alarms.

关 键 词：报警阈值 优化 主元分析法 变量权重 报警数 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]