基于PCA权重的化工报警阈值优化

张桂鑫 田文德 靳满满

doi:10.12034/j.issn.1009-606X.216345

过程工程学报 >

2017 , Vol. 17 >Issue 3: 588 - 593

DOI: https://doi.org/10.12034/j.issn.1009-606X.216345

过程与工艺

基于PCA权重的化工报警阈值优化

张桂鑫田文德靳满满

展开

青岛科技大学化工学院，山东青岛 266042

收稿日期: 2016-11-01

修回日期: 2016-12-01

网络出版日期: 2017-06-14

基金资助

国家自然科学基金;山东省自然科学基金

收起

Thresholds Optimization of Chemical Alarms Based on PCA Weight

Guixin ZHANG Wende TIAN Manman JIN

Expand

College of Chemical Engineering, Qingdao University of Science and Technology, Qingdao, Shandong 266042, China

Received date: 2016-11-01

Revised date: 2016-12-01

Online published: 2017-06-14

Fold

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.

关键词： 报警阈值; 优化; 主元分析法; 变量权重; 报警数目

本文引用格式

张桂鑫田文德靳满满 . 基于PCA权重的化工报警阈值优化[J]. 过程工程学报, 2017 , 17(3) : 588 -593 . DOI: 10.12034/j.issn.1009-606X.216345

Abstract

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.

Key words： alarm threshold; optimization; PCA; variable weight; number of alarms

参考文献

[1] EEMUA.Alarm System: A Guide to Design, Management, and Procurement [Z]..2 Ed. London, UK:EEMUA, 2007, :-
[2]JIANG R.Optimization of alarm threshold and sequential inspection scheme[J].Reliability Engineering and System Safety, 2010, 95(3):208-215
[3]JABLONSKI A, BARSZCZ T, BIELECKA M, et al.Modeling of probability distribution functions for automatic threshold calculation in condition monitoring systems[J].Measurement, 2013, 46(1):727-738
[4]FRICKER R D, BANSCHBACH D.Optimizing biosurveillance systems that use threshold-based event detection methods[J].Information Fusion, 2012, 13(2):117-128
[5]MEZACHE A, SOLTANI F.A novel threshold optimization of ML-CFAR detector in Weibull clutter using fuzzy-neural networks[J].Signal Processing, 2007, 87(9):2100-2110
[6]BROOKS R, THORPE R, WILSON J.A new method for defining and managing process alarms and for correcting process operation when an alarm occurs[J].Journal of Hazardous Materials, 2004, 115(1–3):169-174
[7]VERDIER G, HILGERT N, Vila J P.Adaptive threshold computation for CUSUM-type procedures in change detection and isolation problems[J].Computational Statistics & Data Analysis, 2008, 52(9):4161-4174
[8]王佳, 李宏光.过程参数滋扰报警的一类自适应管理策略[J].化工学报, 2015, 66(10):4085-4091
[9]Chen T.On reducing false alarms in multivariate statistical process control[J].Chemical Engineering Research and Design, 2010, 88(4):430-436
[10] 赵劲松, 朱剑锋.基于数据过滤的化工过程重复报警处理策略[J][J].清华大学学报(自然科学版), 2012, 1(03):277-281
[11] 刘恒, 刘振娟, 李宏光.基于数据驱动的化工过程参数报警阈值优化 [J][J].化工学报, 2012, 1(9):2733-2738
[12]田文德, 史晓楠, 王春利, 等.基于数据过滤的化工过程报警优化[J].现代化工, 2015, 35(4):160-163
[13]肖丹卉, 李宏光, 臧灏.化工过程多变量报警阈值优化方法[J].控制工程, 2015, 22(2):246-251
[14]PORTNOY I, MELENDEZ K, PINZON H, et al.An improved weighted recursive PCA algorithm for adaptive fault detection[J].Control Engineering Practice, 2016, 50(May 2016):69-83

Options

文章导航

模态框（Modal）标题

摘要

本文引用格式

Abstract

参考文献