化工进展 ›› 2020, Vol. 39 ›› Issue (4): 1267-1272.DOI: 10.16085/j.issn.1000-6613.2019-1269

• 化工过程与装备 • 上一篇    下一篇

基于多元统计方法的过程单元缓变故障识别

马方圆1(),林德溪2,许明阳2,王璟德1,孙巍1()   

  1. 1.北京化工大学化学工程学院,北京 100029
    2.中化泉州石化有限公司,福建 泉州 362103
  • 收稿日期:2019-08-06 出版日期:2020-04-05 发布日期:2020-04-28
  • 通讯作者: 孙巍
  • 作者简介:马方圆(1996—),男,硕士研究生,研究方向为化工过程监测。E-mail:2017210051@mail.buct.edu.cn

Early identification of small shift in process unit based on multivariate statistical method

Fangyuan MA1(),Dexi LIN2,Mingyang XU2,Jingde WANG1,Wei SUN1()   

  1. 1.College of Chemical Engineering, Beijing University of Chemical Technology, Beijing 100029, China
    2.Sinochem Quanzhou Petrochemical Co. , Ltd, Quanzhou 362103, Fujian, China
  • Received:2019-08-06 Online:2020-04-05 Published:2020-04-28
  • Contact: Wei SUN

摘要:

在实际化工生产过程中存在一些缓变故障,在发生的初期过程偏离正常工况的程度较少,且受生产数据噪声的影响,不易被传统过程监测方法及时发现。本文针对缓变故障的特点,提出了一种基于偏最小二乘法-主元分析法(PLS-PCA)的过程监测方法。首先利用偏最小二乘法(PLS)回归提取出各变量之间的关系,通过获取变量实测值与回归预测值之间的误差,以放大装置运行状态与预设状态之间的偏差,在此基础上建立基于主元分析法(PCA)的过程监测模型,实现了对缓变故障的早期识别。该过程监测模型被应用在某制氢装置预转化反应器上,结果表明该方法对缓变故障具有较好的早期识别效果,能够比工程师提前13h,比基于传统PCA的过程监测模型提前8h。

关键词: 系统工程, 主元分析法, 偏最小二乘法, 缓变故障, 过程监测

Abstract:

There are some small shifts in the actual chemical production process, which may be resulted by crucial process failure or cause serious problems later on. In the initial stage of this small shift, the process deviation from the normal working conditions is negligible with the appearance of data noise, and it is hard to capture by traditional process monitoring methods. In this work, a process monitoring method based on partial least squares-principal component analysis(PLS-PCA) was proposed. First, the correlation among variables was extracted by PLS method regression. By obtaining the error between the measured value and the predicted value of measurements, the process deviation between the operating state and the pre-set state was amplified. Based on this deviation, the PCA-based process monitoring model was established to realize the early identification of the small shift. A pre-reforming reactor of hydrogen production unit was investigated. The results showed that the small shift, the sulfur content in the diluted steam exceeded the normal level which resulted in catalyst poisoning, could be found at least 13h earlier than human operator and 8h earlier than the traditional PCA-based process monitoring model which could significantly reduce the loss of industrial production.

Key words: systems engineering, principal component analysis, partial least square (PLS), small shift, process monitoring

中图分类号: 

版权所有 © 《化工进展》编辑部


本系统由北京玛格泰克科技发展有限公司设计开发 技术支持:support@magtech.com.cn