SCI和EI收录∣中国化工学会会刊

中国化学工程学报 ›› 2020, Vol. 28 ›› Issue (7): 1875-1883.DOI: 10.1016/j.cjche.2020.05.003

• Process Systems Engineering and Process Safety • 上一篇    下一篇

Identification of abnormal conditions in high-dimensional chemical process based on feature selection and deep learning

Wende Tian1, Zijian Liu1, Lening Li2, Shifa Zhang1, Chuankun Li3   

  1. 1 College of Chemical Engineering, Qingdao University of Science&Technology, Qingdao 266042, China;
    2 Qingdao NOVAN Electromechanical Technology, Qingdao 266100, China;
    3 State Key Laboratory of Safety and Control for Chemicals, SINOPEC Qingdao Research Institute of Safety Engineering, Qingdao 266071, China
  • 收稿日期:2019-12-20 修回日期:2020-04-02 出版日期:2020-07-28 发布日期:2020-08-31
  • 通讯作者: Wende Tian
  • 基金资助:
    Financial support for carrying out this work was provided by the Shandong Provincial Key Research and Development Program (2018YFJH0802).

Identification of abnormal conditions in high-dimensional chemical process based on feature selection and deep learning

Wende Tian1, Zijian Liu1, Lening Li2, Shifa Zhang1, Chuankun Li3   

  1. 1 College of Chemical Engineering, Qingdao University of Science&Technology, Qingdao 266042, China;
    2 Qingdao NOVAN Electromechanical Technology, Qingdao 266100, China;
    3 State Key Laboratory of Safety and Control for Chemicals, SINOPEC Qingdao Research Institute of Safety Engineering, Qingdao 266071, China
  • Received:2019-12-20 Revised:2020-04-02 Online:2020-07-28 Published:2020-08-31
  • Contact: Wende Tian
  • Supported by:
    Financial support for carrying out this work was provided by the Shandong Provincial Key Research and Development Program (2018YFJH0802).

摘要: Identification of abnormal conditions is essential in the chemical process. With the rapid development of artificial intelligence technology, deep learning has attracted a lot of attention as a promising fault identification method in chemical process recently. In the high-dimensional data identification using deep neural networks, problems such as insufficient data and missing data, measurement noise, redundant variables, and high coupling of data are often encountered. To tackle these problems, a feature based deep belief networks (DBN) method is proposed in this paper. First, a generative adversarial network (GAN) is used to reconstruct the random and non-random missing data of chemical process. Second, the feature variables are selected by Spearman's rank correlation coefficient (SRCC) from high-dimensional data to eliminate the noise and redundant variables and, as a consequence, compress data dimension of chemical process. Finally, the feature filtered data is deeply abstracted, learned and tuned by DBN for multi-case fault identification. The application in the Tennessee Eastman (TE) process demonstrates the fast convergence and high accuracy of this proposal in identifying abnormal conditions for chemical process, compared with the traditional fault identification algorithms.

关键词: Chemical process, Deep Belief Networks, Fault identification, Generative Adversarial Networks, Spearman Rank Correlation

Abstract: Identification of abnormal conditions is essential in the chemical process. With the rapid development of artificial intelligence technology, deep learning has attracted a lot of attention as a promising fault identification method in chemical process recently. In the high-dimensional data identification using deep neural networks, problems such as insufficient data and missing data, measurement noise, redundant variables, and high coupling of data are often encountered. To tackle these problems, a feature based deep belief networks (DBN) method is proposed in this paper. First, a generative adversarial network (GAN) is used to reconstruct the random and non-random missing data of chemical process. Second, the feature variables are selected by Spearman's rank correlation coefficient (SRCC) from high-dimensional data to eliminate the noise and redundant variables and, as a consequence, compress data dimension of chemical process. Finally, the feature filtered data is deeply abstracted, learned and tuned by DBN for multi-case fault identification. The application in the Tennessee Eastman (TE) process demonstrates the fast convergence and high accuracy of this proposal in identifying abnormal conditions for chemical process, compared with the traditional fault identification algorithms.

Key words: Chemical process, Deep Belief Networks, Fault identification, Generative Adversarial Networks, Spearman Rank Correlation