SCI和EI收录∣中国化工学会会刊

中国化学工程学报 ›› 2023, Vol. 59 ›› Issue (7): 231-239.DOI: 10.1016/j.cjche.2022.12.007

• Full Length Article • 上一篇    下一篇

Minimax entropy-based co-training for fault diagnosis of blast furnace

Dali Gao1, Chunjie Yang1, Bo Yang2, Yu Chen3, Ruilong Deng1   

  1. 1. The State Key Laboratory of Industrial Control Technology, College of Control Science and Engineering, Zhejiang University, Hangzhou 310027, China;
    2. Jianwei Digital Sphere Co., Ltd, Lianyungang 222113, China;
    3. Jiangsu Binxin Iron and Steel Group Co., Ltd, Chongqing 401220, China
  • 收稿日期:2022-09-21 修回日期:2022-12-21 出版日期:2023-07-28 发布日期:2023-10-14
  • 通讯作者: Chunjie Yang,E-mail:cjyang999@zju.edu.cn
  • 基金资助:
    This work was supported in part by the National Natural Science Foundation of China (61933015), in part by the Central University Basic Research Fund of China under Grant K20200002 (for NGICS Platform, Zhejiang University).

Minimax entropy-based co-training for fault diagnosis of blast furnace

Dali Gao1, Chunjie Yang1, Bo Yang2, Yu Chen3, Ruilong Deng1   

  1. 1. The State Key Laboratory of Industrial Control Technology, College of Control Science and Engineering, Zhejiang University, Hangzhou 310027, China;
    2. Jianwei Digital Sphere Co., Ltd, Lianyungang 222113, China;
    3. Jiangsu Binxin Iron and Steel Group Co., Ltd, Chongqing 401220, China
  • Received:2022-09-21 Revised:2022-12-21 Online:2023-07-28 Published:2023-10-14
  • Contact: Chunjie Yang,E-mail:cjyang999@zju.edu.cn
  • Supported by:
    This work was supported in part by the National Natural Science Foundation of China (61933015), in part by the Central University Basic Research Fund of China under Grant K20200002 (for NGICS Platform, Zhejiang University).

摘要: Due to the problems of few fault samples and large data fluctuations in the blast furnace (BF) ironmaking process, some transfer learning-based fault diagnosis methods are proposed. The vast majority of such methods perform distribution adaptation by reducing the distance between data distributions and applying a classifier to generate pseudo-labels for self-training. However, since the training data is dominated by labeled source domain data, such classifiers tend to be weak classifiers in the target domain. In addition, the features generated after domain adaptation are likely to be at the decision boundary, resulting in a loss of classification performance. Hence, we propose a novel method called minimax entropy-based co-training (MMEC) that adversarially optimizes a transferable fault diagnosis model for the BF. The structure of MMEC includes a dual-view feature extractor, followed by two classifiers that compute the feature’s cosine similarity to representative vector of each class. Knowledge transfer is achieved by alternately increasing and decreasing the entropy of unlabeled target samples with the classifier and the feature extractor, respectively. Transfer BF fault diagnosis experiments show that our method improves accuracy by about 5% over state-of-the-art methods.

关键词: Co-training, Fault diagnosis, Blast furnace, Minimax entropy, Transfer learning

Abstract: Due to the problems of few fault samples and large data fluctuations in the blast furnace (BF) ironmaking process, some transfer learning-based fault diagnosis methods are proposed. The vast majority of such methods perform distribution adaptation by reducing the distance between data distributions and applying a classifier to generate pseudo-labels for self-training. However, since the training data is dominated by labeled source domain data, such classifiers tend to be weak classifiers in the target domain. In addition, the features generated after domain adaptation are likely to be at the decision boundary, resulting in a loss of classification performance. Hence, we propose a novel method called minimax entropy-based co-training (MMEC) that adversarially optimizes a transferable fault diagnosis model for the BF. The structure of MMEC includes a dual-view feature extractor, followed by two classifiers that compute the feature’s cosine similarity to representative vector of each class. Knowledge transfer is achieved by alternately increasing and decreasing the entropy of unlabeled target samples with the classifier and the feature extractor, respectively. Transfer BF fault diagnosis experiments show that our method improves accuracy by about 5% over state-of-the-art methods.

Key words: Co-training, Fault diagnosis, Blast furnace, Minimax entropy, Transfer learning