Correlation analysis between looper system and strip width narrowing in hot strip rolling

• Published in: International journal on interactive design and manufacturing Vol. 17; no. 2; pp. 539 - 551
• Main Authors: Jing, Feng-wei, Chen, Zhao-yu, Li, Jie, Guo, Qiang
• Format: Journal Article
• Language: English
• Published: Paris Springer Paris 01.04.2023; Springer Nature B.V
• Subjects: Accuracy; Artificial neural networks; CAE) and Design; Casting; Computer-Aided Engineering (CAD; Control algorithms; Control equipment; Correlation analysis; Data integration; Electronics and Microelectronics; Engineering; Engineering Design; Fault detection; Fault diagnosis; Hot rolling; Hydraulics; Industrial Design; Instrumentation; Mechanical Engineering; Original Paper; Strip; Strip width narrowing; Fault diagnosis; Data fusion; Hot rolled strip; LSTM; Finishing looper
• ISSN: 1955-2513; 1955-2505
• DOI: 10.1007/s12008-022-00940-y

During the hot rolling industrial production process, strip width narrowing is the main quality problem of product. In this paper, it is obtained through experimental analysis that incoming material temperature, roll eccentricity and the control accuracy of looper equipment are the three main factors leading to strip width narrowing in the hot rolling finishing area. In order to deal with the multi-source data, a fault diagnosis model based on Long Short-Term Memory (LSTM) model and data fusion is proposed, and the diagnosis results are verified. The proposed model is a single LSTM model structure using data fusion twice. Compared with multiple LSTM integrated model structure using data fusion once and multiple Convolutional Neural Networks (CNN) + LSTM integrated model structure that does not use data fusion technology, the model proposed in this paper makes a great improvement in the accuracy of fault detection. This model can be applied to hot strip finishing rolling to determine the causes of width narrowing.