Load prediction method of ball mill based on adaptive network

被引：0
|
作者：
Pan F. [1 ,2 ,3 ]
Yin H. [1 ,2 ,3 ,4 ]
Hou X. [1 ,2 ,3 ]
Li S. [1 ,2 ,3 ]
Liu S. [1 ,2 ,3 ]
Jia D. [1 ,2 ,3 ,4 ]
机构：
[1] Key Laboratory of Networked Control Systems, Chinese Academy of Sciences, Shenyang
[2] Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang
[3] Institute for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang
[4] University of Chinese Academy of Sciences, Beijing
来源：
Jisuanji Jicheng Zhizao Xitong/Computer Integrated Manufacturing Systems, CIMS | 2023年 / 29卷 / 10期
关键词：
adaptive network; ball mill load parameters; multiple working conditions; Transformer model; working condition division;
D O I：
10.13196/j.cims.2023.10.001
中图分类号：
学科分类号：
摘要：
The accurate prediction of the load parameters of the ball mill plays a key role in the monitoring and control of the grinding process, and the data drift in the multi-working environment leads to deep learning and other methods that have limited effect on the load prediction of the ball mill. For this reason, a ball mill load forecasting method based on adaptive network was proposed. A classification model of grinding and grading working conditions based on deep correlation alignment was established; then the relative position encoding was introduced into the Transformer, and the attention mechanism was decoupled to directly encode the position information into the attention mechanism, thereby improving the prediction performance. Furthermore, an adaptive network was proposed, which applied the distribution matching rcgularization term to the hidden layer features of Transformer model to learn the common parameters of the hidden state of the model by reducing the distribution difference between different working conditions, so as to improve the generalization ability of the model. A boosting-based approach was employed to learn the importance of hidden states. The test results showed that the proposed adaptive prediction network could significantly improve the accuracy of ball mill load parameter prediction, and the prediction performance was also a-hcad of the comparison method in the face of unknown working conditions. © 2023 CIMS. All rights reserved.
引用
收藏
页码：3229 / 3238
页数：9
相关论文
共 24 条
  • [1] DAI W, LIU Q., CHAI T X., Particle size estimate of grinding processes using random vector functional link networks with improved robustness [J], Neurocomputing, 169, pp. 361-372, (2015)
  • [2] RAMASAMY M, NARAYANAN S S, RAOC D P., Control of ball mill grinding circuit using model predictive control scheme [J], Journal of Process Control, 15, 3, pp. 273-283, (2005)
  • [3] MA Guang, LI Dong, YANG Xiaodong, Gray prediction model-free adaptive control strategy of ball mill [J], Process Automation Instrumentation, 37, 10, pp. 23-26, (2016)
  • [4] LIU Zhuo, CHAI Tianyou, TANG Jian, Soft sensing modeling of mill load parameter based on adaptive feature selection and extraction of multi-scale vibration and acoustic frequency spec-trum, Control and Decision, 34, 12, pp. 2603-2610, (2019)
  • [5] TANG Jian, ZHAO Lyie, YUE Heng, Et al., Soft sensor for ball mill load based on multi-source data feature fusion, Journal of Zhejiang University : Engineering Science, 44, 7, pp. 1406-1413, (2010)
  • [6] YAN Gaowei, HE Min, TANG Jian, Et al., Soft sensor of wet ball mill load based on maximum mean discrepancy multi-source domain transfer learning [J], Control and Decision, 33, 10, pp. 1795-1800, (2018)
  • [7] TIAN Huixin, LI Kun, MENG Bo, An incremental learning ensemble algorithm for soft sensor modeling, Control and Decision, 30, 8, pp. 1523-1526, (2015)
  • [8] SHAO W M, TIAN X M., Semi-supervised selective ensemble learning based on distance to model for nonlinear soft sensor developmentJ, Neurocomputing, 222, pp. 91-104, (2017)
  • [9] LI K J, ZHANG J X., CHEN X., Building's hourly electrical load prediction based on data clustering and ensemble learning strategy[J], Energy & Buildings, 261, (2022)
  • [10] RANG Shouqiang, LIU Wanghui, WANG Yujing, Et al., Fault diagnosis method of rolling bearing under varying loads based on deep online transfer, Control and Decision, 37, 6, pp. 1521-1530, (2022)
123