Polymerization kettle equipment fault diagnosis based on MPA optimized MKL-FSVDD model

被引：0
|
作者：
Li G. [1 ]
Cai S. [1 ]
Li D. [2 ]
Zhang X. [1 ]
Jia Y. [1 ]
Ning Z. [1 ]
机构：
[1] Engineering Research Center of the Ministry of Education for Intelligent Control System and Intelligent Equipment, Yanshan University, Qinhuangdao
[2] School of Optoelectronic Science and Engineering, University of Electronic Science and Technology of China, Chengdu
来源：
High Technology Letters | 2022年 / 32卷 / 04期
关键词：
Fault diagnosis; Fuzzy membership degree; Marine predators algorithm (MPA); Multiple kernel learning (MKL); Polymerization kettle;
D O I：
10.3772/j.issn.1002-0470.2022.04.006
中图分类号：
学科分类号：
摘要：
In order to solve the problems that the data of chemical process industry has strong non-linearity, easy to be affected by noise and the fault is multi-fault classification, a fault diagnosis method based on marine predators algorithm (MPA) optimized multi-kernel learning and fuzzy support vector machine data description (MKL-FSVDD) is proposed. The multi-kernel function constructed by MKL makes up for the limitation of single kernel function, and has strong adaptability to nonlinear fault data classification. MPA is introduced to optimize the kernel parameters of MKL-FSVDD model efficiently, and the problem of kernel parameter selection can be solved. The effective performance of MPA-MKL-FSVDD model for fault diagnosis is verified through the control experiment on the TE data platform. Finally, the algorithm is applied to the polyvinyl chloride (PVC) polymerization reaction, and the historical data set of the 70m3 polymerization kettle is used for simulation verification. The results show that the proposed method makes full use of the data information of the complex sample set, and the optimal solution can be obtained quickly and stably in the parameter optimization stage, which guarantes the efficiency and accuracy of fault classification. © 2022, Editorial Department of the Journal of Chinese High Technology Letters. All right reserved.
引用
收藏
页码：379 / 391
页数：12
相关论文
共 33 条
  • [1] 27, (2017)
  • [2] (2019)
  • [3] TRITTENBACH H, BOHM K, ASSENT I., Active learning of SVDD hyperparameter values, 2020 IEEE 7th International Conference on Data Science and Advanced Analytics (DSAA), pp. 109-117, (2020)
  • [4] (2018)
  • [5] 14, 1, (1993)
  • [6] WANG J S, LI S X, GAO J., SOM neural network fault diagnosis method of polymerization kettle equipment optimized by improved PSO algorithm, Scientific World Journal, 2014, 7, pp. 1-12, (2014)
  • [7] WU D, ZHAO J., Process topology convolutional network model for chemical process fault diagnosis, Process Safety and Environmental Protection, 150, pp. 93-109, (2021)
  • [8] CHANG J, WANG G C, WANG Y, Et al., Based on genetic algorithm and BP network to fault diagnosis of fermentation process, 2008 Chinese Control and Decision Conference, pp. 578-581, (2008)
  • [9] WEN S, WANG J S, GAO J, Et al., Fault diagnosis strategy of polymerization kettle equipment based on support vector machine and cuckoo search algorithm, Engineering Letters, 25, 4, pp. 2306-2314, (2017)
  • [10] TAX D M, DUIN R P W., Support vector domain description, Machine Learning, 54, 1, pp. 45-66, (2004)
1234