Funct. Mater. 2022; 29 (2): 279-290.

doi:https://doi.org/10.15407/fm29.02.279

Prediction of mechanical properties of hot rolled strip based on DBN and composite expectile regression

Siyu Huang1,2, Xiaoxia He1,2, Xin Zhang1,2, Weigang Li3

1College of Science, Wuhan University of Science and Technology, Wuhan 430065, Hubei, China
2Hubei Province Key Laboratory of Systems Science in Metallurgical Process, Wuhan University of Science and Technology, Wuhan 430065, Hubei, China
3Engineering Research Center of Metallurgical Automation and Measurement Technologyof Ministry ofEducation,Wuhan University of Science and Technology, Wuhan 430081, Hubei, China

Abstract: 

In this paper, a DBN-CER19 model (deep belief network and composite expectile regression) is proposed to predict the tensile strength of hot rolled strip. The model takes full advantage that the quadratic loss function used by expectile regression can be solved by standard gradient optimization algorithm, and that DBN can learn more abstract hidden layer information from the underlying data. At the same time, the model combines the CS algorithm (Cuckoo Search) to select the number of DBN hidden layer nodes to improve the network structure. In order to demonstrate the superiority of this method, we use root mean squared error (RMSE), mean absolute percentage error (MAPE) and mean absolute error (MAE)as measurement indicators for empirical analysis. The results of the proposed model on the test set are 21.4381, 2.5251 and 13.5035, respectively. The empirical results show that the DBN-CER19 model has higher prediction performance than previous models such as BP neural network(BPNN), quantile regression neural network(QRNN), expectile regression neural network(ERNN), and DBN.

Keywords: 
deep belief network; composite expectile regression; CS algorithm; mechanical properties of steel
References: 
1. BorisadeS.G., AjibolaO.O., Adebayo A.O., OyetunjiA.. Materials Today: Proceedings, 38, 1133, 2021.
https://doi.org/10.1016/j.matpr.2020.07.134
 
2. WuSiwei, Liu Zhenyu, Zhou Xiaoguang, Shi Nai-an,.J.Iron Steel Res, 28(12), 1, 2016.
 
3. Khalaj G, Azimzadegan T, M Khoeini, Neural Computing & Applications, 23(7-8): 230, 2013.
https://doi.org/10.1007/s00521-012-1182-0
 
4. Chou P Y, Tsai J T, Chou J H. , IEEE Access, 4, 585, 2017.
https://doi.org/10.1109/ACCESS.2016.2521162
 
5. Adel Saoudi et.al., Intern. J.Pressure Vessels Piping, 186, 104153, 2020.
https://doi.org/10.1016/j.ijpvp.2020.104153
 
6. Hu Shixiong. Mechanical Property Prediction of HotRolled Strip Based on Deep Learning. Wuhan University of Science and Technology, 2019.
 
7. Xie Qian, Suvarna Manu, LiJiali, Zhu Xinzhe, CaiJiajia, WangXiaonan. Mater. & Design, 197, 109201, 2021.
https://doi.org/10.1016/j.matdes.2020.109201
 
8. Roger Koenker. , J .Statist. Planning Infer., 143(7),1134, 2013.
https://doi.org/10.1016/j.jspi.2013.03.014
 
9. Tong Tian, Si Yu Huang, Key Eng Mater, 6234(1794), 29, 2021.
 
10. Dadabada Pradeepkumar, Vadlamani Ravi,. Appl. Soft Computing, 58,35, 2017.
https://doi.org/10.1016/j.asoc.2017.04.014
 
11. ZangHaixiang, LiuChongchong, Teng Jun, Kong Bojun, Smart Power, 48(08), 24, 2020.
 
12. Li Bin, Peng Shurong, Peng Junzhe, Huang Shijun, Zheng Guodong. Electr. Power Autom. Equipment, 38(09),15, 2018. .
 
13. Newey W K, Powell J L., J.Econometric Soc., 819, 1987.
https://doi.org/10.2307/1911031
 
14. HU Zongyi, Li Yi, Wan Chuang, Tang Jianyang, Statist.Inform. Forum, 34(04), 19, 2019.
 
15. Jiang Cuixia, Jiang Ming, Xu Qifa, Huang Xue,. Neurocomputing, 247, 73, 2017
https://doi.org/10.1016/j.neucom.2017.03.040
 
16. JiangMin. Expectile regression neural networkmode1 with applications. Hefei University of Technology, 2018.
 
17. Liu Xiaoqian, Zhou Yong, Systems Engin. Theory Practice, 2020, 40(08), 217, 2020.
 
18. Bao Xuexin, Jiang Dan, Yang Xuefeng, Wang Hongmei, Alexandria Engin J, 60(1): 413, 2021
https://doi.org/10.1016/j.aej.2020.09.003
 
19. Yu Jianbo, Liu Guoliang,. Computers Industry, 121, 103262, 2020.
https://doi.org/10.1016/j.compind.2020.103262
 
20. Xing Haixia, Wang Gongming, Liu Caixia, Suo Minghe, Neural Networks, 2021, 133,157, 2021.
https://doi.org/10.1016/j.neunet.2020.10.013
 
21. Abdel Zaher, Ahmedm., Eldeib, Ayman M. Expert Systems Applic. 2016, 46(15): 139, 2016.
https://doi.org/10.1016/j.eswa.2015.10.015
 
22. Zou Hui, Yuan Ming, The Annals Statis., 36(3),1108, 2008.
https://doi.org/10.1214/07-AOS507
 
23. Hinton G E, Salakhutdinov RR., Science, 313(5786), 504, 2006.
https://doi.org/10.1126/science.1127647
 
24. Yang Xinshe, Suash DEB. Cuckoo Search via Levy Flights. 2009World Congress on Nature & Biologically Inspired Computing, 210, 2009.
https://doi.org/10.1109/NABIC.2009.5393690
 
25. Carreira Perpinan M A, Hinton G E, On contrastive divergence learning. Proceedings of artificial intelligence & statistics, 2005.
 
26. Nicolas Le Roux, Bengio Y. , Neural Computation, 20(6): 1631, 2008.
https://doi.org/10.1162/neco.2008.04-07-510
 
27. Larochelle H, Bengio Y, Louradour J, et al, J.Machine Learning Res., 2009(10), l.
 
28. Chen Kui, Laghrouche Salah, Djerdir Abdesslem, ISA Transactions, 113, 175, 2021.
https://doi.org/10.1016/j.isatra.2020.03.012
 
29. Preeti Malhotra and Dinesh Kumar,. J.Intelligent Systems, 28(2), 321, 2019.
https://doi.org/10.1515/jisys-2017-0127
 
30. Bo Yang, ZhongqiWang, YuanYang, YonggangKang, Cheng Li, Adv. Mechan. Engin., 9(6): 145, 2017.
 

Current number: