正在加载图片...
何飞等:基于多模态和加权支持向量机的热轧轧制力智能预报 ·521* 表3轧制力计算结果 Table3 Computed result of rolling force 机理模型 聚类前加权支持向量机 数据接触面变形抗力/应力 聚类后加权支持向量机 张力 轧制力实 轧制力计算相对误 修正轧制 修正相对 修正轧制 修正相对 序列 积/m2(Nmm2)系数 系数际值MN 值/MN 差/% 力/MN 误差/% 力/MN 误差/% 1 0.15 125.44 1.098 0.986 24.70 20.37 17.5 26.78 8.4 25.28 2.3 2 0.17 134.861.0740.979 27.36 24.11 11.8 25.45 7.0 28.85 5.5 3 0.22 141.25 1.069 0.981 28.43 32.59 14.6 30.63 7.7 29.1 2.4 4 0.21129.381.0780.963 25.85 28.21 9.1 27.69 7.1 26.71 3.3 5 0.19 132.421.067 0.976 28.91 26.2 9.4 31.17 7.8 30.47 5.4 6 0.15 127.611.0720.98 23.64 20.11 14.9 26.08 10.3 24.41 3.3 7 0.13 138.59 1.0850.991 23.19 19.37 16.5 24.97 7.7 23.55 1.6 0.17 135.471.0870.985 27.27 24.66 9.6 29.24 7.2 28.31 3.8 9 0.18 128.361.0770.978 26.74 24.34 9.0 25.13 6.0 25.89 3.2 10 0.2 131.931.0910.986 25.58 28.38 11.0 27.81 8.7 25.97 1.5 平均值 12.3 7.8 3.2 差降至5%以下,多组数据的平均值为3.2%,满足现 28(10):969) 场生产要求 [5]Chen Z M,Luo F,Huang X H,et al.The prediction of rolling force based on chaotic optimized support vector machine.Control Deei,2009,24(6):808 参考文献 (陈治明,罗飞,黄晓红,等.基于混沌优化支持向量机的轧 [Sun Y K.Model and Control of Hot Strip Rolling.Beijing:Metal- 制力预测.控制与决策,2009,24(6):808) lurgical Industry Press,2007 [6]Cortes C,Vapnik V.Support vector networks.Mach Learn, (孙一康.带钢热连轧的模型与控制.北京:治金工业出版 1995,20(3):273 社,2007) 7]Smola A J.Scholkopf B.A tutorial on support vector regression 2]Hong Y,Tang L X,Zhang YY.Optimization of rolling force of Stat Comput,2004,14(3):119 hot rolling by using data subspace PIS modeling technique.Con- [8]Kaufman L.Rousseeuw P J.Finding Groups in Data:an intro- trol Decis,2014,29(7):1199 duction to Cluster Analysis.New York:John Wiley Sons Press, (洪悦,唐立新,张颜颜.基于数据子空间PL$建模技术的热 1990 轧轧制力优化设定.控制与决策,2014,29(7):1199) ]John C P.Using Analytic OP and Sparseness to Speed Training of B3]Zhou FQ,Cao JC,Zhang J,et al.Online calculation model of Support Vector Machines.Cambridge:MIT Press,1999 rolling force for a tandem cold rolling mill.J Univ Sci Technol Bei- [10]Suykens JA K.Least squares support vector machines for classi- ng,2006,28(9):859 fication and nonlinear modeling.Neural Netcork World,2000, (周富强,曹建国,张杰,等。冷连轧机轧制力在线计算模型 10(1):29 北京科技大学学报,2006,28(9):859) [11]Suykens J A K,De Brabanter J,Lukas L,et al.Weighted least 4]Liu HQ,Tang D.Yang Q,et al.Rolling force prediction model squares support vector machine:robustness and sparse approxi- of a multi roll cold tandem mill by fuzzy cerebellum model articula- mation.Neurocomputing,2002,48(1):85 tion controller.J Unin Sci Technol Beijing,2006,28(10):969 [12]Sun J C,Zhang T Y,Liu F.Modeling of chaotie systems based (刘华强,唐获,杨荃,等.模糊小脑模型神经网络在多辊冷 on modified weighted recurrent least squares support vector ma- 连轧机轧制力预报模型中的应用.北京科技大学学报,2006, chines.Chin Phys,2004,13(12):2045何 飞等: 基于多模态和加权支持向量机的热轧轧制力智能预报 表 3 轧制力计算结果 Table 3 Computed result of rolling force 数据 序列 接触面 积/m2 变形抗力/ ( N·mm - 2 ) 应力 系数 张力 系数 轧制力实 际值/MN 机理模型 聚类前加权支持向量机 聚类后加权支持向量机 轧制力计算 值/MN 相对误 差/% 修正轧制 力/MN 修正相对 误差/% 修正轧制 力/MN 修正相对 误差/% 1 0. 15 125. 44 1. 098 0. 986 24. 70 20. 37 17. 5 26. 78 8. 4 25. 28 2. 3 2 0. 17 134. 86 1. 074 0. 979 27. 36 24. 11 11. 8 25. 45 7. 0 28. 85 5. 5 3 0. 22 141. 25 1. 069 0. 981 28. 43 32. 59 14. 6 30. 63 7. 7 29. 1 2. 4 4 0. 21 129. 38 1. 078 0. 963 25. 85 28. 21 9. 1 27. 69 7. 1 26. 71 3. 3 5 0. 19 132. 42 1. 067 0. 976 28. 91 26. 2 9. 4 31. 17 7. 8 30. 47 5. 4 6 0. 15 127. 61 1. 072 0. 98 23. 64 20. 11 14. 9 26. 08 10. 3 24. 41 3. 3 7 0. 13 138. 59 1. 085 0. 991 23. 19 19. 37 16. 5 24. 97 7. 7 23. 55 1. 6 8 0. 17 135. 47 1. 087 0. 985 27. 27 24. 66 9. 6 29. 24 7. 2 28. 31 3. 8 9 0. 18 128. 36 1. 077 0. 978 26. 74 24. 34 9. 0 25. 13 6. 0 25. 89 3. 2 10 0. 2 131. 93 1. 091 0. 986 25. 58 28. 38 11. 0 27. 81 8. 7 25. 97 1. 5 平均值 ― ― ― ― ― ― 12. 3 ― 7. 8 ― 3. 2 差降至 5% 以下,多组数据的平均值为 3. 2% ,满足现 场生产要求. 参 考 文 献 [1] Sun Y K. Model and Control of Hot Strip Rolling. Beijing: Metal￾lurgical Industry Press,2007 ( 孙一康. 带钢热连轧的模型与控制. 北京: 冶金工业出版 社,2007) [2] Hong Y,Tang L X,Zhang Y Y. Optimization of rolling force of hot rolling by using data subspace PLS modeling technique. Con￾trol Decis,2014,29( 7) : 1199 ( 洪悦,唐立新,张颜颜. 基于数据子空间 PLS 建模技术的热 轧轧制力优化设定. 控制与决策,2014,29( 7) : 1199) [3] Zhou F Q,Cao J G,Zhang J,et al. Online calculation model of rolling force for a tandem cold rolling mill. J Univ Sci Technol Bei￾jing,2006,28( 9) : 859 ( 周富强,曹建国,张杰,等. 冷连轧机轧制力在线计算模型. 北京科技大学学报,2006,28( 9) : 859) [4] Liu H Q,Tang D,Yang Q,et al. Rolling force prediction model of a multi roll cold tandem mill by fuzzy cerebellum model articula￾tion controller. J Univ Sci Technol Beijing,2006,28( 10) : 969 ( 刘华强,唐荻,杨荃,等. 模糊小脑模型神经网络在多辊冷 连轧机轧制力预报模型中的应用. 北京科技大学学报,2006, 28( 10) : 969) [5] Chen Z M,Luo F,Huang X H,et al. The prediction of rolling force based on chaotic optimized support vector machine. Control Decis,2009,24( 6) : 808 ( 陈治明,罗飞,黄晓红,等. 基于混沌优化支持向量机的轧 制力预测. 控制与决策,2009,24( 6) : 808) [6] Cortes C,Vapnik V. Support vector networks. Mach Learn, 1995,20( 3) : 273 [7] Smola A J,Scholkopf B. A tutorial on support vector regression. Stat Comput,2004,14( 3) : 119 [8] Kaufman L,Rousseeuw P J. Finding Groups in Data: an intro￾duction to Cluster Analysis. New York: John Wiley & Sons Press, 1990 [9] John C P. Using Analytic QP and Sparseness to Speed Training of Support Vector Machines. Cambridge: MIT Press,1999 [10] Suykens J A K. Least squares support vector machines for classi￾fication and nonlinear modeling. Neural Network World,2000, 10( 1) : 29 [11] Suykens J A K,De Brabanter J,Lukas L,et al. Weighted least squares support vector machine: robustness and sparse approxi￾mation. Neurocomputing,2002,48( 1) : 85 [12] Sun J C,Zhang T Y,Liu F. Modeling of chaotic systems based on modified weighted recurrent least squares support vector ma￾chines. Chin Phys,2004,13( 12) : 2045 · 125 ·
<<向上翻页
©2008-现在 cucdc.com 高等教育资讯网 版权所有