集成机理与数据的复杂模锻过程在线建模方法

中国机械工程 ›› 2015, Vol. 26 ›› Issue (9): 1227-1232.

集成机理与数据的复杂模锻过程在线建模方法

吕文兵;陆新江;黄明辉;雷杰;邹玮

中南大学高性能复杂制造国家重点实验室，长沙，410083

出版日期:2015-05-10 发布日期:2015-05-08
基金资助:
国家重点基础研究发展计划(973计划)资助项目(2011CB706802)；国家自然科学基金资助项目(51205420)；教育部新世纪优秀人才支持计划资助项目(NCET-13-0593)；湖南省自然科学基金资助项目(14JJ3011); 中南大学机电工程学院研究生创新项目(2014scxjj02)

Online Modeling Method with Integrated Mechanism and Data for Whole Complex Forging Process

Lü Wenbing;Lu Xinjiang;Huang Minghui;Lei Jie;Zou Wei

Central South University,Changsha,410083

Online:2015-05-10 Published:2015-05-08
Supported by:
National Program on Key Basic Research Project (973 Program)（No. 2011CB706802）;National Natural Science Foundation of China（No. 51205420）;Program for New Century Excellent Talents in University of Ministry of Education of China（No. NCET-13-0593）;Hunan Provincial Natural Science Foundation of China（No. 14JJ3011）

摘要/Abstract

摘要：

大型航空锻件高精度成形成性依赖于精确的锻造过程模型，然而不规则的锻件形状、复杂的微观流变过程、强非线性与时变的负载力使得高精度的锻造过程模型难以获得。为此，在结合解析建模和数据建模优点的基础上，提出了集成机理与数据的复杂模锻过程在线建模方法。应用物理与过程知识推导了锻造过程的解析模型，在此基础上提出使用在线极限学习机方法构建由于泄漏、不确定性、干扰等引起的偏差模型，实现了锻造过程模型的实时进化，从而满足强非线性与时变性的锻造过程要求。实验结果表明，新方法能有效地建立复杂锻造过程模型，且比现有的方法有更好的建模精度。

关键词: 大型锻件, 锻造过程, 解析模型, 在线极限学习机

Abstract:

Forming of large aviation forging with high dimensional accuracy and high quality was dependent on highly precise model. However, it was hard to obtain precise model with complex forging shape, complex microstructure of forging material, strongly nonlinear and time-varying forging force, and so on. The physical and analytical model was combined with online compensating data model to build a new model. This data model was based on online extreme learning machine (OSELM), with the purpose of reducing modeling errors of analytical model caused by leakage, uncertainty, disturbance, and so on. And new model became less dependent on full data as well. This new model firstly took advantages of the analytical model built with kinetic equation and physical modeling method, and then used the OSELM model to build the deviation model. The forging process model could evolve over time, to satisfy the demands of highly nonlinear and time varying forging process. The simulation results show that, new model can predict the dynamic behavior of forging process well, and has a better prediction precision compared to the solo model which exists as part of the new model.

Key words: large forgings, forging process, analytical model, online sequential extreme learning machine (OSELM)

中图分类号:

TG315.9

吕文兵, 陆新江, 黄明辉, 雷杰, 邹玮. 集成机理与数据的复杂模锻过程在线建模方法[J]. 中国机械工程, 2015, 26(9): 1227-1232.

Lü Wenbing, Lu Xinjiang, Huang Minghui, Lei Jie, Zou Wei. Online Modeling Method with Integrated Mechanism and Data for Whole Complex Forging Process[J]. China Mechanical Engineering, 2015, 26(9): 1227-1232.

参考文献

[1]王淑云,李惠曲,东赟鹏,等. 大型模锻件和模锻液压机与航空锻压技术[J]. 锻压装备与制造技术, 2009, 44(5):31-34.
Wang Shuyun, Li Huiqu, Dong Yunpeng, et al. Development of Large Single-piece Forgings and Heavy Forging Press in Aerospace Forging Industry[J]. China Metalforming Equipment and Manufacturing Technology, 2009,44(5):31-34.
[2]李蓬川. 大型航空模锻件的生产现状及发展趋势[J]. 大型铸锻件, 2011(2):39-34.
Li Pengchuan. Production Status and Development Trend of Heavy Aviation Die Forging[J]. Heavy Casting and Forging, 2011(2): 39-34.
[3]杨家旺. 大型模锻压机运行在线建模及动态特性研究[D]. 长沙：中南大学,2013.
[4]王军生,赵启林,矫志杰,等. 冷连轧过程控制变形抗力模型的自适应学习[J]. 东北大学学报（自然科学版）, 2004, 25(10): 973-976.
Wang Junsheng, Zhao Qilin, Jiao Zhijie, et al. Adaptive Learning of the Model of Deformation Resistance Model for Tandem Cold Rolling Process Control[J]. Journal of Northeastern University (Natural Science), 2004,25(10):973-976.
[5]张猛. 极低速下大型模锻压机系统建模与动态特性分析[D]. 长沙：中南大学,2012.
[6]马璟,陈拂晓. 齿轮钢20CrMnTi在1000℃时的变形抗力数学模型[J]. 材料热处理技术,2012,41(4):49-50.
Ma Jing, Chen Fuxiao. Deformation Resistance Model for Gear Steel 20CrMnTi at 1000℃[J]. Material and Heat Treatment, 2012, 41(4) : 49-50.
[7]刘战英,冯运莉,田薇,等. 45钢低温轧制的变形抗力模型[J]. 轧钢, 2004, 21(1): 12-14.
Liu Zhanying, Feng Yunli, Tian Wei, et al. Model of Deformation Resistance of 45 Steel in Low Temperature Rolling[J]. Steel Rolling, 2004,21(1):12-14.
[8]黄克坚，包忠诩，陈泽中，等. 关于挤压变形规律理论研究方法的一些探讨[J].塑性工程学报, 2003, 10(4): 45-51.
Huang Kejian, Bao Zhongxu, Chen Zezhong, et al. The Discussion of the Theoretical Research Methods Used in the Analysis of Extrusion Forming Laws[J]. Journal of Plasticity Engineering, 2003, 10(4): 45-51.
[9]Lu X J, Li Y B, Huang M H. Operation-region-decomposition-based SVD/NN Modeling Method for Complex Hydraulic Press Machines[J]. Industrial and Engineering Chemistry Research, 2013,52:17221-17228.
[10]林治平. 锻压变形力的工程计算[M]. 北京：机械工业出版社，1986.
[11]Huang G B, Zhu Q Y, Siew C K. Extreme Learning Machine: Theory and Applications[J]. Neurocomputing, 2006, 70(3):489-501.
[12]Huang G B, Chen L, Siew C K. Universal Approximation Using Incremental Constructive Feedforward Networks with Random Hidden Nodes[J]. IEEE Transaction on Neural Networks, 2006, 17(4): 879-892.
[13]Huang G B, Zhou H,Ding X, et al. Extreme Learning Machine for Regression and Multi-class Classification[J]. IEEE Transaction on Systems Man and Cybernetics-Part B：Cybernetics, 2012, 42(2): 51-529.
[14]Liang N Y,Huang G B, Saratchandran P, et al. A Fast and Accurate Online Sequential Learning Algorithm for Feedforward Networks[J]. IEEE Trans. Neural Networks, 2006, 17(6):1411-1423.
[15]Kasun L L C, Zhou H, Huang G B, et al. Representational Learning with Extreme Learning Machine for Big Data[J]. IEEE Intelligent System, 2013,28(6):1-4.