锻压技术

基于改进PSO-BP的拉延筋参数反求优化

英文标题：Parameter inverse optimization for drawbeads based on improved PSO-BP model

作者：王新宝谢延敏王杰乔良

单位：西南交通大学

分类号：

出版年,卷(期):页码：2014,39(4):10-15

摘要：

采用模拟退火算法优化基于加权平均法的拉丁超立方设计，获得了拉延筋阻力样本。采用Dynaform对翼子板成形进行仿真，以最大增厚和最大减薄作为输出目标，采用改进的粒子群优化BP算法，建立拉延筋映射模型，与单纯采用PSO-BP建立的映射模型进行对比，预测精度大幅提高。采用粒子群算法对映射模型进行优化，得到最优拉延筋阻力，采用非线性函数优化方法求得最优的拉延筋几何参数。采用等效拉延筋阻力模型，避免了有限元网格的重划分和采用真实拉延筋带来的计算效率低的问题，成形效果图说明了采用该方法可以获得较好的最优拉延筋几何参数。

LHS was optimized by simulated annealing algorithm based on weighted average method，and drawbead force samples were got. Dynaform was used to simulate the wing. Improved PSO-BP was applied to build drawbead force mapping model by taking the maximum thickening and the maximum thinning as output goal. Compared with the unimproved PSO-BP mapping model，the accuracy of improved model was significantly raised.The optimal drawbead force was obtained by simulating the mapping model using PSO，and the optimal drawbead geometrical parameters were obtained by nonlinear equations. The problem of poor computational efficiency brought by remeshing and adopting real drawbead was avoided by employing equivalent drawbead model. Forming diagram proves that the optimal drawbead geometrical parameters can be obtained by using the method.

基金项目：

国家自然科学基金资助项目(51005193,51275431)

王新宝（1989-），男，硕士研究生

参考文献：

[1]韩利芬，高晖，李光耀，等.神经网络与遗传算法在拉延筋参数反求中的应用[J].机械工程学报，2005,41(5)：171-176.

Han Lifen, Gao Hui, Li Guangyao, et al. Application of neural network and genetic algorithm to inverse solution of parameters of drawbead [J]. Chinese Journal of Mechanical Engineering, 2005, 41(5)：171-176.

[2]郑刚，李光耀，孙光永，等.基于近似模型的拉延筋几何参数反求[J].中国机械工程，2006,17(19)：1988-1992.

Zheng Gang, Li Guangyao, Sun Guangyong, et al. Geometrical parameter inverse problem for drawbeads based on the approximate model [J]. China Mechanical Engineering, 2006, 17(19)：1988-1992.

[3]谢延敏，王新宝，王智，等. 基于灰色理论和GA-BP的拉延筋参数反求[J]. 机械工程学报，2013，49(4):44-50.

Xie Yanmin. Wang Xinbao. Wang Zhi,et al. Parameter inverse problem for drawbeads based on the gray theory and GA-BP[J]. Chinese Journal of Mechanical Engineering, 2013, 49(4)：44-50.

[4]陈涛，高晖，李光耀，等. 真实拉延筋参数化建模及其在薄板冲压仿真中的应用[J].中国机械工程，2006,17(S1)：23-26.

Chen Tao, Gao Hui, Li Guangyao, et al. New method of modeling drawbead geometry and its application in sheet metal forming[J]. China Mechanical Engineering, 2006,17(S1)：23-26.

[5]刘艳芳，施法中，冯云飞.等效拉延筋模型在板料成形数值模拟中的具体实现[J]. 机械工程学报, 2005, 41(1):115-117.

Liu Yanfang, Shi Fazhong, Feng Yunfei. Elaborate implementation of an equivalent drawbead model in the numerical simulation of sheet metal forming[J]. Chinese Journal of Mechanical Engineering, 2005, 41(1):115-117.

[6]Rumelhart D E, Hinton G E, Williams R J. Learning representations by back-propagation error[J]. Nature,1986, 323(9):533-536.

[7] Belytschkot Ted,Jerry I Lin. Explicit algorithms for the nonlinear dynamics of shells [J].Compute Methods in Applied Mechanics and Engineering, 1984, 20: 225-251.

[8]李群，李娜，郭宝峰，等.半圆筋循环载荷作用下的板材变形特征研究[J],锻压技术，2013,38（2）：31-33.

Li Qun, Li Na, Guo Baofeng, et al. Study of sheet deformation characteristics under the cyclic loading applied by semicircle drawbead[J]. Forging & Stamping Technology,2013,38(2):31-33.

[9]Hickernell F J. A generalized discrepancy and quadrature error bound[J]. Mathematics of Computation, 1998, 67: 299-322.

[10]谢延敏，于沪平，陈军，等.基于灰色系统理论的方盒件拉深稳健设计[J].机械工程学报，2007,43(3)：55-58.

Xie Yanmin, Yu Huping, Chen Jun, et al. Application of grey theory in deep drawing robust design [J]. Chinese Journal of Mechanical Engineering, 2007, 43(3)：55-58.

服务与反馈：

【文章下载】【加入收藏】