Transactions of Materials and Heat Treatment

Title：Research on improving prediction accuracy of maximum reduction by open-die cold extrusion

Authors： ZHANG Shuang-jie GAO Ying WANG Li-juan LIU Yu-zhong LI Jun (College of Material Science and Engineering Hebei University of Science and Technology Shijiazhuang Hebei 050054 China)

Unit：

KeyWords： genetic algorithm neural network BP algorithm open-die cold extrusion reduction

ClassificationCode：TG376.3

year,vol(issue):pagenumber：2005,30(3):46-49

Abstract：

Combination of genetic algorithm and neural network to optimize the network weight value can make full use of the advantage of the ability of BP algorithm's function approximation, GA algorithm's high search efficiency and its global search ability, and the new prediction model about maximum reduction is published in view of above.The new prediction model has the advantages of least iterative times, shorter training time and better performance of the prediction compared with prediction model based on BP algorithm.The calculated and analyzed results can be applied to the practical production.

Funds：

河北省自然科学基金项目(E2004000296)资助

Reference：

[1]李军.开式冷挤压成形的理论建模及工艺参数的应用研究[D].天津大学博士学位论文,1997.
[2]高颖.基于BP网络的开式冷挤压关键参数预测研究[D].燕山大学硕士论文,2001.
[3]高颖,张双杰,韩鹏彪,等.用神经网络对开式冷挤压极限变形程度的预测[J].河北工业科技,2002,19(2):1720.
[4]张立明.人工神经网络的模型及其应用[M].上海:复旦大学出版社,1993.
[5]徐丽娜.神经网络控制[M].哈尔滨:哈尔滨工业大学出版社,1999.
[6](日)玄光男,程润伟著,汪定伟,等译.遗传算法与工程设计[M].北京:科学出版社,2000.
[7]袁慧梅.具有自适应交换率和变异率的遗传算法[J].首都师范大学学报(自然科学版),2000,21(3):1420.
[8]张大志,李谋渭,孙一康,等.用改进的遗传神经网络预报冷连轧轧机的轧制压力[J].钢铁研究,2000,114(3):2731.
[9]梁化楼,戴贵亮.人工神经网络与遗传算法的结合进展与展望[J].电子学报,1995,(10):194200.
[10]SchafferJD,MorishmaA.Anadaptivecrossovermechanismfor geneticalgorithms[C].InProcessdingsofthesecondInternation alConferenceonGeneticAlgorithms,1987:3640.
[11]Michalecz.ZA.ModifiedGeneticAlgorithmforOptimalControl problems[J].ComputersMathApplic,1992,23(12):8394.
[12]SchafferJD,MorisbmaA.Anadaptivecrossovermechamism forgeneticalgorithms[C].InProceedingofthesecondInter nationalconferenceonGeneticAlgorithms,1987:3640.

Service：

【This site has not yet opened Download Service】【Add Favorite】