网站首页期刊简介编委会过刊目录投稿指南广告合作征订与发行联系我们English
基于神经网络的机械油泵轴锻造工艺优化
英文标题:Forging process optimization on mechanical oil pump shafts
作者:莫洪武 塔金星 
单位:广西农业职业技术学院 东北林业大学 
关键词:神经网络 锻造工艺优化 机械油泵轴 训练误差 预测误差 
分类号:TU599
出版年,卷(期):页码:2018,43(9):21-24
摘要:
为了优化机械油泵轴锻造工艺,提升机械油泵轴的综合性能,基于5×25×1三层拓扑结构,以坯料加热温度、始锻温度、终锻温度、模具预热温度和锻造变形量为5个输入参数,以磨损性能为输出参数,以tansig函数为传递函数,构建了机械油泵轴锻造工艺神经网络优化模型,并进行了神经网络模型的训练、预测与验证。结果表明该模型平均相对训练误差为3.2%,相对预测误差低于5%,具有较高预测精度和较强预测能力。与生产线现用工艺相比,采用模型优化工艺锻造的SKH-51高速钢机械油泵轴的磨损体积减小了51.9%,磨损性能得到明显提高。
In order to optimize the forging process and improve the comprehensive performance of mechanical oil pump shaft, based on the 5×25×1 three-level topological structure, the neural network optimization model for forging process of mechanical oil pump shaft was designed and constructed by taking the heating temperature of billet, initial forging temperature, final forging temperature, mold preheating temperature and forging deformation amount as five input parameters, wear performance as output parameter, and tansig function as transfer function, and the training, prediction and validation of the neural network model were carried out. The results show that the average relative training error of the model is 3.2%, the relative prediction error is lower than 5%, and the model has high prediction accuracy and strong prediction ability. Compared with the present production line, the wear volume of the high speed steel SKH-51 mechanical oil pump shaft forged by model optimization is reduced by 51.9%, and the wear performance is obviously improved.
基金项目:
广西高校科研项目(KY2016YB686)
作者简介:
莫洪武(1980-),男,硕士,副教授 E-mail:nzymhw@163.com
参考文献:

[1]魏伊伦,束学道,王雨,等.模具参数对闭开联合轧制汽车油泵轴力能参数的影响
[J].中国机械工程,2017,28(10):1238-1244.

Wei Y L,Shu X D,Wang Y,et al. Influences of mold parameters on mechanics parameters in closed-open type cross wedge rolling of automobile oil pump shaft part
[J]. China Mechanical Engineering,2017,28(10):1238-1244.


[2]石岩,李云峰,刘佳,等.高压油泵凸轮轴激光增材制造梯度耐磨层研究
[J].机械工程学报,2017,53(6):80-87.

Shi Y,Li Y F,Liu J,et al. Research of gradient wear-resisting coating produced by laser additive manufacturing on high-pressure pump camshaft
[J]. Journal of Mechanical Engineering,2017,53(6):80-87.


[3]苏永利,刘进军,刘永伟,等.乙烯装置急冷油泵泵轴失效断裂原因分析
[J].机械工程师,2016,24(8):78-79.

Su Y L,Liu J J,Liu Y W,et al. Reason analysis on shaft fracture of ethylene quench oil pump pump
[J]. Mechanical Engineer,2016,24(8):78-79.


[4]李卫战,陈莉,牛星军,等.一种间接驱动润滑油泵轴断裂失效的分析
[J].移动电源与车辆,2016,(2):26-31.

Li W Z,Chen L,Niu X J,et al. An analysis of fracture failure of oil pump shaft in indirect drive
[J]. Movable Power Station & Vehicle,2016,(2):26-31.


[5]徐仕海,邓晓晖,李诚,等.百万核电汽轮发电机组顶轴油泵振动问题处理
[J].东方汽轮机,2015,39(3):56-58,74.

Xu S H,Deng X H,Li C,et al. Analysis and disposal of vibration problem for jacking-up oil pump in 1000 MW nuclear turbine
[J]. Dangfang Turbine,2015,39(3):56-58,74.


[6]王宝雨,左斌,杨乐毅,等.油泵齿轮轴精锻成形工艺研究及缺陷分析
[J].北京科技大学学报,2014,36(8):1087-1093.

Wang B Y,Zuo B,Yang L Y,et al. Forming process study and defect analysis of precision forging for the gear shafts of oil pumps
[J]. Journal of University of Science and Technology Beijing,2014,36(8):1087-1093.


[7]史小云,张晓园,毛友川,等.锻造温度对TC21钛合金锻板组织和力学性能的影响
[J].锻压技术,2015,40(1):14-16,42.

Shi X Y,Zhang X Y,Mao Y C,et al. Influence of forging temperature on structure and mechanical property for TC21 titanium plate
[J]. Forging & Stamping Technology,2015,40(1):14-16,42.


[8]莫玉梅.基于神经网络的镁合金汽车车轮锻压工艺优化
[J].轻合金加工技术,2016,44(5):28-32,62.

Mo Y M. Forging process optimization of magnesium alloy automobile wheels based on neural network
[J]. Light Alloy Fabrication Technology,2016,44(5):28-32,62.


[9]舒阳,顾炳伟,张义桥,等.BP神经网络算法对建筑保温材料性能的预测
[J].建筑节能,2017,45(4):52-55.

Shu Y,Gu B W,Zhang Y Q,et al. Performance prediction of building thermal insulation materials based on BP neural network algorithm
[J]. Building Energy Efficiency,2017,45(4):52-55.


[10]佘勇,樊文欣,陈东宝,等.基于RBF神经网络的强力旋压连杆衬套力学性能预测研究
[J].锻压技术,2016,41(6):128-132,145.

She Y,Fan W X,Chen D B,et al. Study on mechanical property prediction of power spinning connecting rod bushing based on RBF neural network
[J]. Forging & Stamping Technology,2016,41(6):128-132,145.
服务与反馈:
本网站尚未开通全文下载服务】【加入收藏
《锻压技术》编辑部版权所有

中国机械工业联合会主管 北京机电研究所有限公司 中国机械工程学会塑性工程分会主办
联系地址:北京市海淀区学清路18号 邮编:100083
电话:+86-010-62920652 +86-010-82415085 传真:+86-010-62920652
E-mail: fst@263.net(稿件) dyjsjournal@163.com(广告)
京ICP备09032115号-5