摘要
运用BP神经网络方法建立了铅黄铜超塑性状态下伸长率与变形参数之间的预测模型,采用标准前馈式神经网络原理建立了铅黄铜超塑性拉伸试验参数与其伸长率之间的神经网络模型,以试验数据为样本,对所建模型进行训练,较好的预测了铅黄铜超塑拉伸的伸长率,最大误差只有4·95%。实现了不同变形工艺参数与伸长率之间的非线性映射,也为优化铅黄铜轴承保持架的超塑性成形参数提供理论和试验依据。伸长率预测值与试验结果吻合良好。
The predicting model for relationship between elongation and deformation parameters of superplastic lead brass has been established based on standard feed-forward BP neural network.The elongation of the superplastic lead brass can be effectively predicted,such as maximum error lower than 4.81%,by training the established model based on experimental data,realizing the non-linear mapping between different deformation parameters and elongation.The predicted values of the elongation is well in agreement with experimental results,which provides a reference to theory and experiment of superplastic forming parameters of lead brass bearing cage.
出处
《特种铸造及有色合金》
CAS
CSCD
北大核心
2007年第6期430-432,共3页
Special Casting & Nonferrous Alloys
基金
河南省科技攻关资助项目(0624260035)
河南省高校杰出科研人才创新工程资助项目(2004KYCX020)
关键词
铅黄铜
超塑性
伸长率
BP神经网络
预测模型
Lead Brass,Superplastic,Elongation,BP Neural Network,Prediction Model
