摘要
为提高复合磁体材料热处理传统工艺效率、降低研发成本,利用基于MATLAB软件的反向传播神经网络BPNN平台,将材料制备工艺与数学预测方法相结合。研究结果表明,BPNN能够较好地预测出,当热处理条件改变时材料剩磁性能的变化规律。通过实验数据,可以预测出复合磁体材料最佳剩磁性能所需的热处理温度和保温时间,这可以进一步地表明,BPNN具有训练方法简单、训练速度快、预测结果精确稳定等优点。
In order to improve the traditional process efficiency of composite magnetmaterial heat treatment and reduce the research and development cost,the material preparation process and mathematical prediction method are combined by BPNN platform of back propagation neural network based on MATLAB software.The results show that BPNN can better predict the change of remanence properties of the material when the heat treatment conditions change.Through the experimental data,the heat treatment temperature and holding time required to obtain the optimum remanence performance of the composite magnet material can be predicted,and it is further shown that BPNN has the advantages of simple training method,fast training speed and accurate and stable prediction results.
作者
张椿英
赵浩峰
周健
王冬玲
黄俊杰
ZHANG Chun-ying;ZHAO Hao-feng;ZHOU Jian;WANG Dong-ling;HUANG Jun-jie(Anhui Institute of Information Engineering,Wuhu Anhui 241000,China;Nanjing University of Information Scinence&Tecnology,Najing Jiangsu 210044,China)
出处
《铜陵学院学报》
2021年第4期105-107,共3页
Journal of Tongling University
基金
安徽省信息工程学院校级一流课程建设“金属材料及热处理”(AXG2019-jwc-3126)。
关键词
BPNN
剩磁
复合磁体材料
热处理工艺
BPNN
remanence
composite magnet material
heat treatment process