机构地区: 西北工业大学
出 处: 《中国机械工程》 2003年第23期2056-2059,共4页
摘 要: 快速凝固是一种获得高强度和高导电铜合金的有效途径。通过对时效温度和时间与硬度和导电率样本集的学习,采用Levenberg—Mar-quardt算法和四层BP网络建立了通用的快速凝固Cu—Cr—Zr合金时效工艺神经网络模型。实验结果与预测值吻合良好,从而为预测和控制该工艺性能开辟了新的途径。 Rapid solidifiation makes it possible for ag-ing Cu--Cr--Zr alloy to enhance the hardness and electri- cal conductivity. This paper presents a new method forpredicting the properties of rapidly solidified aging, whichwas developed on the basis of the four layers of BackPropagation neural network and the Levenberg -- Mar-quardt training algorithm. The non -- linear relationshipbetween properties of Cu--Cr--Zr alloy and aging param-eters is available. The results show that the predictedvalues are well identical with the experimental ones.