铜合金时效性能的神经网络预测模型

Neural Network Predictive Model for Ageing Properties of Copper Alloys

运用BP神经网络算法 ,对时效试验数据进行训练 ,建立了Cu 0 .30Cr 0 .15Zr合金时效后硬度和导电性与时效时间和时效温度的映射模型 ,从而可预测铜合金在一定时效条件下的硬度和导电性 ,预测结果与实测值吻合较好 ,表明神经网络用于铜合金的时效性能预测是可行的 。

A predictive model for ageing properties of copper alloys(Cu-0.30Cr-0.15Zr) by BP artificial neural net was developed.The non-linear relationship between hardness,conductivity and ageing time, ageing temperature were established . Hardness and conductivity performances of copper alloys can be predicted by means of the trained neural net from the ageing data. It shows that the errors between the predictive value and the measured value are very small, which proves the predictive model for hardness and conductivity performances of copper alloys is feasible and effective.The prediction accuracy depends on the quality and quantity of allay ageing data for training neural network.