摘要
采用人工神经网络的径向基函数网络对压铸AZ91镁合金的力学性能进行预测,结合GAP方法和梯度下降法,提出优化算法。结果表明,该优化网络算法具有较快的收敛速度和良好的预测性能,能够成为合金设计有力的辅助手段。
The mechanical properties of AZ91 Magnesium alloys were predicted with the artificial neural network and radial basis function network. An optimal algorithm was made by applying the growing and pruning method （GAP） and gradient descent method. The results show that the network with the optimal algorithm has a high speed of convergence and good quality of prediction ,which will be an effective assistant method for allov design.
出处
《金属热处理》
EI
CAS
CSCD
北大核心
2007年第4期19-22,共4页
Heat Treatment of Metals
基金
江苏省重点基金项目(P0251-061)
关键词
力学性能预测
神经网络
镁合金
径向基函数(RBF)网络
prediction of mechanical properties
artificial neural network
magnesium alloy
radial basis function （RBF） network
