论文部分内容阅读
基于BP神经网络的Fe81Ga19合金磁致伸缩性能,在不同的晶体偏离角度、预压力和磁场强度下,得到了Fe81Ga19合金的磁致伸缩应变的实验值。以试凑法来确定BP网络的中间隐层,确定网络结构为3-9-1。预测结果表明,采用BP神经网络法,可预测Fe81Ga19合金的磁致伸缩应变,预测误差均低于6%。
The magnetostrictive properties of Fe81Ga19 alloy based on BP neural network are obtained under different angles of crystal deviation, pre-pressure and magnetic field, and the experimental results of magnetostrictive strain of Fe81Ga19 alloy are obtained. To try to determine the middle of the BP network hidden layer, to determine the network structure is 3-9-1. The prediction results show that using BP neural network method can predict the magnetostrictive strain of Fe81Ga19 alloy, the prediction errors are less than 6%.