基于BP神经网络的Fe81Ga19合金磁致伸缩性能研究

Magnetostriction Properties of Fe81Ga19 Alloy Based on BP Neural Network

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 magnetostriction properties of Fe81Ga19 alloy was investigated by using BP neural network. Under different crystal deviation angles, pre-pressures and magnetic field intensities, the experimental magnetostriction strain values of Fe81Ga19 alloy were measured. Using a trial and error method. The intermediate hidden layers were determined and the network structure on 3-9-1 was obtained. The predicted results show that BP neural network is a good method for predicting the magnetostriction strain of Fe81Ga19 alloy and the prediction errors is less than 6%.