人工神经网络对NaNbO_(3)基陶瓷介电性能的预测研究

Prediction of Dielectric Properties of NaNbO_(3) Based Ceramics via Artificial Neural

NaNbO_(3)基陶瓷在电介质储能领域具有极大的应用潜力。研究在对NaNbO_(3)基复合陶瓷材料开展实验研究的基础上,基于人工神经网络方法构建BP神经网络与优化的GA-BP神经网络模型,以磷酸盐玻璃相的添加量、烧结温度、烧结时间作为输入,介电性能(介电常数与介电损耗)作为输出,对NaNbO3基复合陶瓷材料的介电性能开展预测研究。结果表明,通过GA-BP网络预测的介电常数相对误差最大仅为1.03%,介电损耗预测结果最大值仅为-3.18%,完全符合应用需求。

NaNbO_(3)-based ceramics have great potential in dielectric energy storage applications.On the basis of experimental study of NaNbO_(3)-based ceramic composites,this study has predicted their dielectric properties by artificial neural network method.The BP neural network and its optimized GA-BP neural network model were constructed based on the artificial neural network method.The addition amount of phosphate glass phase,sintering temperature and dwelling time of the NaNbO_(3)-based ceramic composites were taken as inputs,and the dielectric properties including dielectric constant and dielectric loss were taken as outputs.The results show that the maximum relative error of the dielectric constant predicted by the GA-BP neural network is only 1.03%,and the maximum value of the dielectric loss predicted by the GA-BP neural network is only-3.18%,which fully meet the requirement of applications.