人工神经网络在优化BaTiO_3陶瓷配方研究中的应用  

Application of Artificial Neural Network (ANN) Technique to the Formulation Design of BaTiO_3 Dielectric Ceramics

作　　者：郭栋[1] 王永力[1] 李龙土[1] 桂治轮[1] 夏军涛[2] 

机构地区：[1]清华大学材料系新型陶瓷与精细工艺国家重点实验室,北京100084 [2]北京理工大学化工与材料学院,北京100081

出　　处：《无机材料学报》2002年第4期845-851,共7页Journal of Inorganic Materials

摘　　要：首次将人工神经网络技术用于介电陶瓷的配方性能分析.以BaTiO3为研究对象选取了几种掺杂剂,在均匀实验设计的基础上,用BP人工神经网络对所得实验结果进行了分析,建立了相应配方的数学模型并将其与多重非线形回归模型的结果进行了比较.通过对人工神经网络配方数学模型的二次分析,得到了比多重非线形回归模型更加丰富的配方信息和内在规律,并且用图形化方式直观地表达了出来.在进一步对配方结果的优化和验证的基础上发现实验结果能够较好地符合理论预测,说明人工神经网络对于获得多性能指标要求介电陶瓷的最优化配方具有较好的指导作用.Application of the artificial neural network (ANN) to the formulation design of BaTiO3 based dielectrics was carried through for the first time. Based on the homogenous experimental design, the experimental results of 21 samples were analyzed by a three-layered BP network model. The results were also expressed by intuitionistic graphics. In addition, optimized formulations were calculated and the optimized epsilon(25) output values were in accordance with experiments. The three-layer BP network proved to be a very useful tool in dealing with problems with serious non-linearity encountered in the formulation design of dielectric ceramics.

关 键 词：人工神经网络 优化 BATIO3 陶瓷 配方 钛酸钡 介电性能 人工神经网络 BP算法 

分 类 号：TQ174.758[化学工程—陶瓷工业] TP183[化学工程—硅酸盐工业]