BP网络优化纳米Fe_3O_4的制备工艺  

Optimization with BP network for the preparation condition of Fe_3O_4 nanoparticle

作　　者：崔升[1] 沈晓冬[1] 林本兰[1] 

机构地区：[1]南京工业大学材料科学与工程学院学院,江苏南京210009

出　　处：《云南大学学报（自然科学版）》2005年第S3期404-408,共5页Journal of Yunnan University(Natural Sciences Edition)

基　　金：江苏省高等学校研究生创新计划项目资助;江苏省人民医院合作项目资助;中国人民解放军总装备部预研管理中心项目资助(41312030312)

摘　　要：采用液相共沉淀法制备出纳米级Fe3O4颗粒,并根据BP神经网络理论,建立了相应模型,采用正交设计实验的结果进行模型的训练,用训练后的模型对预测样本进行了预测与优选.模型优选结果表明,以胶溶化时加入HCl为2.3 mL,Fe3+/Fe2+的比例0.77、共沉淀时的pH值10.8和制备Fe3O4的晶化温度65℃为最佳实验条件,预测纳米粉体平均粒径为10.3 nm,而实际实验结果为10.5 nm,相对误差为1.94%.It is introduced that the preparation of Fe_3O_4 nanoparticle with liquid-phase coprecipitation.An optitimizing selection model for preparation was proposed based on the BP net theory.The model had been trained using the data from the experiment design of processing conditions.Then it was used for optimizing test samples.The optimized value was 10.3nm,with a relative error of only 1.94% compared with the actual value,i.e.,10.5?nm when proportion of HCl was 2.3?mL,proportion of Fe^(3+)/Fe^(2+) was 0.77,pH was 10.8 and crystal temperature was 65?℃.

关 键 词：BP神经 Fe3O4纳米粉 液相共沉淀法 制备 

分 类 号：TB383.1[一般工业技术—材料科学与工程]