开放骨架磷酸铝定向合成反应预测研究

Prediction research on open-framework aluminophosphate syntheses

针对国内外材料领域和分子工程学领域中如何定向合成无机微孔材料的热点问题.基于开放骨架磷酸铝具有独特结构特征的特点,于吉林大学“无机合成与制备国家重点实验室”构建的磷酸铝合成反应数据库的研究基础上,采用基于机器学习的算法对磷酸铝合成反应进行预测.首先提取富信息样本建立预测模型和补值模型.然后利用BP补值模型对数据库中的缺失数据进行校正.最后根据已建模型建立了一个简单、易用的定向合成(12,6)元环分子筛系统.

Oriented synthesis of inorganic microporous materials is a hot issue in the field of international materials and molecular engineering.In view of the rich structure of open-framework aluminophosphates,we use machine learning-based theories and methods to analyze and predict aluminum phosphate syntheses in this paper based on the database established by"State Key Laboratory of Inorganic Synthesis and Preparative in Jilin University".Firstly,the rich information samples are used to the establishment of the prediction model and the missing value estimation model.Then,the missing data in the database is corrected by the BP missing value estimation model.Finally,a system of the rational design and synthesis of the formation of(6,12)-ringcontaining aluminophosphates was established according to the established model.