摘要
本文应用主成分分析(PCA)、遗传算法(GA)等方法进行变量选择确定对含能材料爆轰性能影响显著的分子结构描述符,再用偏最小二乘(PLS)和人工神经元网络等方法建立含能材料预测模型。通过预测模型进行计算,预测密度、爆速、生成焓和爆压等爆轰性能参数。与已知的爆轰性能参数比较,其准确度可达到98%。这说明了模型的准确性,可以用于未知含能材料的爆轰性能的预测。
The paper focuses on employing partial least squares(PLS) and artificial neural network( ANN ) based on principal component analysis(PCA) and generic algorithm( GA) input selection of molecular structure descriptors of energetic materials, a quantitative model about quantitative structure-detonation relationship(QSDR) was established. The QSDR model could predict the detonation property of energetic compound. The results showed that the yield model reflected the complex relationship between the structure and the detonation property, and had high predicting accuracy.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2007年第12期1714-1716,共3页
Computers and Applied Chemistry
基金
国防科工委重大专项及国家自然科学基金资助项目(No.20675063).
关键词
含能材料
分子设计
性能预估
energetic materials, molecular design, property prediction