基于QSPR方法的活性化合物热稳定性预测研究

Study on prediction of thermal stability of reactive chemicals based on quantitative structure-property relationship method

为形成安全、可靠、便捷的活性化合物热稳定性预测方法,快速获取活性化合物热稳定性参数,采用定量结构-性质相关性(QSPR)方法,针对38种有机过氧化物和104种硝基化合物的起始放热温度和分解热,结合遗传函数算法(GFA)和“断点原则”筛选出的分子描述符,利用遗传算法(GA)优化的BP神经网络,建立活性化合物的热稳定性GA-BP预测模型,验证分析模型的性能和应用域。研究结果表明:所建立的GA-BP模型具有良好的拟合能力、稳定性和预测能力,优于线性模型,说明活性化合物热稳定性与分子结构之间存在非线性关系;同时,得出影响活性化合物热稳定性参数的主要结构因素。

In order to form a safe,reliable and convenient prediction method for the thermal stability of reactive chemicals,and quickly obtain the thermal stability parameters of reactive chemicals,a quantitative structure-property relationship(QSPR)method was applied to predict the detected exothermic onset temperature and the heat of decomposition of 38 organic peroxides and 104 nitro compounds.The molecular descriptors selected by the genetic function algorithm(GFA)and the breaking point rule were adopted to establish the GA-BP prediction model of thermal stability for the reactive chemicals by using the error back propagation(BP)neural network optimized by genetic algorithm(GA),and the performance and application domain of the model were verified and analyzed.The results showed that the GA-BP model had good fitting ability,stability and prediction ability,which was superior to the linear model,indicating that there was a non-linear relationship between the thermal stability and molecular structure of reactive chemicals.In addition,the main structural factors affecting the thermal stability parameters of reactive chemicals were revealed.